Monitoring Managing Data Centre
This document was uploaded by user and they confirmed that they have the permission to share it. If you are author or own the copyright of this book, please report to us by using this DMCA report form. Report DMCA
Overview
Download & View Monitoring Managing Data Centre as PDF for free.
More details
- Words: 17,175
- Pages: 93
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Section 5 - Monitoring and Managing the Data Center
Introduction
© 2006 EMC Corporation. All rights reserved.
Welcome to Section 5 of Storage Technology Foundations – Monitoring and Managing the Data Center. Copyright © 2006 EMC Corporation. All rights reserved. These materials may not be copied without EMC's written consent. EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice. THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Use, copying, and distribution of any EMC software described in this publication requires an applicable software license. EMC2, EMC, Navisphere, CLARiiON, and Symmetrix are registered trademarks and EMC Enterprise Storage, The Enterprise Storage Company, The EMC Effect, Connectrix, EDM, SDMS, SRDF, Timefinder, PowerPath, InfoMover, FarPoint, EMC Enterprise Storage Network, EMC Enterprise Storage Specialist, EMC Storage Logix, Universal Data Tone, E-Infostructure, Access Logix, Celerra, SnapView, and MirrorView are trademarks of EMC Corporation. All other trademarks used herein are the property of their respective owners.
Monitoring and Managing the Data Center - 1
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Section Objectives Upon completion of this section, you will be able to: y Describe areas of the data center to monitor y Discuss considerations for monitoring the data center y Describe techniques for managing the data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 2
The objectives for this section are shown here. Please take a moment to read them.
Monitoring and Managing the Data Center - 2
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
In This Section This section contains the following modules: y Monitoring in the Data Center y Managing in the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 3
This section contains 2 modules, monitoring in the data center and managing in the data center.
Monitoring and Managing the Data Center - 3
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge The following modules contain Apply Your Knowledge information (Available in the Student Resource Guide): y Monitoring in the Data Center y Managing in the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 4
Please note that certain modules of this section contain Apply Your Knowledge information that is only available in the Student Resource Guide
Monitoring and Managing the Data Center - 4
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring in the Data Center After completing this module, you will be able to: y Discuss data center areas to monitor y List metrics to monitor for different data center components y Describe the benefits of continuous monitoring y Describe the challenges in implementing a unified and centralized monitoring solution in heterogeneous environments y Describe industry standards for data center monitoring
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 5
In this module, you will learn about different aspects of monitoring data center components, including the benefits of pro-active monitoring and the challenges of managing a heterogeneous environment (multiple hardware/software from various vendors).
Monitoring and Managing the Data Center - 5
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Data Center Components Client
HBA Port HBA Keep Alive
Port IP
IP
Network
SAN Storage Arrays Health Capacity Performance
Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Security Monitoring and Managing the Data Center - 6
The Business Continuity Overview module discussed the importance of resolving all single points of failure when designing data centers. Having designed a resilient data center, the next step is to ensure that all components that make up the data center are functioning properly and are available on a 24x7 basis. The way to achieve this is by monitoring the data center on a continual basis. System Monitoring is essential to ensure that the underlying IT infrastructure business critical applications are operational and optimized. The main objective is to ensure that the various hosts, network systems and storage are running smoothly and to know how loaded each system and component is and how effectively it is being utilized. The major components within the data center that should be monitored include: y Servers, databases and applications y Network ((SAN) and IP Networks (switches, routers, bridges)) y Storage Arrays Each of these components should be monitored for health, capacity, performance, and security.
Monitoring and Managing the Data Center - 6
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Why Monitor Data Centers y Availability – Continuous monitoring ensures availability – Warnings and errors are fixed proactively
y Scalability – Monitoring allows for capacity planning/trend analysis which in turn helps to scale the data center as the business grows
y Alerting – Administrators can be informed of failures and potential failures – Corrective action can be taken to ensure availability and scalability
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 7
Continuous monitoring of health, capacity, performance and security of all data center components is critical to ensure data availability and scalability. For example, information about component failures can be sent to appropriate personnel for corrective actions. Ongoing trends show that the data storage environment continues to grow at a rapid pace. According to the International Data Corporation (IDC), external storage-system capacity growth will increase at a compound annual growth rate (CAGR) of approximately 50% through 2007. This represents a doubling of the current capacity every 2 years or so. Automated monitoring and alerting solutions are becoming increasingly important. Monitoring the data center closely and effectively optimizes data center operations and avoids downtime.
Monitoring and Managing the Data Center - 7
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health y Why monitor health of different components – Failure of any hardware/software component can lead to outage of a number of different components ¾ Example: A failed HBA could cause degraded access to a number of data devices in a multi-path environment or to loss of data access in a single path environment
y Monitoring health is fundamental and is easily understood and interpreted – At the very least health metrics should be monitored – Typically health issues would need to be addressed on a high priority
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 8
Health deals with the status/availability of a particular hardware component or a software process. (i.e., status of SAN device or port, database instance up/down, HBA status, disk/drive failure, etc.) If a component has failed, it could lead to down time unless redundancy exists. Monitoring the health of data center components is very important and is easy to understand and interpret (i.e., a component is either available or it has failed). Monitoring for capacity, performance, and security depend on the health and availability of different components.
Monitoring and Managing the Data Center - 8
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity y Why monitor capacity – Lack of proper capacity planning can lead to data un-availability and the ability to scale – Trend reports can be created from all the capacity data ¾ Enterprise is well informed of how IT resources are utilized
y Capacity monitoring prevents outages before they can occur – More preventive and predictive in nature than health metrics ¾ Based on reports one knows that 90% of a file system is full and that the file system is filling up at a particular rate ¾ 95% of all the ports have been utilized in a particular SAN fabric, a new switch should added if more arrays/servers are to be added to the same fabric
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 9
From a monitoring perspective, capacity deals with the amount of resources available. Examples: y Available free/used space on a file system or a database table space y Amount of space left in a RAID Group y Amount of disk space available on storage arrays y Amount of file system or mailbox quota allocated to users. y Number of available ports in a switch (e.g., 52 out of 64 ports in use, leaving 12 free ports for expansion)
Monitoring and Managing the Data Center - 9
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Performance y Why monitor Performance metrics – Want all data center components to work efficiently/optimally – See if components are pushing performance limits or if they are being under utilized – Can be used to identify performance bottlenecks
y Performance Monitoring/Analysis can be extremely complicated – Dozens of inter-related metrics depending on the component in question – Most complicated of the various aspects of monitoring
Monitoring and Managing the Data Center - 10
© 2006 EMC Corporation. All rights reserved.
Performance monitoring measures the efficiency of operation of different data center components. Examples: y Number of I/Os thorough a front-end port of a storage array y Number of I/Os to disks in a storage array y Response time of an application y Bandwidth utilization y Server CPU utilization
Monitoring and Managing the Data Center - 10
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security y Why monitor security – Prevent and track unauthorized access ¾ Accidental or malicious
y Enforcing security and monitoring for security breaches is a top priority for all businesses
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 11
Security prevents and tracks unauthorized access. Examples of security monitoring are: y Login failures y Unauthorized storage array configuration/re-configuration y Monitoring physical access (via badge readers, biometric scans, video cameras, etc.) y Unauthorized Zoning and LUN masking in SAN environments or changes to existing zones
Monitoring and Managing the Data Center - 11
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Servers y Health – Hardware components ¾ HBA, NIC, graphic card, internal disk …
– Status of various processes/applications
y Capacity – File system utilization
HBA HBA
– Database ¾ Table space/log space utilization
– User quota
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 12
Any failure of a hardware component such as HBA or NIC, should be immediately detected and rectified. As seen earlier, component redundancy can prevent total outage. Mission critical applications running on the servers should also be monitored continuously. A database might spawn a number of processes that are required to ensure operations. Failure of any of these processes can cause non-availability of the database. Databases and applications usually have mechanisms to detect such errors and report them. Capacity monitoring on a server will involve monitoring file system space utilization. By continuously monitoring file system free space, estimate the growth rate of the file system and effectively predict as to when it will become a 100% full. Corrective action such as extending the space of a file system can be taken well ahead of time to avoid a file system full condition. In many environments, system administrators enforce space utilization quota on users. For example, a user cannot exceed 10 GB of space or a particular file cannot be greater than 100 MB.
Monitoring and Managing the Data Center - 12
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Servers y Performance – CPU utilization – Memory utilization – Transaction response times
y Security – Login – Authorization
HBA HBA
– Physical security ¾ Data center access
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 13
Two key metrics of performance of servers are the CPU and memory utilization. A continuously high value (above 80%) for CPU utilization is an indication that the server is running out of processing power. During periods of high CPU utilization, applications running on the server, and consequently end-users of the application, will experience slower response times. Corrective action could include upgrading processors, adding more processors, shifting some applications to different Servers, or restricting the number of simultaneous client access. Databases, applications, and file systems utilize Server physical memory (RAM) to stage data for manipulation. When sufficient memory is not available, data has to be paged in and out of disks. This process will also result in slower response times. Login failures and attempts by unauthorized users to execute code or launch applications should be closely monitored to ensure secure operations.
Monitoring and Managing the Data Center - 13
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Health – Fabrics ¾ Fabric errors, zoning errors
– Ports ¾ Failed GBIC, status/attribute change
– Devices ¾ Status/attribute Change
– Hardware Components ¾ Processor cards, fans, power supplies
y Capacity – ISL utilization – Aggregate switch utilization – Port utilization © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 14
Uninterrupted access to data over the SAN depends on the health of its physical and logical components. The GBICs, power supplies, and fans in switches and cables are the physical components. Any failure in these must be immediately reported. Constructs such as zones and fabrics are the logical components. Errors in zoning such as specifying the wrong WWN of a port will result in failure to access that port. These have to be monitored, reported, and rectified as well. By way of capacity, the number of ports on different switches that are currently used/free should be monitored. This will aid in planning expansion by way of adding more Servers or connecting to more storage array ports. Utilization metrics at the switch level and port level, along with utilization of Interswitch Links (ISLs), are also a part of SAN capacity measurements. These can be viewed as being a part of performance metrics as well.
Monitoring and Managing the Data Center - 14
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Performance – Connectivity ports ¾ Link failures ¾ Loss of signal ¾ Loss of synchronization ¾ Link utilization ¾ Bandwidth MB/s or frames/s
– Connectivity devices ¾ Statistics are usually a cumulative value of all the port statistics
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 15
A number of SAN performance/statistical metrics can be used to determine/predict hardware failure (health). For example, an increasing number of link failures may indicate that a port is about to fail. The following are metrics which describe these failures: y Link Failures - the number of link failures occurring on a connectivity device port. A high number of failure could indicate a hardware problem (bad port, bad cable …) y Loss of Signal - the number of loss of signal events occurring on a connectivity device port. A high number indicates a possible hardware failure. y Loss of Synchronization - the number of loss of synchronization events occurring on a connectivity device port. High counts may indicate hardware failure. Connectivity device port performance can be measured with the Receive or Transmit Link Utilization metrics. These calculated values give a good indicator of how busy the switch port is based on the assumed maximum throughput. Heavily used ports can cause queuing delays on the host.
Monitoring and Managing the Data Center - 15
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Security – Zoning ¾ Ensure communication between dedicated sets of ports (HBA and Storage Ports)
– LUN Masking ¾ Ensure the only certain hosts have access to certain Storage Array volumes
– Administrative Tasks ¾ Restrict administrative tasks to a select set of users ¾ Enforce strict passwords
– Physical Security ¾ Access to Data Center should be monitored
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 16
SAN Security includes monitoring the fabrics for any zoning changes. Any errors in the zone set information can lead to data inaccessibility. Unauthorized zones can compromise data security. User login/authentication to switches should be monitored to audit administrative changes. Ensure that only authorized users are allowed to perform LUN masking tasks. Any such tasks performed should be audited for proper authorization.
Monitoring and Managing the Data Center - 16
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Health – All hardware components ¾ Front End ¾ Back End ¾ Memory ¾ Disks ¾ Power Supplies ¾…
– Array Operating Environment ¾ RAID processes ¾ Environmental Sensors ¾ Replication processes
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 17
Storage arrays typically have redundant components to function when individual components fail. Performance of the array might be affected during such failures. Failed components should be replaced quickly to restore optimal performance. Some arrays include the capability to send a message to the vendor’s support center in the event of hardware failures. This feature is typically known as “call-home”. It is equally important to monitor the various processes of the storage array operating environment. For example, failure of replication tasks will compromise disaster recovery capabilities.
Monitoring and Managing the Data Center - 17
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Capacity – Configured/unconfigured capacity – Allocated/unallocated storage – Fan-in/fan-out ratios
y Performance – Front End utilization/throughput – Back End utilization/throughput – I/O profile – Response time – Cache metrics
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 18
Physical disks in a storage array are partitioned into LUNs for use by hosts. y Configured capacity is the amount of space that has been partitioned into LUNs y Unconfigured capacity is the remaining space on the physical disks Allocated storage refers to LUNs that have been masked for use by specific hosts/servers. Unallocated storage refers to LUNs that have been configured, but not yet been masked for host use. Monitoring storage array capacity enables you to predict and react to storage needs as they occur. Fan-in/fan-out ratios and availability of unused front end ports (ports to which no host has yet been connected) is useful when new hosts/servers have to be given access to the storage array. Performance: Numerous performance/statistical metrics can be monitored for storage arrays. Some of the key metrics to monitor are the utilization rates of the various components that make up the storage arrays. Extremely high utilization rates can lead to performance degradation.
Monitoring and Managing the Data Center - 18
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Security – LUN Access ¾ Ensure the only certain hosts have access to certain Storage Array volumes ¾ Disallow WWN spoofing
– Administrative tasks ¾ Most arrays allow the restriction of various array configuration tasks Device configuration LUN masking Replication operations Port configuration
– Physical Security ¾ Monitor access to data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 19
World Wide Name (WWN) spoofing is a security concern. For example, an unauthorized host can be configured with a HBA that has the same WWN as another authorized host. If this host is now connected to the storage array via the same SAN, then zoning and LUN Masking restrictions will be bypassed. Storage arrays have mechanisms in place which can prevent such security breaches. Auditing of array device configuration tasks, as well as replication operations is important, to ensure that only authorized personnel are performing these.
Monitoring and Managing the Data Center - 19
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring IP Networks y Health – Hardware Components ¾ Processor cards, fans, Power Supplies, ...
– Cables
y Performance – Bandwidth – Latency
IP
– Packet Loss – Errors – Collisions
y Security © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 20
Network performance is vital in a storage environment. Monitor network latency, packet loss, availability, traffic, and bandwidth utilization for: − I/O (Bandwidth Usage) − Errors − Collisions
Monitoring and Managing the Data Center - 20
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the Data Center as a Whole y Monitor data center environment – Temperature, humidity, airflow, hazards (water, smoke, etc.) – Voltage – power supply
y Physical security – Facility access (Monitoring cameras, access cards, etc.)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 21
Monitoring the environment of a data center is just as crucial as monitoring the different components. Most electrical/electronic equipment are extremely sensitive to heat, humidity, voltage fluctuations, etc. Data center layout and design have to account for correct levels of ventilation, accurate control of temperature/humidity, uninterrupted power supplies, and corrections to voltage fluctuations. Any changes to the environment should be monitored and reported immediately. Physical security is easy to understand.
Monitoring and Managing the Data Center - 21
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
End-to-End Monitoring Client
HBA Port HBA IP
Network
Keep Alive
Port IP
SAN Storage Arrays Single Failure Multiple Symptoms Root Cause Analysis
Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Business Impact Monitoring and Managing the Data Center - 22
A good end-to-end monitoring system should be able to quickly analyze the impact that a single failure can cause. The monitoring system should be able to deduce that a set of seemingly unrelated symptoms are result of a root cause. It should also be able to alert on the impact to business arising from different component failures.
Monitoring and Managing the Data Center - 22
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: Array Port Failure H1
Degraded HBA HBA
SW1 H2
Degraded HBA
Port
HBA
Port
SW2 Storage Arrays
H3
Degraded HBA HBA
SAN
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 23
Here is an example of the importance of end-to-end monitoring. In this example, 3 Servers (H1, H2, and H3) have 2 HBA each and are connected to the storage array via two switches (SW1 and SW2). The three servers share the same storage ports on the Storage Array. If one of the storage array ports fails it will have the following effect on the whole data center: y Since all servers are sharing the ports, all the storage volumes that were accessed via SW1 will be unavailable. y The servers will experience path failures. Redundancy enables them to continue operations via SW2. y The applications will experience reduced performance (degraded), because the number of available paths to the storage devices has been cut in half. y If the applications belong to different business units all of these would be affected even though only a single port has failed. This example illustrates the importance of monitoring the health of storage arrays. By constantly monitoring the array, you can detect the fault as soon as it happens and fix it right away so as to minimize the time that applications have to run in a degraded mode.
Monitoring and Managing the Data Center - 23
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: HBA failure Degraded H1
HBA HBA
SW1 H2
HBA
Port
HBA
Port
SW2 Storage Arrays H3
HBA HBA
SAN
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 24
The scenario presented here is the same as the previous (3 Servers H1, H2 and H3 have 2 HBA each and are connected to the storage array via two switches SW1 and SW2. The three servers share the same storage ports on the storage array). In this example, if there is a single HBA failure, the server with the failed HBA will experience path failures to the storage devices that it had access to. Application performance on this server will be affected.
Monitoring and Managing the Data Center - 24
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: Switch Failure
SW1 Port Port
All Hosts Degraded
Port Port
SW2 Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Storage Arrays SAN
Monitoring and Managing the Data Center - 25
In this example, a number of servers (with 2 HBAs each) are connected to the storage array via two switches (SW1 and SW2). Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. What happens if there is a complete switch failure of switch SW1? All the hosts that were accessing storage volume via switch SW1 will experience path failures. All applications on the servers will run in a degraded mode. Notice that the failure of a single component (a switch in this case) has a ripple effect on many data center components.
Monitoring and Managing the Data Center - 25
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity: Array New Server SAN
SW1
Storage Array Port Port
SW2
Port Port
Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Can the Array provide the required storage to the new server?
Monitoring and Managing the Data Center - 26
This example illustrates the importance of monitoring the capacity of arrays. A number of servers (with 2 HBAs each) are connected to the storage array via two switches (SW1 and SW2). Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. Each of the servers has been allocated storage on the storage array. An application on the new server has to be given access to storage devices from the array, via switches SW1 and SW2.A new server has to be deployed. Monitoring the amount of configured and unconfigured space on the array is critical for deciding if this is possible. Proactive monitoring will help from the initial planning stages to final deployment.
Monitoring and Managing the Data Center - 26
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity: Servers File System Space No Monitoring
FS Monitoring File System
File System
Extend FS Warning: FS is 66% Full Critical: FS is 80% Full
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 27
This example illustrates the importance of monitoring capacity on servers. y On the left is an application server which is writing to a file system without monitoring the file system capacity. Once the file system is full, the application will no longer be able to function. y On the right is a similar setup. An application server is writing to a file system. In this case, the file system is monitored. A warning is issued at 66%, then a critical message at 80%. We can take action and extend the file system before the file system full condition is reached. Proactively monitoring the file system can prevent application outages caused by lack of file system space.
Monitoring and Managing the Data Center - 27
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
New Server H4
SW1
Port Util. %
Monitoring Performance: Array Port Utilization
SW2
Port
HBA HBA
H1
HBA HBA
H2
H3
100%
H1 + H2 + H3
HBA HBA
Port Port
HBA HBA
SAN Storage Arrays
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 28
This example illustrates the importance of monitoring performance metrics on storage arrays. Three Servers (H1, H2 and H3) have two HBAs each and are connected to the storage array via two switches (SW1 and SW2). The three servers share the same storage ports on the storage array. A new server H4 has to be deployed and must share the same storage ports as H1, H2 and H3. To ensure that the new server does not adversely affect the performance of the others, it is important to monitor the array port utilization. In this example, the utilization for the shared ports is shown using the green and red lines in the line graph. If the actual utilization prior to deploying the new server is the green line, then there is room to add the new server. Otherwise, the deployment of the new server will impact performance of all servers.
Monitoring and Managing the Data Center - 28
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Performance: Servers Critical: CPU Usage above 90% for the last 90 minutes
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 29
Most servers have tools that allow you to interactively monitor CPU usage. For example, Windows Task Manager displays the CPU and Memory usage (as shown above). Interactive tools are fine if only a few servers are being managed. In a data center, with potentially hundreds of servers, the tool must be capable of monitoring many servers simultaneously. Tools tool should send a warning to the System Administrator whenever the CPU utilization exceeds a specified threshold.
Monitoring and Managing the Data Center - 29
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security: Servers
Login 1 Login 2 Login 3 Critical: Three successive login failures for username “Bandit” on server “H4”, possible security threat
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 30
Login failures could be accidental (mistyping) or could be the result of a deliberate attempt to break into a system. Most servers will usually allow two successive login failures and will not allow any more attempts after a third successive login failure. In most environments, this information may simply be logged in a system log file. Ideally, you should monitor for such security events. In a monitored environment when there are three successive login failures, a message could be sent to the System Administrator to warn them of a possible security threat.
Monitoring and Managing the Data Center - 30
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security: Array – Local Replication
SAN SW1
Storage Array Port WG2
Workgroup 2 (WG1)
Port
SW2 Replication CMD
Workgroup 1 (WG2)
© 2006 EMC Corporation. All rights reserved.
Port
WG1
Port
Warning: Attempted replication of WG2 devices by WG1 user – Access denied Monitoring and Managing the Data Center - 31
This example illustrates the importance of monitoring security breaches in a storage array. A storage array is a shared resource. In this example, the array is being shared between two workgroups. The data of WG1 should not be accessible by WG2. Likewise, WG2 should not be accessible by WG1. A user from WG1 may try to make a local replica of the data that belongs to WG2. Typically, mechanisms will be in place to prevent such an action. If this action is not monitored or recorded in some fashion, be unaware that some one is trying to violate security protocols. But if this action is monitored, a warning message can be sent to the Storage Administrator.
Monitoring and Managing the Data Center - 31
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Alerting of Events y Warnings require administrative attention – File systems becoming full – Soft media errors
y Errors require immediate administrative attention – Power failures – Disk failures – Memory failures – Switch failures
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 32
Monitoring systems allow administrators to assign different severity levels for different conditions in the data center. Health related alerts will usually be classified as being critical or fatal – meaning that a failure in a component has immediate adverse consequences. Other alerts can be arranged in a spectrum from Information to Fatal. Generically: y Information – useful information requiring no administrator intervention, e.g. an authorized user has logged in y Warning – administrative attention is required, but the situation is not critical. An example may be that a file system has reached the 75% full mark. The administrator has time to decide what action should be taken y Fatal – immediate attention is required, because the condition will affect system performance or availability. If a disk fails, for example, the administrator must ensure that it is replaced quickly. The sources of monitoring messages may include hardware components, such as servers and storage systems, and software components, such as applications. Continuous monitoring, in combination with automated alerting, enables administrators to: y Reactively respond to failures quickly y Proactively avert failures by looking at trends in utilization and performance
Monitoring and Managing the Data Center - 32
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Challenges EMC Hitachi
Storage Arrays
NetApp CAS
NAS
HP
DAS IBM
SAN
Cisco
TLU
SUN
Network
Servers
McData MF
UNIX WIN
Databases Oracle
Informix
© 2006 EMC Corporation. All rights reserved.
SAN
Applications
IP Brocade
MS SQL Monitoring and Managing the Data Center - 33
The core elements of the data center are the storage arrays, networks, servers, databases, and applications. y Storage arrays could be NAS, CAS, DAS, SAN attached or Tape/Disk Library Units y The network consists of the SAN and the IP Network y Servers could be Open Systems (UNIX or Windows) or Mainframe. There are numerous vendors who supply these data center components The challenge is to monitor and manage each of these components. Typically, each vendor will provide monitoring/management tools for each of the components. As a consequence, in order to successfully monitor and manage a data center, learn multiple tools and terminologies. In an environment where multiple tools are in use, it is almost impossible to get a complete picture of what is going on a single page. Most data center components are inter-related (i.e. a SUN host is connected to a EMC storage array via a Cisco SAN). In an ideal world, the monitoring tool should be able to correlate the information from all objects in one place, so that you can make an informed decision on any of the metrics that is monitored.
Monitoring and Managing the Data Center - 33
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Ideal Solution
Monitoring/Management
One UI
Engine Storage Arrays
Storage Arrays
Network
CAS NAS
Servers, Databases, DAS
SAN
Applications TLU
Servers
Network MF
UNIX WIN
SAN Databases © 2006 EMC Corporation. All rights reserved.
IP
Applications Monitoring and Managing the Data Center - 34
The ideal solution to monitoring all data center objects from all the vendors would be a Monitoring/Management engine that would be able to gather information on all the objects and be able to manage all the same via a single user interface. The engine should also be able to perform root cause analysis and indicate how individual component failures affect various business units. y Single interface to monitor all objects in the data center y Root cause analysis - multiple symptoms may be triggered by single root cause y How to individual component failures affect various business units y Should have mechanism to inform administrators of events via e-mail/page/SNMP traps etc. y Should provide the ability to generate reports
Monitoring and Managing the Data Center - 34
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Without Standards y No common access layer between managed objects and applications – vendor specific y No common data model
Network Management Applications Management
y No interconnect independence Host Management
y Multi-layer management difficulty y Legacy systems can not be accommodated
Storage Management
y No multi-vendor automated discovery y Policy-based management is not possible across entire classes of devices
© 2006 EMC Corporation. All rights reserved.
Database Management
Interoperability!
Monitoring and Managing the Data Center - 35
SAN Administrators have often been faced with the dilemma of integrating multi-vendor hardware and software under a single management umbrella. It is relatively easy for administrators to monitor individual switches. But, monitoring a set of switches together and correlating data is a more complex challenge. Users and administrators want the flexibility to select the most suitable products for a particular application or set of applications and then easily integrate those products into their computing environments. Traditionally this has not been possible for the reasons listed above. Without standards, policy-based management is not possible across entire classes of devices. This poses a big dilemma for diverse environments.
Monitoring and Managing the Data Center - 35
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Simple Network Management Protocol (SNMP) y SNMP – Meant for network management – Inadequate for complete SAN Management
y Limitations of SNMP – No Common Object Model – Security - only newer SAN devices support v3 – Positive response mechanism – Inflexible - No auto discovery functions – No ACID (Atomicity, Consistency, Isolation, and Durability) properties – Richness of canonical intrinsic methods – Weak modeling constructs © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 36
Until recently, Simple Network Management Protocol (SNMP) has been a protocol of choice that has been used quite effectively to manage multi-vendor SAN environments. However, SNMP, being primarily a network management protocol, is inadequate when it comes to providing a detailed treatment on the fine grain elements in a SAN. Some of the limitations of SNMP are shown here. While SNMP still retains a predominant role in SAN management, newer and emerging standards may change this.
Monitoring and Managing the Data Center - 36
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Storage Management Initiative (SMI) y Created by the Storage Networking Industry Association (SNIA) y Integration of diverse multi-vendor storage networks
Management Application
y Development of more powerful management applications
Integration Infrastructure Object Model Mapping Vendor Unique Features
y Common interface for vendors to develop products that incorporate the management interface technology y Key components – – – – – – –
Inter-operability testing Education and collaboration Industry and customer promotion Promotions and demonstrations Technology center SMI specification Storage industry architects and developers
© 2006 EMC Corporation. All rights reserved.
•Platform Independent •Distributed •Automated Discovery •Security •Locking •Object Oriented
SMI-S Interface
CIM/WBEM Technology
Tape Library
Switch
Array
Many Other
MOF
MOF
MOF
MOF
Standard Object Model per Device Vendor Unique Function
Monitoring and Managing the Data Center - 37
The Storage Networking Industry Association (SNIA) has been engaged in an initiative to develop a common, open storage and SAN management interface based on the Distributed Management Task Force’s (DMTF) Common Information Model. This initiative is known as the Storage Management Initiative (SMI). One of the core objectives of this initiative is to create a standard that will be adopted by all Storage and SAN vendors, hardware and software alike, that will bring about true interoperability and allow administrators to manage multi-vendor and diverse storage networks using a single console or interface. The Storage Management Initiative Specification (SMI-S) offers substantial benefits to users and vendors. With SMI-S, developers have one complete, unified and rigidly specified object model, and can turn to one document to understand how to manage the breadth of SAN components. Management application vendors are relieved of the tedious task of integrating incompatible management interfaces, letting them focus on building management engines that reduce cost and extend functionality. And device vendors are empowered to build new features and functions into subsystems. SMI-S-compliant products will lead to easier, faster deployment and accelerated adoption of policy-based storage management frameworks. A test suite developed by the SNIA will certify compliance of hardware components and management applications with the specification. Certified components also will be subjected to rigorous interoperability testing in an SMI laboratory. Monitoring and Managing the Data Center - 37
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Storage Management Initiative Specification (SMI-S) y Based on: – Web Based Enterprise Management (WBEM) architecture – Common Information Model (CIM)
y Features: – A common interoperable and extensible management transport – A complete, unified and rigidly specified object model that provides for the control of a SAN – An automated discovery system – New approaches to the application of the CIM/WBEM technology
© 2006 EMC Corporation. All rights reserved.
Management
Graphical User Storage Resource Management Performance
Management Tools Container Management Volume Management Media Management Other
Capacity Planning Removable Media
Users Data Management File System Database Manager Backup and HSM
Storage Management Interface Specification Managed Objects Physical Components Removable Media Tape Drive Disk Drive Robot Enclosure Host Bus Adapter Switch
Logical Components Volume Clone Snapshot Media Set Zone Other
Monitoring and Managing the Data Center - 38
SMI-S forms a layer that resides between managed objects and managed applications. The following features of SMI-S provide the key to simplifying SAN management: y Common data model: SMI-S is based on Web Based Enterprise Management (WBEM) technology and the Common Information Model (CIM). SMI-S agents interrogate a device, such as a switch, host or storage array, extract the relevant management data from CIMenabled devices, and provide it to the requester. y Interconnect independence: SMI-S eliminates the need to redesign the management transport and lets components be managed using in-band or out-of-band communications, or a mix of the two. SMI-S offers further advantages by specifying the CMI-XML over HTTP protocol stack and utilizing the lower layers of the TCP/IP stack, both of which are ubiquitous in today's networking world. y Multilayer management: SMI-S has been developed to work with server-based volume managers, RAID systems and network storage appliances, a combination that most storage environments currently employ. y Legacy system accommodation: SMI-S has been developed to incorporate the management mechanisms in legacy devices with existing proprietary interfaces through the use of a proxy agent. Other devices and subsystems also can be integrated into an SMI-S network using embedded software or a CIM object manager. y Automated discovery: SMI-S-compliant products announce their presence and capabilities to other constituents. Combined with the automated discovery systems in WBEM to support object model extension, this will simplify management and give network managers the freedom to add components to their SAN more easily. y Policy-based management: SMI-S includes object models applicable across entire classes of devices, which lets SAN managers implement policy-based management for entire storage networks. Monitoring and Managing the Data Center - 38
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Common Information Model (CIM) y Describes the management of data y Details requirements within a domain y Information model with required syntax
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 39
The Common Information Model (CIM) is the language and methodology for describing management data. Information used to perform tasks is organized or structured to allow disparate groups of people to use it. This can be accomplished by developing a model or representation of the details required by people working within a particular domain. Such an approach can be referred to as an information model. An information model requires a set of legal statement types or syntax to capture the representation, and a collection of actual expressions necessary to manage common aspects of the domain. A CIM schema includes models for systems, applications, Networks (LAN), and devices. The CIM schema will enable applications from different developers on different platforms to describe management data in a standard format so that it can be shared among a variety of management applications.
Monitoring and Managing the Data Center - 39
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Web Based Enterprise Management (WBEM)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 40
Web Based Enterprise Management (WBEM) is a set of management and internet standard architectures developed by the Distributed Management Task Force (DMTF) to unify the management of enterprise computing environments, traditionally administered through traditional management stacks like SNMP and CMIP. WBEM provides the ability for the industry to deliver a well-integrated set of standard-based management tools leveraging emerging web technologies. The DMFT has developed a core set of standards that make up WBEM, which includes a data model, the CIM standard; an encoding specification, xml CIM encoding specification; and a transport mechanism, CIM Operation over HTTP .
Monitoring and Managing the Data Center - 40
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Enterprise Management Platforms (EMPs) y Graphical applications y Monitoring of many (if not all) data center components y Alerting of errors reported by those components y Management of many (if not all) data center components y Can often launch proprietary management applications y May include other functionality – Automatic provisioning – Scheduling of maintenance activities
y Proprietary architecture © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 41
Enterprise Management Platforms (EMPs) are complex applications, or suites of applications, that simplify the tasks of managing and monitoring data center environments. They will monitor data center components such as network switches, SAN switches, hosts, and alert the user of any problems with those components. At a minimum, the icon associated with the component in the GUI will change color to indicate its condition. Other forms of alerting, such as email or paging, may also be used. In addition to the monitoring functionality, management functionality is usually included as well. This may take the form of ‘native’ management by code embedded into the EMP, or may involve launching the proprietary management utility supplied by the manufacturer of the component. Other included functionality often allows easy scheduling of operations that must be performed on a regular basis, as well as provisioning of resources such as disk capacity.
Monitoring and Managing the Data Center - 41
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Module Summary Key points covered in this module: y It is important to continuously monitoring of data center components to support the availability and scalability initiatives of any business – Components include the server, SAN, network, and storage arrays
y The four areas of monitoring: – – – –
Health Capacity Performance Security
y There are attempts to define a common monitoring and management model © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 42
These are the key points covered in the module. Please take a moment to review them.
Monitoring and Managing the Data Center - 42
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge Upon completion of this topic, you will be able to: y Describe how EMC ControlCenter can be used to monitor the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 43
Monitoring and Managing the Data Center - 43
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Architecture User Interface Tier • Console (many) • Optional applications
Agent Tier • Master Agent (1) • Application Agents (many)
Infrastructure Tier • Server (one) • Repository (one) • Store (many) © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 44
EMC ControlCenter is a multi-tiered application, with multiple hosts running processes at each tier to support monitoring and management functions. The three tiers are: y User Interface Tier - the ControlCenter Console is an application that runs on a host and provides the main user interface for monitoring and managing the Storage Environment. y Infrastructure Tier - the ControlCenter Server, Repository, and one or more Stores provide central data storage and agent coordination at this tier. Infrastructure components can be installed on different hosts to allow a single infrastructure to scale to manage a large datacenter environment. ControlCenter Server and Store(s) are actually processes that run on a (usually) dedicated host. This host is referred to as the Infrastructure host. The repository is a database on the Infrastructure host. y Agent Tier - agents are responsible for gathering data about, and the management of different objects in the storage environment. Objects in a Storage Environment can be physical such as host, storage array, SAN switches or logical such as database, file system. Each host in the Storage Environment that needs to be monitored/managed via ControlCenter must have one Master Agent, and one Host Agent specific to that host type. The hosts can also have other agents to monitor/manage physical objects connected to them or logical objects residing on them. ControlCenter commands are passed from the console to the ControlCenter Server over a TCP/IP network, for execution. The ControlCenter Server then either retrieves the requested data from the Repository and returns it to the Console for display, or forwards the command to an agent. Agents pass the data they collect from the customer environment to the Store, which writes it to the Repository. Agents can also collect transient data, such as alerts and real-time performance data; they pass this directly to the ControlCenter Server for display on the console.
Monitoring and Managing the Data Center - 44
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Console y Primary interface through which the storage environment is viewed and managed y Java-based application supported on Windows and Solaris platforms y Objects managed by various agents are organized into groups such as Storage, Hosts, and Connectivity y Information about an object can be retrieved by the Console from the Repository or in real-time directly from the agent y Any command issued for the object is passed from the Console to the ControlCenter Server and handled appropriately y There can be several Consoles spread across the network © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 45
The ControlCenter console is the main management and monitoring interface for ControlCenter. It is a Java application supported on a Windows or Solaris host. The Console retrieves most of its information from data stored in the Repository via the ControlCenter Server. Property and configuration information about managed objects is reported by the agents to the Repository on a periodic basis, and a simple database query retrieves the data for use in Console displays. To provide immediate updates, real-time data such as object status changes or alert information is passed from the Agent directly to running Consoles via the ControlCenter Server. In all cases, the ControlCenter Server manages Console information gathering and presentation.
Monitoring and Managing the Data Center - 45
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Server y ControlCenter Server is the primary interface between the Console and the ControlCenter infrastructure y ControlCenter Server provides a diverse collection of services including: – Web Applications Server – used for installing the Java Console – Security and access management, such as licensing, login, authentication, and authorization – Communication with the Console – Alert and event management – Real-time statistics – Object management to maintain a list of managed objects – Agent management to maintain a list of available agents
y ControlCenter Server retrieves data from the Repository for display by the Java and Web Console y User initiated real-time data requests from some agents, are also handled by the ControlCenter Server y Balances Agent to Store communication based on workload © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 46
The ControlCenter Server handles data transfers between the Console and the Infrastructure. Much of the information presented to the Console is retrieved from stored records in the Repository. Real-time status updates, and alerts, are generated by the Agents and transferred directly to the Console by the ControlCenter Server.
Monitoring and Managing the Data Center - 46
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Repository y Licensed, embedded Oracle 9i database that holds current and historical information about the managed environment y ControlCenter Server executes transactions on the Repository to retrieve information requested by the Console y Store(s) populate the Repository with persistent data from the agents y Repository requires minimal user interaction or maintenance. The database has restricted access and can be updated only by ControlCenter applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 47
The ControlCenter Repository is an Oracle database, used for storing information about the managed environment and the objects therein. Data is entered and retrieved only through ControlCenter components: the Server retrieves information for user displays, while the Store populates the database with information from the agents. The Repository is a protected database—users can not directly access or change the tables or records. All access, including administration rights, is reserved by ControlCenter components. An automated task backs up the database daily.
Monitoring and Managing the Data Center - 47
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Store y Store receives the data sent by the agents, processes the data and updates the Repository y There can be multiple Stores in the environment, providing load balancing, scaling, and failover
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 48
The ControlCenter Store records data delivered by the Agents into the Repository. The Infrastructure can have multiple Stores. The ControlCenter Server load balances multiple Stores by choosing the Store with the lowest load for each Agent transaction. This provides fail over as well, since a new Store can be chosen at any time. ControlCenter scalability for large environments is in part achieved by adding multiple Stores as the number of managed objects grows.
Monitoring and Managing the Data Center - 48
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Agents y Master agent: – One per host – Manages other agents on the host – start/stop, monitor agent status and health
y ControlCenter Agents: – Runs on hosts to collect data and monitor object health – Generate alerts – Multiple agents can exist on a host – Passes information to the ControlCenter Store and the ControlCenter Server.
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 49
Agents monitor and issue management commands to objects. An agent of some type is needed for any object-related activity or information. A Master Agent must be running on any host that has any other ControlCenter Agents. The Master Agent starts, stops, and monitors the status of the other agents. The other ControlCenter Agents typically monitor and manage objects. Their primary function is to scan their managed object(s) at regular intervals set by ControlCenter Data Collection Policies (DCPs). Data is typically reported to a Store for addition to the Repository. Agents can also route information through the ControlCenter Server directly to a Console for immediate updates. Most Agents are very specific in their focus. A Storage Agent for CLARiiON can only manage CLARiiON arrays, for instance. Many Agents can monitor multiple objects at the same time.
Monitoring and Managing the Data Center - 49
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Storage Arrays The following Storage Arrays are supported by EMC ControlCenter y EMC Symmetrix y EMC CLARiiON y EMC Centera y EMC Celerra and Network Appliances NAS servers y EMC Invista y Hitachi Data Systems (including the HP and Sun resold versions) y HP Storageworks y IBM ESS y SMI-S (Storage Management Initiative Specification) compliant arrays © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 50
A large number of enterprise storage arrays are supported by ControlCenter, including EMC Symmetrix, CLARiiON, Centera, Invista, and the Celerra NAS (Network Attached Storage) server. Some level of support is also available for other vendor’s storage, including Network Appliance NAS servers, Hitachi Data Systems arrays, Hewlett-Packard Storageworks arrays, and IBM ESS (Shark) arrays. Further support is provided for any array that is SMI-S (Storage Management Initiative Specification) compliant. This new SNIA (Storage Industry Networking Association) initiative specifies a standard set of storage management commands. Arrays that are SMI-S compliant can be managed in the same way by one client—in this case, a ControlCenter agent. EMC and many other vendors are already building SMI-S compliance into their arrays.
Monitoring and Managing the Data Center - 50
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter support for SAN Devices The following SAN devices are supported by ControlCenter y EMC Connectrix y Brocade y McData y Cisco y Inrange (CNT) y IBM Blade Server (IBM-branded Brocade models only) y Dell Blade Server (Dell-branded Brocade models only)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 51
A variety of SAN devices can be managed with ControlCenter, including any EMC Connectrix switches. Fibre Channel switches from other vendors such as Brocade, McData, Cisco, and Inrange are also supported. Re-branded Brocade switches sold by IBM and Dell are also supported.
Monitoring and Managing the Data Center - 51
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Hosts The following hosts are supported by ControlCenter y Dedicated Host agents – – – – – – –
Microsoft Windows Hewlett-Packard HP-UX IBM AIX IBM mainframe Linux Novell Netware Sun Solaris
y Proxy management via Common Mapping Agent (CMA) – Compaq Tru64 – Fujitsu-Siemens BS2000 – Windows, Solaris, AIX, Linux, and HP-UX hosts can also be monitored by Common Mapping Agent proxy
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 52
Most host management is handled by dedicated ControlCenter agents. Usually, an operatingsystem specific agent must be installed on each host that you want to manage. Common Mapping Agent allows the management of hosts for which dedicated agents are not provided. Functionality provided by the Common Mapping Agent is limited in comparison to that provided by dedicated agents.
Monitoring and Managing the Data Center - 52
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Database and Backup The following databases are supported by ControlCenter y Dedicated database agent – Oracle – DB2 on mainframe
y Proxy management via Common Mapping Agent (CMA) – – – –
SQL Server Sybase Informix DB2
y Dedicated backup agent – – – –
EMC EDM IBM Tivoli EMC Networker Veritas Netbackup
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 53
ControlCenter can monitor several types of database and backup applications. Most are managed by dedicated agents that can manage only one type of application. However, the Common Mapping Agent can also manage several types of databases by proxy.
Monitoring and Managing the Data Center - 53
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Discovery of Managed Objects by Agents y Automatic Discovery: Many agents discover data objects automatically y Assisted Discovery: These agents must discover their objects by administrator action – Common Mapping Agent – Database Agent for Oracle – Fibre Channel Connectivity Agent – Storage Agents for CLARiiON, Centera, Invista, NAS, SMI, HP StorageWorks, HDS and ESS
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 54
Many ControlCenter agents automatically discover the object they monitor as soon as they are installed and started (e.g. Symmetrix, SDM, and Host Agents). The agent collects information about the object and forwards it to the Store. The Store then populates the Repository with this information. Some Objects, listed under Assisted Discovery, must be manually discovered. Typically, this happens when the agent must monitor the object via a network connection. An administrator can issue a manual discover command through the Console (the appropriate agent must be installed and running first). The dialogs available under this menu allow the administrator to choose the discovered object type, enter information (network address, version, etc.), and monitor the results of the discovery.
Monitoring and Managing the Data Center - 54
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Collection Policies (DCP) y Formal set of statements used to manage the data collected by ControlCenter agents y Policies specify the data to collect and the frequency of collection y ControlCenter agents have predefined collection policy definitions and templates – Default definitions can be easily modified, or new definitions can be created from the templates provided
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 55
Data Collection Policies are a formal set of statements that define how ControlCenter Agents gather and report configuration information about the objects that they manage. Data Collection Policies define which objects should be monitored and with what frequency they should be polled. By default, all agents of the same type discover their information at the same time every day. The defaults can be changed very easily to suit the environment.
Monitoring and Managing the Data Center - 55
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Console View of the Storage Environment
SAN Switch Server Dual HBAs
WWN of HBAs
Storage Array
Storage Array Front-end Directors and Ports © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 56
After discovery, the managed objects can be displayed in a number of ways in the ControlCenter Console. Shown here is the Topology View. This view shows the Server with its two HBAs (and their WWN), connected to a Storage Array via a SAN switch.
Monitoring and Managing the Data Center - 56
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Alerts - Overview y Why Alert? - Data availability – Monitor and report on events that could lead to application outages – Every ControlCenter agent can monitor a number of metrics ¾30 agents and 700+ alerts
y Alert categories – Health ¾Examples - Database instance up/down, Symmetrix service processor down, Connectivity device port status
– Capacity ¾Examples - File System Space, File/Directory Size Change
– Performance ¾Examples – Symmetrix Total Hit %, Host CPU Usage © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 57
Alerts are categorized by ControlCenter as being related to health, capacity, or performance. Alerts can be configured and customized by the administrator. Customization includes: y Setting threshold values to trigger alerts y Assigning severity (from Information to Fatal) based on threshold values y Specifying different means of Notification of Alerts y Include/exclude objects to be monitored for Alert conditions
Monitoring and Managing the Data Center - 57
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Alert Notification Notification capabilities y Messages are directed to the ControlCenter console by default y Messages can be directed to a Management Framework via Integration Gateway (SNMP) – governed by Management Policy associated with the Alert y E-mail notification as specified in the Management Policy
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 58
Listed are the notification capabilities of ControlCenter. In addition to these, one can specify custom scripts to be executed upon encountering different alert conditions.
Monitoring and Managing the Data Center - 58
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Console View of Alerts Message
Object Name
Alert state
© 2006 EMC Corporation. All rights reserved.
Severity Alert severity
Monitoring and Managing the Data Center - 59
This is the Alerts View. It is the main functional view for alerts, enabling you can quickly find the cause of the alert, which object is affected, and what the level of the alert is. The columns in this view include: y Alert State: Yellow Bell – New alert (Text will be in Bold font). Gray Bell – Acknowledged or Assigned Alert (Text is in normal font). y Severity: Ranges from 1 to 5. 1 = Fatal, 2= Critical, 3 = Warning, 4 = Minor, 5 = Information y Object Name: Host, storage array, network component (such as a switch), or other managed object for which the alert triggered. y Message: A description of the condition that caused the alert. Look here for information about the specific resources affected.
Monitoring and Managing the Data Center - 59
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Managing in the Data Center After completing this module, you will be able to: y Describe individual component tasks that would have to be performed in order to achieve overall data center management objectives y Explain the concept of Information Lifecycle Management
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 60
The objectives for this module are shown here. Please take a moment to review them.
Monitoring and Managing the Data Center - 60
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Managing Key Data Center Components Client
HBA Port HBA Keep Alive
Port IP
IP
Network
SAN Storage Arrays Availability
Performance Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Reporting
Capacity
Security Monitoring and Managing the Data Center - 61
In the module on Monitoring, we learned about the importance of monitoring the various data center components for Health, Capacity, Performance, and Security. In this section, we will focus on the various management tasks that need to be performed in order to ensure that Capacity, Availability, Performance, and Security requirements are met. The major components within the data center to be managed are: y IP Networks y Servers and all applications and databases running on the servers y Storage Area Network (SAN) y Storage Arrays Data Center Management can be broadly categorized as Capacity Management, Availability Management, Security Management, Performance Management and Reporting. Specific management tasks could address one or more of the categories. E.g. A LUN Masking task, addresses Capacity (storage capacity is provided to a specific host), Availability (if a device is masked via more than one path then single point of failure is eliminated), Security (masking prevents other hosts from accessing a given device) and Performance (if a device is accessible via multiple paths then host based multipathing software can improve performance by load balancing).
Monitoring and Managing the Data Center - 61
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management y Capacity Management – Allocation of adequate resources
y Availability Management – Business Continuity ¾ Eliminate single points of failure ¾ Backup & Restore ¾ Local & Remote Replication
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 62
Capacity Management ensures that there is adequate allocation of resources for all applications at all times. Capacity Management involves tasks that need to be performed on all data center components in order to achieve this goal. Let us take the example of allocating storage to a new applications that will be deployed on a new server from an intelligent storage array (we will explore this specific example in much more detail later in this module). To achieve this objective the following tasks would have to be performed on the storage array, the SAN and on the server: y Storage Array: Device configuration, LUN Masking y SAN: Unused Ports, Zoning y Server: HBA Configuration, host reconfiguration, file system management, application/database management Availability Management ensures business continuity by eliminating single points of failure in the environment and ensuring data availability though the use of backups, local replication and remote replication. Backup, local and remote replication have been discussed in Section 4 – Business Continuity. Availability management applies to all data center components. In this example, of a new application/server, availability is achieved as follows: y Server: At least two HBAs, multi-pathing software with path failover capability, Cluster, Backup. y SAN: Server is connected to the storage array via two independent SAN Fabrics, SAN switches themselves have built-in redundancy of various components. y Storage Array: Devices have some RAID protection, Array devices are made available to the host via at least two front-end ports (via independent SAN fabrics), Array has built-in redundancy for various components, local and remote replication, backup.
Monitoring and Managing the Data Center - 62
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management, continued y Security Management – Prevent unauthorized activities or access
y Performance Management – Configure/Design for optimal operational efficiency – Performance analysis ¾ Identify bottlenecks ¾ Recommend changes to improve performance
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 63
Security Management prevents unauthorized access to, and configuration tasks on, the data center components. Unauthorized access to data is prevented as well. In the new application/server deployment example, security management is addressed as follows: y Server: Creation of user logins, application/database logins, user privileges. Volume/Application/Database management can only be performed by authorized users. y SAN: Zoning (restricts access to front-end ports by specific HBAs). Administrative/Configuration operations can only be performed by authorized users. y Storage Array: LUN Masking (restrict access to specific devices by specific HBAs). Administrative/Configuration operations can only be performed by authorized users. Replication operations are restricted to authorized users as well. Performance Management ensures optimal operational efficiency of all data center components. Performance analysis of metrics collected is an important part of performance management and can be complicated because data center components are all inter-related. The performance of one component will have an impact on other components. In the new application/server deployment example performance management will involve: y Server: Volume Management, Database/Application layout, writing efficient applications, multiple HBAs and multi-pathing software with intelligent load balancing. y SAN: Design sufficient ISLs in a multi-switch fabric. Fabric design – core-edge, full mesh partial mesh … y Storage Array: Choice of RAID type and layout of the devices (LUNs) on the back-end of the array, choice of front-end ports (are the front-end ports being shared by multiple servers, are the ports maxed out), LUN Masking devices on multiple ports for multi-pathing. Monitoring and Managing the Data Center - 63
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management, continued y Reporting – Encompasses all data center components is used to provide information for Capacity, Availability, Security and Performance Management – Examples ¾ Capacity Planning Storage Utilization File System/Database Tablespace Utilzation Port usage
¾ Configuration/Asset Management Device Allocation Local/Remote Replica Fabric configuration – Zone and Zonesets Equipment on lease/rotation/refresh
¾ Chargeback Based on Allocation or Utilization
¾ Performance reports © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 64
Reports can be generated for all data center components. Data center reports can be used for trend analysis, capacity planning, chargeback, basic configuration information, etc.
Monitoring and Managing the Data Center - 64
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 1 – Storage Allocation to a New Server SAN
Host Storage Allocation Tasks
File / Database Mgmt
File System Mgmt
File System Host / Database Used Used
Volume Mgmt
Host Allocated
SAN Zoning
Array Assign Volumes Ports
Allocate Volumes Hosts
Reserved
Mapped
Config New Volumes
Unconfigured
Configured Volume Group Allocated
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 65
Let us explore the various management tasks with the help of an example. Let us assume that a new server has to be deployed in an existing SAN environment and has to be allocated storage from a storage array. The allocated storage is to be used by an application which uses a relational database. The database uses file systems. The picture breaks down the individual allocation tasks. We will explore the individual tasks in the next few slides. Storage Array Management y Configure new volumes on the array for use by the new server y Assign new volumes to the array front end ports SAN Management y Perform SAN Zoning – Zone the new servers HBAs via redundant fabrics to the front end ports of the storage Array y Perform LUN Masking on the storage array – Give the new server access to the new volumes via the array front end ports Host Storage Management y Configure HBAs on new server y Configure server to see new devices after zoning and LUN Masking is done y Volume Management (LVM tasks) y File System Management y Database/Application Management Monitoring and Managing the Data Center - 65
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Management – Allocation Tasks y Configure new volumes (LUNs) – Choose RAID type, size and number of volumes – Physical disks must have the required space available
y Assign volumes to array front end ports – This is automatic on some arrays while on others this step must be explicitly performed Intelligent Storage System Front End Host
Connectivity
Back End Cache
Physical Disks LUN 0
LUN 1 RAID 0 RAID 1 RAID 5 … © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 66
As we learned previously, the physical disks at the backend of the storage array are not directly presented as LUNs to a Host. Typically, a RAID Group or RAID set would be created and then LUNs could be created within the RAID set. These LUNs are then eventually presented to a host. These LUNs appear as physical disks from a host point of view. The space on the array physical disks that has not been configured for use as a host LUN is considered un-configured space and can be used to create more LUNs. Based on the storage requirements configure enough LUNs of the required size and RAID type. On many arrays, when the LUN is created, it is automatically assigned to the Front End ports of the array. On some arrays, the LUNs have to be explicitly assigned to array front end ports – this operation is called Mapping.
Monitoring and Managing the Data Center - 66
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Server Management – HBA Configuration y Server must have HBA hardware installed and configured
HBA Driver
New Server
Multi-path
– Install the HBA hardware and the software (device driver) and configure
HBA
HBA
y Optionally install multi-pathing software – Path failover and load balancing © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 67
The installation of the HBA hardware, software, and HBA configuration has to be performed before the server can be connected to the SAN. Multi-pathing software can be optionally installed. Most enterprises would opt to use multi-pathing because of availability requirements. Multi-pathing software can also perform load balancing, which will help performance.
Monitoring and Managing the Data Center - 67
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
SAN Management – Allocation Tasks y Perform Zoning – Zone the HBAs of the new server to the designated array front end ports via redundant fabrics ¾ Are there enough free ports on the switch? ¾ Did you check the array port utilization? Storage Array SW1 HBA
Port Port
HBA
New Server
Port
SW2
Port
y Perform LUN Masking – Grant the HBAs on the new server access to the LUNs on the array © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 68
Zoning and LUN Masking operations have been discussed in detail in the section on FC SAN. Zoning tasks are performed on the SAN Fabric. LUN Masking operations are typically performed on the storage array. The switches should have free ports available for the new server. Check the array port utilization if the port is shared between many servers.
Monitoring and Managing the Data Center - 68
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Server Management – Allocation y Reconfigure Server to see new devices y Perform Volume Management tasks y Perform Database/Application tasks
VG LV
HBA
FS HBA
DB App © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 69
Reconfigure Server to see new devices y Bus rescan or a reboot Perform Volume Management tasks y Create Volume Groups/Logical Volumes/File Systems − # of Logical Volumes/File Systems depends on how the database/application is to be laid out Database/Application tasks y Install database/application on the Logical Volumes/File Systems that were created y Startup database/application
Monitoring and Managing the Data Center - 69
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 2 – Running out of File System Space Solutions y Offload non-critical data File System
– Delete non-essential data – Move older/seldom used data to other media ¾ ILM/HSM strategy ¾ Easy retrieval if needed
y Extend File System
Warning: FS is 66% Full
– Operating System and Logical Volume Manager dependent – Management tasks seen in Scenario 1 will apply here as well
Critical: FS is 80% Full © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 70
In this scenario, we will explore data center management tasks that you would possibly have to execute to prevent a file system from getting 100% full. When a file system is running out of space, either: y Actively perform tasks which off load data from the existing file system (keep file system the same size) − Delete unwanted files − Offload files that have not been accessed for a long time to tape or to some other media from which it can be easily retrieved if necessary y Extend the file system to make it bigger − Considerations for extending file systems ¾ Dynamic extension of file systems is dependent on the specific operating system or logical volume manager (LVM) in use − The possible tasks to extend file systems is discussed in more detain in the next slide In reality, a good data center administrator should constantly monitor file systems and offload non-critical data and also be ready to extend the file system, if necessary.
Monitoring and Managing the Data Center - 70
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 2 – Running out of File System Space, continued Correlate File System with Volume Group or Disk Group.
Done No
Is there free space available in the VG?
Yes
No
Execute Command to extend File System.
Is the File System being replicated? Yes
Does the server have additional devices available?
Yes
Execute Command to extend VG.
No Does the Array have configured LUNs that can be allocated?
Yes
Perform tasks to ensure that the larger File System and Volume Group are replicated correctly
Allocate LUNs to server
No Does the array have unconfigured capacity? No © 2006 EMC Corporation. All rights reserved.
Yes
Configure new LUNs
Identify/Procure another array Monitoring and Managing the Data Center - 71
The steps/considerations prior to the extension of a file system have been illustrated in the flow chart. The goal is to increase the size of the file system to avoid application outage. Other considerations revolve around local/remote replication/protection employed for the application. For instance, if the application is protected via remote/local replication and a new device is added to the Volume Group, ensure that this new device is replicated as well. The steps include: y Correlate the file system to the logical volume and volume group if an LVM is in use y If there is enough space in the volume group – extend the file system y If the volume group does not have space – does the server have access to other devices which can be use to extend the volume group – extend the volume group – extend the file system y If the server does not have access to additional devices – allocate additional devices to the server – many or all of the steps discussed in scenario 1 will have to be used to do this (configure new LUNs on array, LUN mask, reconfigure server to recognize new devices – extend volume group – extend file system)
Monitoring and Managing the Data Center - 71
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report Storage Arrays VG
Production (Green)
LV FS VG
Remote Replica (Red)
DB App LV FS
Port DB AppVG
SW1
LV
Port
FS
Port
DB App
SW2 Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Port
Local Replica (Blue)
Monitoring and Managing the Data Center - 72
Scenario 3: In this scenario, we will explore the various data center tasks that will be necessary to create a specific report. A number of servers (50 – only 3 shown in picture) with 2 HBAs each and are connected to a Storage Array via two switches SW1 and SW2. Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. Applications are running on each of the servers, array replication technology is used to create local and remote replicas. The Production devices are represented by the green devices, local replica by the blue devices and the remote replicas by the red devices. A report documenting the exact amount of storage used by each application (including that used for local and remote replication) has to be created. The amount of raw storage used must be reported as well. The cost of the raw storage consumed by each application must be billed to the application owners. A sample report is shown in the picture. The report shows the information for two applications. Application Payroll_1 has been allocated 100 GB of storage. Production volumes are RAID 1 volumes hence the raw space used by the production volumes is 200 GB. Local replicas are on unprotected (no fault tolerance) volumes, hence raw space used by local replicas is 100 GB. The remote replicas are on RAID5 (5 disk group) volumes, hence raw space used for remote replicas is 125 GB. What are the various data center management steps to perform in order to create such a report?
Monitoring and Managing the Data Center - 72
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report – Tasks – Correlate Application Æ File Systems Æ Logical Volumes Æ Volume Groups Æ Host Physical Devices Æ Array Devices (Production) – Determine Array Devices used for Local Replication – Determine Array Devices used for Remote Replication – Determine storage allocated to application based on the size of the array devices
Example: VG
DB App
Array 1
Remote Array
LV
Source Vol 1
Local Replica Vol 1
Remote Replica Vol 1
FS
Source Vol 2
Local Replica Vol 2
Remote Replica Vol 2
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 73
The first step in determining the chargeback costs associated with an application is to correlate the application with the array devices that are in use. As indicated in the picture, trace the application to the file systems, logical volumes, volume groups, and eventually to the array devices. Since the applications are being replicated, determine the array devices used for local replication and the array devices used for remote replication. In the example shown, the application is using “Source Vol 1&2” (in Array 1). The replication devices are “Local Replica Vol 1&2” (in Array 1) and “Remote Replica Vol 1&2” (in the Remote Array). Keep in mind that this can change over time. As the application grows, more file systems and devices may be used. Thus, before a new report is generated, the correlation of application to the array devices should be done to ensure that the most current information is used. After the array devices are identified, the amount of storage allocated to the application can be easily computed. In this case “Source Vol 1&2” are each 10GB in size. Thus the storage allocated to the application is 20GB (10+10). The allocated storage for replication would be 20GB for local and 20GB for remote. The allocated storage is the actual storage that can be used, it does not represent the actual raw storage used by the application. To determine the raw space, determine the RAID protection that is used to the various array devices.
Monitoring and Managing the Data Center - 73
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report – Tasks, continued – Determine RAID type for Production/Local Replica/Remote Replica devices – Determine the total raw space allocated to application for production/local replication/remote replication – Compute the chargeback amount based of price/raw GB of storage – Repeat steps for each application and create report – Repeat the steps each time the report is to be created (weekly/monthly)
Example: 2 Source Vols = 2*10GB RAID 1 = 2* 20GB raw = 40GB 2 Local Replica Vols = 2*10GB = 2*10GB raw = 20GB 2 Remote Replica Vols = 2*10 GB RAID 5 = 2*12.5 GB raw = 25GB Total raw storage = 40+20+25 = 85GB Chargeback cost = 85*0.25/GB = 21.25 © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 74
To determine the raw space, review the steps displayed on the slide using the example listed. Determine RAID type for Production/Local Replica/Remote Replica devices. In the example shown, production devices are 10GB RAID 1, Local replica devices are 10GB with no protection, and remote replica devices are 10GB RAID 5 (5 disk group) devices. Determine the total raw space allocated to application for production, local replication, and remote replication. Based on the values from step 1, you can determine that the total raw space used by the application is 85GB. (Total raw storage = 40+20+25 = 85GB). Compute the chargeback amount based on price per raw GB of storage. Based on the cost per GB of storage (for the example this equals .25/GB), the chargeback cost can be computed. (Chargeback cost = 85*0.25/GB = 21.25). Repeat these steps for each application and create a report. Repeat the steps each time the report is to be created (weekly/monthly). The exercise would have to repeated for every single application in the enterprise in order to generate the require report. These tasks can be done manually. Manual creation of the report may be acceptable if only one or two applications exist. The process can become extremely tedious if many applications exist. The best way to create this report would be to automate these various tasks.
Monitoring and Managing the Data Center - 74
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management y Information Management Challenges y Information Lifecycle y Information Lifecycle Management – Definition – Process – Benefits – Implementation
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 75
Information Lifecycle Management (ILM) is a key approach for assuring availability, capacity, and performance. Let’s look at some of the aspects of ILM.
Monitoring and Managing the Data Center - 75
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Key Challenges of Information Management CHALLENGE
Scaling infrastructure within budget constraints
Information growth is relentless
CHALLENGE
Scaling resources to manage complexity
CHALLENGE
Information is more strategic than ever
Access, availability, and protection of critical information assets at optimal cost
Information changes in value over time
Ability to prioritize information management based on data value
CHALLENGE
Reducing risk of non-compliance CHALLENGE
© 2006 EMC Corporation. All rights reserved.
1 2 3 4 5
Monitoring and Managing the Data Center - 76
Companies face three key challenges related to information management: Strong growth of information: y Post-dot com rate of growth is around 50%, driven by digitization, increased use of e-mail, etc. y Just planning for growth can take up to 50% of storage resources y Meeting growth needs has increased the complexity of a customer environment Information is playing a more important role in determining business success: y New business applications provide more ways to extract a competitive advantage in the marketplace, e.g., companies like Dell, WalMart, and Amazon, where, at the heart of their respective business models, is the strategic use of information. Finally, information changes in value, and many times not necessarily in a linear fashion. y For example, customers become inactive, reducing the need for account information; pending litigation makes certain information more valuable, etc. y Understanding the value of information should be at the heart of managing information in general
Monitoring and Managing the Data Center - 76
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
The Information Lifecycle Sales Order Application Example
New Order Record
Order Processing
Warranty Claim
TIME
VALUE Orders Fulfilled
Protect Create
© 2006 EMC Corporation. All rights reserved.
Warranty Voided
Migrate Access
Dispose Archive
Monitoring and Managing the Data Center - 77
Information that is stored on a computer has a different value to a company, depending on how long it is stored on the network. In the above example, this sales order goes through differing value to the company from the time that it is created to the time that the warrantee is eventually voided. In a typical sales example as this one, the value of information is highest when a new order is created and processed. After order fulfillment, there is potentially less need to have real-time access to customer/order data, unless a warranty claim or other event triggers that need. Similarly, after the product has entered EOL, or after the account is closed, there is little value in the information and it can be disposed.
Monitoring and Managing the Data Center - 77
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Definition
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 78
Information Lifecycle Management is a strategy, not a product or service in itself; further, this strategy is proactive and dynamic in helping plan for IT growth as it relates to business needs, and reflects the value of information in a company. A successful information lifecycle management strategy must be: y Business-centric by tying closely with key processes, applications, and initiatives of the business y Centrally managed, providing an integrated view into all information assets of the business, both structured and unstructured y Policy-based, anchored in enterprise-wide information management policies that span all processes, applications, and resources y Heterogeneous, encompassing all types of platforms and operating systems y Aligned with the value of data, matching storage resources to the value of the data to the business at any given point in time
Monitoring and Managing the Data Center - 78
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Process Policy-based Alignment of Storage Infrastructure with Data Value
AUTOMATED Classify data / applications based on business rules
Implement policies with information management tools
Manage storage environment
Tier storage resources to align with data classes
FLEXIBLE Storage infrastructure that is
Application and Lifecycle Aware © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 79
The process of implementing the ongoing modification of an Information Lifecycle Management strategy consists of four activities: y Classify data and applications on the basis of business rules and policies to enable differentiated treatment of information y Implement policies with information management tools—from creation to disposal of data y Manage the environment with integrated tools that interface with multi-vendor platforms, and reduce operational complexity y Tier storage resources to align with data classes - storing information in the right type of infrastructure based on its current value.
Monitoring and Managing the Data Center - 79
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Benefits Information growth is relentless
Information is more strategic than ever Information changes in value over time
© 2006 EMC Corporation. All rights reserved.
1. Improve utilization of assets through tiered storage platforms 2. Simplify and automate management of information and storage infrastructure 3. Provide more cost-effective options for access, business continuity and protection 4. Ensure easy compliance through policy-based management 5. Deliver maximum value at lowest TCO by aligning storage infrastructure and management with information value
Monitoring and Managing the Data Center - 80
Implementing an ILM strategy delivers key benefits that directly address the challenges of information management. y Improved utilization, by the use of tiered platforms, and increased visibility into all enterprise information y Simplified management by integration of process steps and interfaces to individual tools in place today, and by increased automation y A wider range of options backup, protection, and recovery to balance the need for continuity with the cost of losing specific information y Painless compliance by having better control upfront in knowing what data needs to be protected and for how long y Lower TCO while meeting required service levels through aligning the infrastructure and management costs with information value so resources are not wasted or complexity introduced by managing low-value data at the cost of high-value data
Monitoring and Managing the Data Center - 80
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Path to Enterprise – Wide ILM App
App
App
Data
Data
Data
App
Automated Networked Storage
Step 1 Networked Tiered Storage y y y
Enable networked storage Automate environment Classify applications / data
App
App
Data
Data
App
ILM for Specific Applications
Step 2 Application-specific ILM y y
Define business policies for various information types Deploy ILM components into principal applications
App
App
Cross-Application ILM
Step 3 Enterprise-wide ILM y y y
Implement ILM across applications Policy-based automation Full visibility into all information
Lower cost through increased automation © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 81
Implementing ILM enterprise wide will take time, and no one believes it can be done instantaneously. A three step roadmap to enterprise-wide ILM is illustrated. y Step 1 and 2 are tuned to products and solutions available today, with the goal to be “ILMenabled” across a few enterprise-critical applications. In step 1, the goal is to get the environment to an automated networked storage environment. This is the basis for any policy-based information management. The value of tiered storage platforms can be exploited manually. In fact, many enterprises are already in this state. y Step 2 takes ILM to the next level with detailed application/data classification and linkage to business policies. While done in a manual way, the resultant policies can be automatically executed with tools for one or more applications, resulting in better management and optimal allocation of storage resources. y Step 3 of the vision is to automate more of the “front-end” or classification and policy management activities so as to scale to a wider set of enterprise applications. It is consistent with the need for more automation and greater simplicity of operations.
Monitoring and Managing the Data Center - 81
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Module Summary Key points covered in this module: y Individual component tasks that would have to be performed in order to achieve overall data center management objectives were illustrated – Allocation of storage to a new application server – Running out of file system space – Creating a chargeback report
y Concept of Information Lifecycle Management
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 82
These are the key points covered in this module. Please take a moment to review them.
Monitoring and Managing the Data Center - 82
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge Upon completion of this topic, you will be able to: y Describe how EMC ControlCenter can be used to manage the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 83
The next set of slides illustrate a small subset out of a vast set of management tasks that can be performed using EMC ControlCenter. Representative examples of EMC Symmetrix and CLARiiON array configurations, as well as reporting are presented.
Monitoring and Managing the Data Center - 83
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Configuration – Symmetrix
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 84
The drop-down menus shown allow users to perform an exhaustive set of array configuration tasks on the Symmetrix. The two tasks listed in the module – namely Configure new volumes and Assign volumes to array front end ports can be accomplished via the menu choices Logical Device Configuration and SDR Device Mapping respectively.
Monitoring and Managing the Data Center - 84
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Configuration - CLARiiON Some configuration options are only available when you select a specific RAID Group, Storage Group or LUN
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 85
Shown in this slide are the array configuration options for the EMC CLARiiON array. These include creation of new RAID groups, Binding LUNs within a RAID group, as well as certain operations on a RAID group.
Monitoring and Managing the Data Center - 85
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
SAN Management – Zoning Operations
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 86
As seen previously in this module, zoning and LUN masking are the two key SAN management allocation tasks. Zoning tasks such as creating/modifying/deleting zones and zone sets can be performed from EMC ControlCenter. Fabric level management tasks such as activating/deactivating zoning and importing zone sets can also be performed.
Monitoring and Managing the Data Center - 86
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Device Masking - Symmetrix
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 87
Shown here is an example of LUN masking on the EMC Symmetrix array from EMC ControlCenter. The menu driven user interface lets the administrator grant access to devices to different hosts connected to the Symmetrix via the same front-end director and port.
Monitoring and Managing the Data Center - 87
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Device Masking - CLARiiON
1
2 © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 88
This slide shows the LUN masking operation for a CLARiiON array. A host connected to a CLARiiON array is given access to LUNs by placing the host and the desired LUNs in the same Storage Group.
Monitoring and Managing the Data Center - 88
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Allocation Reports How much storage is allocated, and where is it?
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 89
EMC ControlCenter StorageScope is a part of the EMC ControlCenter software. StorageScope provides extensive reports on various components of the data center. Shown here are examples of two such reports – one from the hosts perspective (top) and one from the arrays perspective (bottom). Host-based allocation data shows how much storage has been allocated to a host, breaking the measures down by type: y Primary or Replica (mirror or copy storage) y Logical structure allocation like volume groups, logical volumes, filesystems, and databases (not shown in this example) y Array-based allocation data shows how much storage has been allocated for use. Many categories of allocated storage are reported: − Unconfigured (raw) or Configured (configured for use by hosts or the array itself) − Allocated (presented to hosts or used by the array itself) − Primary, Local Replica (same-array mirror or copy storage), or Remote Replica (cross array mirror) The array based report shows corresponding information for each of the arrays managed by EMC ControlCenter.
Monitoring and Managing the Data Center - 89
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Hosts Chargeback Report
y Chargeback summarizes host accessible storage devices, filesystems, and databases, independently dragged to the group
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 90
As seen previously in this module, Chargeback reporting is a key task in managing the data center. EMC ControlCenter StorageScope can generate summary reports on the amount of storage allocated to hosts/groups of hosts/functional organizations which own many groups of hosts etc.
Monitoring and Managing the Data Center - 90
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Summary This topic introduced a very small subset of data center management tasks that can be performed using EMC ControlCenter.
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 91
Monitoring and Managing the Data Center - 91
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Section Summary Key points covered in this section: y Areas of the data center to monitor y Considerations for monitoring the data center y Techniques for managing the data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 92
This completes Section 5 – Monitoring and Managing the Data Center. Please take a moment to review the key points covered in this section.
Monitoring and Managing the Data Center - 92
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Course Summary Key points covered in this course: y Storage concepts and architecture y Evolution of storage and storage environments y Logical and physical components of storage systems y Storage technologies and solutions y Core data center infrastructure elements and activities for monitoring and managing the data center y Options for Business Continuity
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 93
This completes the Storage Technology Foundations training. Please take a moment to review the key points covered in this course.
Monitoring and Managing the Data Center - 93
Section 5 - Monitoring and Managing the Data Center
Introduction
© 2006 EMC Corporation. All rights reserved.
Welcome to Section 5 of Storage Technology Foundations – Monitoring and Managing the Data Center. Copyright © 2006 EMC Corporation. All rights reserved. These materials may not be copied without EMC's written consent. EMC believes the information in this publication is accurate as of its publication date. The information is subject to change without notice. THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NO REPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THIS PUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Use, copying, and distribution of any EMC software described in this publication requires an applicable software license. EMC2, EMC, Navisphere, CLARiiON, and Symmetrix are registered trademarks and EMC Enterprise Storage, The Enterprise Storage Company, The EMC Effect, Connectrix, EDM, SDMS, SRDF, Timefinder, PowerPath, InfoMover, FarPoint, EMC Enterprise Storage Network, EMC Enterprise Storage Specialist, EMC Storage Logix, Universal Data Tone, E-Infostructure, Access Logix, Celerra, SnapView, and MirrorView are trademarks of EMC Corporation. All other trademarks used herein are the property of their respective owners.
Monitoring and Managing the Data Center - 1
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Section Objectives Upon completion of this section, you will be able to: y Describe areas of the data center to monitor y Discuss considerations for monitoring the data center y Describe techniques for managing the data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 2
The objectives for this section are shown here. Please take a moment to read them.
Monitoring and Managing the Data Center - 2
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
In This Section This section contains the following modules: y Monitoring in the Data Center y Managing in the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 3
This section contains 2 modules, monitoring in the data center and managing in the data center.
Monitoring and Managing the Data Center - 3
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge The following modules contain Apply Your Knowledge information (Available in the Student Resource Guide): y Monitoring in the Data Center y Managing in the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 4
Please note that certain modules of this section contain Apply Your Knowledge information that is only available in the Student Resource Guide
Monitoring and Managing the Data Center - 4
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring in the Data Center After completing this module, you will be able to: y Discuss data center areas to monitor y List metrics to monitor for different data center components y Describe the benefits of continuous monitoring y Describe the challenges in implementing a unified and centralized monitoring solution in heterogeneous environments y Describe industry standards for data center monitoring
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 5
In this module, you will learn about different aspects of monitoring data center components, including the benefits of pro-active monitoring and the challenges of managing a heterogeneous environment (multiple hardware/software from various vendors).
Monitoring and Managing the Data Center - 5
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Data Center Components Client
HBA Port HBA Keep Alive
Port IP
IP
Network
SAN Storage Arrays Health Capacity Performance
Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Security Monitoring and Managing the Data Center - 6
The Business Continuity Overview module discussed the importance of resolving all single points of failure when designing data centers. Having designed a resilient data center, the next step is to ensure that all components that make up the data center are functioning properly and are available on a 24x7 basis. The way to achieve this is by monitoring the data center on a continual basis. System Monitoring is essential to ensure that the underlying IT infrastructure business critical applications are operational and optimized. The main objective is to ensure that the various hosts, network systems and storage are running smoothly and to know how loaded each system and component is and how effectively it is being utilized. The major components within the data center that should be monitored include: y Servers, databases and applications y Network ((SAN) and IP Networks (switches, routers, bridges)) y Storage Arrays Each of these components should be monitored for health, capacity, performance, and security.
Monitoring and Managing the Data Center - 6
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Why Monitor Data Centers y Availability – Continuous monitoring ensures availability – Warnings and errors are fixed proactively
y Scalability – Monitoring allows for capacity planning/trend analysis which in turn helps to scale the data center as the business grows
y Alerting – Administrators can be informed of failures and potential failures – Corrective action can be taken to ensure availability and scalability
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 7
Continuous monitoring of health, capacity, performance and security of all data center components is critical to ensure data availability and scalability. For example, information about component failures can be sent to appropriate personnel for corrective actions. Ongoing trends show that the data storage environment continues to grow at a rapid pace. According to the International Data Corporation (IDC), external storage-system capacity growth will increase at a compound annual growth rate (CAGR) of approximately 50% through 2007. This represents a doubling of the current capacity every 2 years or so. Automated monitoring and alerting solutions are becoming increasingly important. Monitoring the data center closely and effectively optimizes data center operations and avoids downtime.
Monitoring and Managing the Data Center - 7
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health y Why monitor health of different components – Failure of any hardware/software component can lead to outage of a number of different components ¾ Example: A failed HBA could cause degraded access to a number of data devices in a multi-path environment or to loss of data access in a single path environment
y Monitoring health is fundamental and is easily understood and interpreted – At the very least health metrics should be monitored – Typically health issues would need to be addressed on a high priority
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 8
Health deals with the status/availability of a particular hardware component or a software process. (i.e., status of SAN device or port, database instance up/down, HBA status, disk/drive failure, etc.) If a component has failed, it could lead to down time unless redundancy exists. Monitoring the health of data center components is very important and is easy to understand and interpret (i.e., a component is either available or it has failed). Monitoring for capacity, performance, and security depend on the health and availability of different components.
Monitoring and Managing the Data Center - 8
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity y Why monitor capacity – Lack of proper capacity planning can lead to data un-availability and the ability to scale – Trend reports can be created from all the capacity data ¾ Enterprise is well informed of how IT resources are utilized
y Capacity monitoring prevents outages before they can occur – More preventive and predictive in nature than health metrics ¾ Based on reports one knows that 90% of a file system is full and that the file system is filling up at a particular rate ¾ 95% of all the ports have been utilized in a particular SAN fabric, a new switch should added if more arrays/servers are to be added to the same fabric
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 9
From a monitoring perspective, capacity deals with the amount of resources available. Examples: y Available free/used space on a file system or a database table space y Amount of space left in a RAID Group y Amount of disk space available on storage arrays y Amount of file system or mailbox quota allocated to users. y Number of available ports in a switch (e.g., 52 out of 64 ports in use, leaving 12 free ports for expansion)
Monitoring and Managing the Data Center - 9
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Performance y Why monitor Performance metrics – Want all data center components to work efficiently/optimally – See if components are pushing performance limits or if they are being under utilized – Can be used to identify performance bottlenecks
y Performance Monitoring/Analysis can be extremely complicated – Dozens of inter-related metrics depending on the component in question – Most complicated of the various aspects of monitoring
Monitoring and Managing the Data Center - 10
© 2006 EMC Corporation. All rights reserved.
Performance monitoring measures the efficiency of operation of different data center components. Examples: y Number of I/Os thorough a front-end port of a storage array y Number of I/Os to disks in a storage array y Response time of an application y Bandwidth utilization y Server CPU utilization
Monitoring and Managing the Data Center - 10
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security y Why monitor security – Prevent and track unauthorized access ¾ Accidental or malicious
y Enforcing security and monitoring for security breaches is a top priority for all businesses
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 11
Security prevents and tracks unauthorized access. Examples of security monitoring are: y Login failures y Unauthorized storage array configuration/re-configuration y Monitoring physical access (via badge readers, biometric scans, video cameras, etc.) y Unauthorized Zoning and LUN masking in SAN environments or changes to existing zones
Monitoring and Managing the Data Center - 11
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Servers y Health – Hardware components ¾ HBA, NIC, graphic card, internal disk …
– Status of various processes/applications
y Capacity – File system utilization
HBA HBA
– Database ¾ Table space/log space utilization
– User quota
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 12
Any failure of a hardware component such as HBA or NIC, should be immediately detected and rectified. As seen earlier, component redundancy can prevent total outage. Mission critical applications running on the servers should also be monitored continuously. A database might spawn a number of processes that are required to ensure operations. Failure of any of these processes can cause non-availability of the database. Databases and applications usually have mechanisms to detect such errors and report them. Capacity monitoring on a server will involve monitoring file system space utilization. By continuously monitoring file system free space, estimate the growth rate of the file system and effectively predict as to when it will become a 100% full. Corrective action such as extending the space of a file system can be taken well ahead of time to avoid a file system full condition. In many environments, system administrators enforce space utilization quota on users. For example, a user cannot exceed 10 GB of space or a particular file cannot be greater than 100 MB.
Monitoring and Managing the Data Center - 12
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Servers y Performance – CPU utilization – Memory utilization – Transaction response times
y Security – Login – Authorization
HBA HBA
– Physical security ¾ Data center access
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 13
Two key metrics of performance of servers are the CPU and memory utilization. A continuously high value (above 80%) for CPU utilization is an indication that the server is running out of processing power. During periods of high CPU utilization, applications running on the server, and consequently end-users of the application, will experience slower response times. Corrective action could include upgrading processors, adding more processors, shifting some applications to different Servers, or restricting the number of simultaneous client access. Databases, applications, and file systems utilize Server physical memory (RAM) to stage data for manipulation. When sufficient memory is not available, data has to be paged in and out of disks. This process will also result in slower response times. Login failures and attempts by unauthorized users to execute code or launch applications should be closely monitored to ensure secure operations.
Monitoring and Managing the Data Center - 13
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Health – Fabrics ¾ Fabric errors, zoning errors
– Ports ¾ Failed GBIC, status/attribute change
– Devices ¾ Status/attribute Change
– Hardware Components ¾ Processor cards, fans, power supplies
y Capacity – ISL utilization – Aggregate switch utilization – Port utilization © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 14
Uninterrupted access to data over the SAN depends on the health of its physical and logical components. The GBICs, power supplies, and fans in switches and cables are the physical components. Any failure in these must be immediately reported. Constructs such as zones and fabrics are the logical components. Errors in zoning such as specifying the wrong WWN of a port will result in failure to access that port. These have to be monitored, reported, and rectified as well. By way of capacity, the number of ports on different switches that are currently used/free should be monitored. This will aid in planning expansion by way of adding more Servers or connecting to more storage array ports. Utilization metrics at the switch level and port level, along with utilization of Interswitch Links (ISLs), are also a part of SAN capacity measurements. These can be viewed as being a part of performance metrics as well.
Monitoring and Managing the Data Center - 14
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Performance – Connectivity ports ¾ Link failures ¾ Loss of signal ¾ Loss of synchronization ¾ Link utilization ¾ Bandwidth MB/s or frames/s
– Connectivity devices ¾ Statistics are usually a cumulative value of all the port statistics
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 15
A number of SAN performance/statistical metrics can be used to determine/predict hardware failure (health). For example, an increasing number of link failures may indicate that a port is about to fail. The following are metrics which describe these failures: y Link Failures - the number of link failures occurring on a connectivity device port. A high number of failure could indicate a hardware problem (bad port, bad cable …) y Loss of Signal - the number of loss of signal events occurring on a connectivity device port. A high number indicates a possible hardware failure. y Loss of Synchronization - the number of loss of synchronization events occurring on a connectivity device port. High counts may indicate hardware failure. Connectivity device port performance can be measured with the Receive or Transmit Link Utilization metrics. These calculated values give a good indicator of how busy the switch port is based on the assumed maximum throughput. Heavily used ports can cause queuing delays on the host.
Monitoring and Managing the Data Center - 15
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the SAN y Security – Zoning ¾ Ensure communication between dedicated sets of ports (HBA and Storage Ports)
– LUN Masking ¾ Ensure the only certain hosts have access to certain Storage Array volumes
– Administrative Tasks ¾ Restrict administrative tasks to a select set of users ¾ Enforce strict passwords
– Physical Security ¾ Access to Data Center should be monitored
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 16
SAN Security includes monitoring the fabrics for any zoning changes. Any errors in the zone set information can lead to data inaccessibility. Unauthorized zones can compromise data security. User login/authentication to switches should be monitored to audit administrative changes. Ensure that only authorized users are allowed to perform LUN masking tasks. Any such tasks performed should be audited for proper authorization.
Monitoring and Managing the Data Center - 16
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Health – All hardware components ¾ Front End ¾ Back End ¾ Memory ¾ Disks ¾ Power Supplies ¾…
– Array Operating Environment ¾ RAID processes ¾ Environmental Sensors ¾ Replication processes
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 17
Storage arrays typically have redundant components to function when individual components fail. Performance of the array might be affected during such failures. Failed components should be replaced quickly to restore optimal performance. Some arrays include the capability to send a message to the vendor’s support center in the event of hardware failures. This feature is typically known as “call-home”. It is equally important to monitor the various processes of the storage array operating environment. For example, failure of replication tasks will compromise disaster recovery capabilities.
Monitoring and Managing the Data Center - 17
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Capacity – Configured/unconfigured capacity – Allocated/unallocated storage – Fan-in/fan-out ratios
y Performance – Front End utilization/throughput – Back End utilization/throughput – I/O profile – Response time – Cache metrics
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 18
Physical disks in a storage array are partitioned into LUNs for use by hosts. y Configured capacity is the amount of space that has been partitioned into LUNs y Unconfigured capacity is the remaining space on the physical disks Allocated storage refers to LUNs that have been masked for use by specific hosts/servers. Unallocated storage refers to LUNs that have been configured, but not yet been masked for host use. Monitoring storage array capacity enables you to predict and react to storage needs as they occur. Fan-in/fan-out ratios and availability of unused front end ports (ports to which no host has yet been connected) is useful when new hosts/servers have to be given access to the storage array. Performance: Numerous performance/statistical metrics can be monitored for storage arrays. Some of the key metrics to monitor are the utilization rates of the various components that make up the storage arrays. Extremely high utilization rates can lead to performance degradation.
Monitoring and Managing the Data Center - 18
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Storage Arrays y Security – LUN Access ¾ Ensure the only certain hosts have access to certain Storage Array volumes ¾ Disallow WWN spoofing
– Administrative tasks ¾ Most arrays allow the restriction of various array configuration tasks Device configuration LUN masking Replication operations Port configuration
– Physical Security ¾ Monitor access to data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 19
World Wide Name (WWN) spoofing is a security concern. For example, an unauthorized host can be configured with a HBA that has the same WWN as another authorized host. If this host is now connected to the storage array via the same SAN, then zoning and LUN Masking restrictions will be bypassed. Storage arrays have mechanisms in place which can prevent such security breaches. Auditing of array device configuration tasks, as well as replication operations is important, to ensure that only authorized personnel are performing these.
Monitoring and Managing the Data Center - 19
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring IP Networks y Health – Hardware Components ¾ Processor cards, fans, Power Supplies, ...
– Cables
y Performance – Bandwidth – Latency
IP
– Packet Loss – Errors – Collisions
y Security © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 20
Network performance is vital in a storage environment. Monitor network latency, packet loss, availability, traffic, and bandwidth utilization for: − I/O (Bandwidth Usage) − Errors − Collisions
Monitoring and Managing the Data Center - 20
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring the Data Center as a Whole y Monitor data center environment – Temperature, humidity, airflow, hazards (water, smoke, etc.) – Voltage – power supply
y Physical security – Facility access (Monitoring cameras, access cards, etc.)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 21
Monitoring the environment of a data center is just as crucial as monitoring the different components. Most electrical/electronic equipment are extremely sensitive to heat, humidity, voltage fluctuations, etc. Data center layout and design have to account for correct levels of ventilation, accurate control of temperature/humidity, uninterrupted power supplies, and corrections to voltage fluctuations. Any changes to the environment should be monitored and reported immediately. Physical security is easy to understand.
Monitoring and Managing the Data Center - 21
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
End-to-End Monitoring Client
HBA Port HBA IP
Network
Keep Alive
Port IP
SAN Storage Arrays Single Failure Multiple Symptoms Root Cause Analysis
Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Business Impact Monitoring and Managing the Data Center - 22
A good end-to-end monitoring system should be able to quickly analyze the impact that a single failure can cause. The monitoring system should be able to deduce that a set of seemingly unrelated symptoms are result of a root cause. It should also be able to alert on the impact to business arising from different component failures.
Monitoring and Managing the Data Center - 22
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: Array Port Failure H1
Degraded HBA HBA
SW1 H2
Degraded HBA
Port
HBA
Port
SW2 Storage Arrays
H3
Degraded HBA HBA
SAN
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 23
Here is an example of the importance of end-to-end monitoring. In this example, 3 Servers (H1, H2, and H3) have 2 HBA each and are connected to the storage array via two switches (SW1 and SW2). The three servers share the same storage ports on the Storage Array. If one of the storage array ports fails it will have the following effect on the whole data center: y Since all servers are sharing the ports, all the storage volumes that were accessed via SW1 will be unavailable. y The servers will experience path failures. Redundancy enables them to continue operations via SW2. y The applications will experience reduced performance (degraded), because the number of available paths to the storage devices has been cut in half. y If the applications belong to different business units all of these would be affected even though only a single port has failed. This example illustrates the importance of monitoring the health of storage arrays. By constantly monitoring the array, you can detect the fault as soon as it happens and fix it right away so as to minimize the time that applications have to run in a degraded mode.
Monitoring and Managing the Data Center - 23
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: HBA failure Degraded H1
HBA HBA
SW1 H2
HBA
Port
HBA
Port
SW2 Storage Arrays H3
HBA HBA
SAN
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 24
The scenario presented here is the same as the previous (3 Servers H1, H2 and H3 have 2 HBA each and are connected to the storage array via two switches SW1 and SW2. The three servers share the same storage ports on the storage array). In this example, if there is a single HBA failure, the server with the failed HBA will experience path failures to the storage devices that it had access to. Application performance on this server will be affected.
Monitoring and Managing the Data Center - 24
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Health: Switch Failure
SW1 Port Port
All Hosts Degraded
Port Port
SW2 Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Storage Arrays SAN
Monitoring and Managing the Data Center - 25
In this example, a number of servers (with 2 HBAs each) are connected to the storage array via two switches (SW1 and SW2). Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. What happens if there is a complete switch failure of switch SW1? All the hosts that were accessing storage volume via switch SW1 will experience path failures. All applications on the servers will run in a degraded mode. Notice that the failure of a single component (a switch in this case) has a ripple effect on many data center components.
Monitoring and Managing the Data Center - 25
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity: Array New Server SAN
SW1
Storage Array Port Port
SW2
Port Port
Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Can the Array provide the required storage to the new server?
Monitoring and Managing the Data Center - 26
This example illustrates the importance of monitoring the capacity of arrays. A number of servers (with 2 HBAs each) are connected to the storage array via two switches (SW1 and SW2). Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. Each of the servers has been allocated storage on the storage array. An application on the new server has to be given access to storage devices from the array, via switches SW1 and SW2.A new server has to be deployed. Monitoring the amount of configured and unconfigured space on the array is critical for deciding if this is possible. Proactive monitoring will help from the initial planning stages to final deployment.
Monitoring and Managing the Data Center - 26
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Capacity: Servers File System Space No Monitoring
FS Monitoring File System
File System
Extend FS Warning: FS is 66% Full Critical: FS is 80% Full
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 27
This example illustrates the importance of monitoring capacity on servers. y On the left is an application server which is writing to a file system without monitoring the file system capacity. Once the file system is full, the application will no longer be able to function. y On the right is a similar setup. An application server is writing to a file system. In this case, the file system is monitored. A warning is issued at 66%, then a critical message at 80%. We can take action and extend the file system before the file system full condition is reached. Proactively monitoring the file system can prevent application outages caused by lack of file system space.
Monitoring and Managing the Data Center - 27
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
New Server H4
SW1
Port Util. %
Monitoring Performance: Array Port Utilization
SW2
Port
HBA HBA
H1
HBA HBA
H2
H3
100%
H1 + H2 + H3
HBA HBA
Port Port
HBA HBA
SAN Storage Arrays
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 28
This example illustrates the importance of monitoring performance metrics on storage arrays. Three Servers (H1, H2 and H3) have two HBAs each and are connected to the storage array via two switches (SW1 and SW2). The three servers share the same storage ports on the storage array. A new server H4 has to be deployed and must share the same storage ports as H1, H2 and H3. To ensure that the new server does not adversely affect the performance of the others, it is important to monitor the array port utilization. In this example, the utilization for the shared ports is shown using the green and red lines in the line graph. If the actual utilization prior to deploying the new server is the green line, then there is room to add the new server. Otherwise, the deployment of the new server will impact performance of all servers.
Monitoring and Managing the Data Center - 28
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Performance: Servers Critical: CPU Usage above 90% for the last 90 minutes
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 29
Most servers have tools that allow you to interactively monitor CPU usage. For example, Windows Task Manager displays the CPU and Memory usage (as shown above). Interactive tools are fine if only a few servers are being managed. In a data center, with potentially hundreds of servers, the tool must be capable of monitoring many servers simultaneously. Tools tool should send a warning to the System Administrator whenever the CPU utilization exceeds a specified threshold.
Monitoring and Managing the Data Center - 29
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security: Servers
Login 1 Login 2 Login 3 Critical: Three successive login failures for username “Bandit” on server “H4”, possible security threat
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 30
Login failures could be accidental (mistyping) or could be the result of a deliberate attempt to break into a system. Most servers will usually allow two successive login failures and will not allow any more attempts after a third successive login failure. In most environments, this information may simply be logged in a system log file. Ideally, you should monitor for such security events. In a monitored environment when there are three successive login failures, a message could be sent to the System Administrator to warn them of a possible security threat.
Monitoring and Managing the Data Center - 30
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring Security: Array – Local Replication
SAN SW1
Storage Array Port WG2
Workgroup 2 (WG1)
Port
SW2 Replication CMD
Workgroup 1 (WG2)
© 2006 EMC Corporation. All rights reserved.
Port
WG1
Port
Warning: Attempted replication of WG2 devices by WG1 user – Access denied Monitoring and Managing the Data Center - 31
This example illustrates the importance of monitoring security breaches in a storage array. A storage array is a shared resource. In this example, the array is being shared between two workgroups. The data of WG1 should not be accessible by WG2. Likewise, WG2 should not be accessible by WG1. A user from WG1 may try to make a local replica of the data that belongs to WG2. Typically, mechanisms will be in place to prevent such an action. If this action is not monitored or recorded in some fashion, be unaware that some one is trying to violate security protocols. But if this action is monitored, a warning message can be sent to the Storage Administrator.
Monitoring and Managing the Data Center - 31
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Alerting of Events y Warnings require administrative attention – File systems becoming full – Soft media errors
y Errors require immediate administrative attention – Power failures – Disk failures – Memory failures – Switch failures
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 32
Monitoring systems allow administrators to assign different severity levels for different conditions in the data center. Health related alerts will usually be classified as being critical or fatal – meaning that a failure in a component has immediate adverse consequences. Other alerts can be arranged in a spectrum from Information to Fatal. Generically: y Information – useful information requiring no administrator intervention, e.g. an authorized user has logged in y Warning – administrative attention is required, but the situation is not critical. An example may be that a file system has reached the 75% full mark. The administrator has time to decide what action should be taken y Fatal – immediate attention is required, because the condition will affect system performance or availability. If a disk fails, for example, the administrator must ensure that it is replaced quickly. The sources of monitoring messages may include hardware components, such as servers and storage systems, and software components, such as applications. Continuous monitoring, in combination with automated alerting, enables administrators to: y Reactively respond to failures quickly y Proactively avert failures by looking at trends in utilization and performance
Monitoring and Managing the Data Center - 32
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Challenges EMC Hitachi
Storage Arrays
NetApp CAS
NAS
HP
DAS IBM
SAN
Cisco
TLU
SUN
Network
Servers
McData MF
UNIX WIN
Databases Oracle
Informix
© 2006 EMC Corporation. All rights reserved.
SAN
Applications
IP Brocade
MS SQL Monitoring and Managing the Data Center - 33
The core elements of the data center are the storage arrays, networks, servers, databases, and applications. y Storage arrays could be NAS, CAS, DAS, SAN attached or Tape/Disk Library Units y The network consists of the SAN and the IP Network y Servers could be Open Systems (UNIX or Windows) or Mainframe. There are numerous vendors who supply these data center components The challenge is to monitor and manage each of these components. Typically, each vendor will provide monitoring/management tools for each of the components. As a consequence, in order to successfully monitor and manage a data center, learn multiple tools and terminologies. In an environment where multiple tools are in use, it is almost impossible to get a complete picture of what is going on a single page. Most data center components are inter-related (i.e. a SUN host is connected to a EMC storage array via a Cisco SAN). In an ideal world, the monitoring tool should be able to correlate the information from all objects in one place, so that you can make an informed decision on any of the metrics that is monitored.
Monitoring and Managing the Data Center - 33
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Monitoring: Ideal Solution
Monitoring/Management
One UI
Engine Storage Arrays
Storage Arrays
Network
CAS NAS
Servers, Databases, DAS
SAN
Applications TLU
Servers
Network MF
UNIX WIN
SAN Databases © 2006 EMC Corporation. All rights reserved.
IP
Applications Monitoring and Managing the Data Center - 34
The ideal solution to monitoring all data center objects from all the vendors would be a Monitoring/Management engine that would be able to gather information on all the objects and be able to manage all the same via a single user interface. The engine should also be able to perform root cause analysis and indicate how individual component failures affect various business units. y Single interface to monitor all objects in the data center y Root cause analysis - multiple symptoms may be triggered by single root cause y How to individual component failures affect various business units y Should have mechanism to inform administrators of events via e-mail/page/SNMP traps etc. y Should provide the ability to generate reports
Monitoring and Managing the Data Center - 34
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Without Standards y No common access layer between managed objects and applications – vendor specific y No common data model
Network Management Applications Management
y No interconnect independence Host Management
y Multi-layer management difficulty y Legacy systems can not be accommodated
Storage Management
y No multi-vendor automated discovery y Policy-based management is not possible across entire classes of devices
© 2006 EMC Corporation. All rights reserved.
Database Management
Interoperability!
Monitoring and Managing the Data Center - 35
SAN Administrators have often been faced with the dilemma of integrating multi-vendor hardware and software under a single management umbrella. It is relatively easy for administrators to monitor individual switches. But, monitoring a set of switches together and correlating data is a more complex challenge. Users and administrators want the flexibility to select the most suitable products for a particular application or set of applications and then easily integrate those products into their computing environments. Traditionally this has not been possible for the reasons listed above. Without standards, policy-based management is not possible across entire classes of devices. This poses a big dilemma for diverse environments.
Monitoring and Managing the Data Center - 35
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Simple Network Management Protocol (SNMP) y SNMP – Meant for network management – Inadequate for complete SAN Management
y Limitations of SNMP – No Common Object Model – Security - only newer SAN devices support v3 – Positive response mechanism – Inflexible - No auto discovery functions – No ACID (Atomicity, Consistency, Isolation, and Durability) properties – Richness of canonical intrinsic methods – Weak modeling constructs © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 36
Until recently, Simple Network Management Protocol (SNMP) has been a protocol of choice that has been used quite effectively to manage multi-vendor SAN environments. However, SNMP, being primarily a network management protocol, is inadequate when it comes to providing a detailed treatment on the fine grain elements in a SAN. Some of the limitations of SNMP are shown here. While SNMP still retains a predominant role in SAN management, newer and emerging standards may change this.
Monitoring and Managing the Data Center - 36
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Storage Management Initiative (SMI) y Created by the Storage Networking Industry Association (SNIA) y Integration of diverse multi-vendor storage networks
Management Application
y Development of more powerful management applications
Integration Infrastructure Object Model Mapping Vendor Unique Features
y Common interface for vendors to develop products that incorporate the management interface technology y Key components – – – – – – –
Inter-operability testing Education and collaboration Industry and customer promotion Promotions and demonstrations Technology center SMI specification Storage industry architects and developers
© 2006 EMC Corporation. All rights reserved.
•Platform Independent •Distributed •Automated Discovery •Security •Locking •Object Oriented
SMI-S Interface
CIM/WBEM Technology
Tape Library
Switch
Array
Many Other
MOF
MOF
MOF
MOF
Standard Object Model per Device Vendor Unique Function
Monitoring and Managing the Data Center - 37
The Storage Networking Industry Association (SNIA) has been engaged in an initiative to develop a common, open storage and SAN management interface based on the Distributed Management Task Force’s (DMTF) Common Information Model. This initiative is known as the Storage Management Initiative (SMI). One of the core objectives of this initiative is to create a standard that will be adopted by all Storage and SAN vendors, hardware and software alike, that will bring about true interoperability and allow administrators to manage multi-vendor and diverse storage networks using a single console or interface. The Storage Management Initiative Specification (SMI-S) offers substantial benefits to users and vendors. With SMI-S, developers have one complete, unified and rigidly specified object model, and can turn to one document to understand how to manage the breadth of SAN components. Management application vendors are relieved of the tedious task of integrating incompatible management interfaces, letting them focus on building management engines that reduce cost and extend functionality. And device vendors are empowered to build new features and functions into subsystems. SMI-S-compliant products will lead to easier, faster deployment and accelerated adoption of policy-based storage management frameworks. A test suite developed by the SNIA will certify compliance of hardware components and management applications with the specification. Certified components also will be subjected to rigorous interoperability testing in an SMI laboratory. Monitoring and Managing the Data Center - 37
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Storage Management Initiative Specification (SMI-S) y Based on: – Web Based Enterprise Management (WBEM) architecture – Common Information Model (CIM)
y Features: – A common interoperable and extensible management transport – A complete, unified and rigidly specified object model that provides for the control of a SAN – An automated discovery system – New approaches to the application of the CIM/WBEM technology
© 2006 EMC Corporation. All rights reserved.
Management
Graphical User Storage Resource Management Performance
Management Tools Container Management Volume Management Media Management Other
Capacity Planning Removable Media
Users Data Management File System Database Manager Backup and HSM
Storage Management Interface Specification Managed Objects Physical Components Removable Media Tape Drive Disk Drive Robot Enclosure Host Bus Adapter Switch
Logical Components Volume Clone Snapshot Media Set Zone Other
Monitoring and Managing the Data Center - 38
SMI-S forms a layer that resides between managed objects and managed applications. The following features of SMI-S provide the key to simplifying SAN management: y Common data model: SMI-S is based on Web Based Enterprise Management (WBEM) technology and the Common Information Model (CIM). SMI-S agents interrogate a device, such as a switch, host or storage array, extract the relevant management data from CIMenabled devices, and provide it to the requester. y Interconnect independence: SMI-S eliminates the need to redesign the management transport and lets components be managed using in-band or out-of-band communications, or a mix of the two. SMI-S offers further advantages by specifying the CMI-XML over HTTP protocol stack and utilizing the lower layers of the TCP/IP stack, both of which are ubiquitous in today's networking world. y Multilayer management: SMI-S has been developed to work with server-based volume managers, RAID systems and network storage appliances, a combination that most storage environments currently employ. y Legacy system accommodation: SMI-S has been developed to incorporate the management mechanisms in legacy devices with existing proprietary interfaces through the use of a proxy agent. Other devices and subsystems also can be integrated into an SMI-S network using embedded software or a CIM object manager. y Automated discovery: SMI-S-compliant products announce their presence and capabilities to other constituents. Combined with the automated discovery systems in WBEM to support object model extension, this will simplify management and give network managers the freedom to add components to their SAN more easily. y Policy-based management: SMI-S includes object models applicable across entire classes of devices, which lets SAN managers implement policy-based management for entire storage networks. Monitoring and Managing the Data Center - 38
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Common Information Model (CIM) y Describes the management of data y Details requirements within a domain y Information model with required syntax
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 39
The Common Information Model (CIM) is the language and methodology for describing management data. Information used to perform tasks is organized or structured to allow disparate groups of people to use it. This can be accomplished by developing a model or representation of the details required by people working within a particular domain. Such an approach can be referred to as an information model. An information model requires a set of legal statement types or syntax to capture the representation, and a collection of actual expressions necessary to manage common aspects of the domain. A CIM schema includes models for systems, applications, Networks (LAN), and devices. The CIM schema will enable applications from different developers on different platforms to describe management data in a standard format so that it can be shared among a variety of management applications.
Monitoring and Managing the Data Center - 39
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Web Based Enterprise Management (WBEM)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 40
Web Based Enterprise Management (WBEM) is a set of management and internet standard architectures developed by the Distributed Management Task Force (DMTF) to unify the management of enterprise computing environments, traditionally administered through traditional management stacks like SNMP and CMIP. WBEM provides the ability for the industry to deliver a well-integrated set of standard-based management tools leveraging emerging web technologies. The DMFT has developed a core set of standards that make up WBEM, which includes a data model, the CIM standard; an encoding specification, xml CIM encoding specification; and a transport mechanism, CIM Operation over HTTP .
Monitoring and Managing the Data Center - 40
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Enterprise Management Platforms (EMPs) y Graphical applications y Monitoring of many (if not all) data center components y Alerting of errors reported by those components y Management of many (if not all) data center components y Can often launch proprietary management applications y May include other functionality – Automatic provisioning – Scheduling of maintenance activities
y Proprietary architecture © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 41
Enterprise Management Platforms (EMPs) are complex applications, or suites of applications, that simplify the tasks of managing and monitoring data center environments. They will monitor data center components such as network switches, SAN switches, hosts, and alert the user of any problems with those components. At a minimum, the icon associated with the component in the GUI will change color to indicate its condition. Other forms of alerting, such as email or paging, may also be used. In addition to the monitoring functionality, management functionality is usually included as well. This may take the form of ‘native’ management by code embedded into the EMP, or may involve launching the proprietary management utility supplied by the manufacturer of the component. Other included functionality often allows easy scheduling of operations that must be performed on a regular basis, as well as provisioning of resources such as disk capacity.
Monitoring and Managing the Data Center - 41
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Module Summary Key points covered in this module: y It is important to continuously monitoring of data center components to support the availability and scalability initiatives of any business – Components include the server, SAN, network, and storage arrays
y The four areas of monitoring: – – – –
Health Capacity Performance Security
y There are attempts to define a common monitoring and management model © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 42
These are the key points covered in the module. Please take a moment to review them.
Monitoring and Managing the Data Center - 42
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge Upon completion of this topic, you will be able to: y Describe how EMC ControlCenter can be used to monitor the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 43
Monitoring and Managing the Data Center - 43
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Architecture User Interface Tier • Console (many) • Optional applications
Agent Tier • Master Agent (1) • Application Agents (many)
Infrastructure Tier • Server (one) • Repository (one) • Store (many) © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 44
EMC ControlCenter is a multi-tiered application, with multiple hosts running processes at each tier to support monitoring and management functions. The three tiers are: y User Interface Tier - the ControlCenter Console is an application that runs on a host and provides the main user interface for monitoring and managing the Storage Environment. y Infrastructure Tier - the ControlCenter Server, Repository, and one or more Stores provide central data storage and agent coordination at this tier. Infrastructure components can be installed on different hosts to allow a single infrastructure to scale to manage a large datacenter environment. ControlCenter Server and Store(s) are actually processes that run on a (usually) dedicated host. This host is referred to as the Infrastructure host. The repository is a database on the Infrastructure host. y Agent Tier - agents are responsible for gathering data about, and the management of different objects in the storage environment. Objects in a Storage Environment can be physical such as host, storage array, SAN switches or logical such as database, file system. Each host in the Storage Environment that needs to be monitored/managed via ControlCenter must have one Master Agent, and one Host Agent specific to that host type. The hosts can also have other agents to monitor/manage physical objects connected to them or logical objects residing on them. ControlCenter commands are passed from the console to the ControlCenter Server over a TCP/IP network, for execution. The ControlCenter Server then either retrieves the requested data from the Repository and returns it to the Console for display, or forwards the command to an agent. Agents pass the data they collect from the customer environment to the Store, which writes it to the Repository. Agents can also collect transient data, such as alerts and real-time performance data; they pass this directly to the ControlCenter Server for display on the console.
Monitoring and Managing the Data Center - 44
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Console y Primary interface through which the storage environment is viewed and managed y Java-based application supported on Windows and Solaris platforms y Objects managed by various agents are organized into groups such as Storage, Hosts, and Connectivity y Information about an object can be retrieved by the Console from the Repository or in real-time directly from the agent y Any command issued for the object is passed from the Console to the ControlCenter Server and handled appropriately y There can be several Consoles spread across the network © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 45
The ControlCenter console is the main management and monitoring interface for ControlCenter. It is a Java application supported on a Windows or Solaris host. The Console retrieves most of its information from data stored in the Repository via the ControlCenter Server. Property and configuration information about managed objects is reported by the agents to the Repository on a periodic basis, and a simple database query retrieves the data for use in Console displays. To provide immediate updates, real-time data such as object status changes or alert information is passed from the Agent directly to running Consoles via the ControlCenter Server. In all cases, the ControlCenter Server manages Console information gathering and presentation.
Monitoring and Managing the Data Center - 45
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Server y ControlCenter Server is the primary interface between the Console and the ControlCenter infrastructure y ControlCenter Server provides a diverse collection of services including: – Web Applications Server – used for installing the Java Console – Security and access management, such as licensing, login, authentication, and authorization – Communication with the Console – Alert and event management – Real-time statistics – Object management to maintain a list of managed objects – Agent management to maintain a list of available agents
y ControlCenter Server retrieves data from the Repository for display by the Java and Web Console y User initiated real-time data requests from some agents, are also handled by the ControlCenter Server y Balances Agent to Store communication based on workload © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 46
The ControlCenter Server handles data transfers between the Console and the Infrastructure. Much of the information presented to the Console is retrieved from stored records in the Repository. Real-time status updates, and alerts, are generated by the Agents and transferred directly to the Console by the ControlCenter Server.
Monitoring and Managing the Data Center - 46
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Repository y Licensed, embedded Oracle 9i database that holds current and historical information about the managed environment y ControlCenter Server executes transactions on the Repository to retrieve information requested by the Console y Store(s) populate the Repository with persistent data from the agents y Repository requires minimal user interaction or maintenance. The database has restricted access and can be updated only by ControlCenter applications © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 47
The ControlCenter Repository is an Oracle database, used for storing information about the managed environment and the objects therein. Data is entered and retrieved only through ControlCenter components: the Server retrieves information for user displays, while the Store populates the database with information from the agents. The Repository is a protected database—users can not directly access or change the tables or records. All access, including administration rights, is reserved by ControlCenter components. An automated task backs up the database daily.
Monitoring and Managing the Data Center - 47
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Store y Store receives the data sent by the agents, processes the data and updates the Repository y There can be multiple Stores in the environment, providing load balancing, scaling, and failover
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 48
The ControlCenter Store records data delivered by the Agents into the Repository. The Infrastructure can have multiple Stores. The ControlCenter Server load balances multiple Stores by choosing the Store with the lowest load for each Agent transaction. This provides fail over as well, since a new Store can be chosen at any time. ControlCenter scalability for large environments is in part achieved by adding multiple Stores as the number of managed objects grows.
Monitoring and Managing the Data Center - 48
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Agents y Master agent: – One per host – Manages other agents on the host – start/stop, monitor agent status and health
y ControlCenter Agents: – Runs on hosts to collect data and monitor object health – Generate alerts – Multiple agents can exist on a host – Passes information to the ControlCenter Store and the ControlCenter Server.
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 49
Agents monitor and issue management commands to objects. An agent of some type is needed for any object-related activity or information. A Master Agent must be running on any host that has any other ControlCenter Agents. The Master Agent starts, stops, and monitors the status of the other agents. The other ControlCenter Agents typically monitor and manage objects. Their primary function is to scan their managed object(s) at regular intervals set by ControlCenter Data Collection Policies (DCPs). Data is typically reported to a Store for addition to the Repository. Agents can also route information through the ControlCenter Server directly to a Console for immediate updates. Most Agents are very specific in their focus. A Storage Agent for CLARiiON can only manage CLARiiON arrays, for instance. Many Agents can monitor multiple objects at the same time.
Monitoring and Managing the Data Center - 49
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Storage Arrays The following Storage Arrays are supported by EMC ControlCenter y EMC Symmetrix y EMC CLARiiON y EMC Centera y EMC Celerra and Network Appliances NAS servers y EMC Invista y Hitachi Data Systems (including the HP and Sun resold versions) y HP Storageworks y IBM ESS y SMI-S (Storage Management Initiative Specification) compliant arrays © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 50
A large number of enterprise storage arrays are supported by ControlCenter, including EMC Symmetrix, CLARiiON, Centera, Invista, and the Celerra NAS (Network Attached Storage) server. Some level of support is also available for other vendor’s storage, including Network Appliance NAS servers, Hitachi Data Systems arrays, Hewlett-Packard Storageworks arrays, and IBM ESS (Shark) arrays. Further support is provided for any array that is SMI-S (Storage Management Initiative Specification) compliant. This new SNIA (Storage Industry Networking Association) initiative specifies a standard set of storage management commands. Arrays that are SMI-S compliant can be managed in the same way by one client—in this case, a ControlCenter agent. EMC and many other vendors are already building SMI-S compliance into their arrays.
Monitoring and Managing the Data Center - 50
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter support for SAN Devices The following SAN devices are supported by ControlCenter y EMC Connectrix y Brocade y McData y Cisco y Inrange (CNT) y IBM Blade Server (IBM-branded Brocade models only) y Dell Blade Server (Dell-branded Brocade models only)
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 51
A variety of SAN devices can be managed with ControlCenter, including any EMC Connectrix switches. Fibre Channel switches from other vendors such as Brocade, McData, Cisco, and Inrange are also supported. Re-branded Brocade switches sold by IBM and Dell are also supported.
Monitoring and Managing the Data Center - 51
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Hosts The following hosts are supported by ControlCenter y Dedicated Host agents – – – – – – –
Microsoft Windows Hewlett-Packard HP-UX IBM AIX IBM mainframe Linux Novell Netware Sun Solaris
y Proxy management via Common Mapping Agent (CMA) – Compaq Tru64 – Fujitsu-Siemens BS2000 – Windows, Solaris, AIX, Linux, and HP-UX hosts can also be monitored by Common Mapping Agent proxy
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 52
Most host management is handled by dedicated ControlCenter agents. Usually, an operatingsystem specific agent must be installed on each host that you want to manage. Common Mapping Agent allows the management of hosts for which dedicated agents are not provided. Functionality provided by the Common Mapping Agent is limited in comparison to that provided by dedicated agents.
Monitoring and Managing the Data Center - 52
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Support for Database and Backup The following databases are supported by ControlCenter y Dedicated database agent – Oracle – DB2 on mainframe
y Proxy management via Common Mapping Agent (CMA) – – – –
SQL Server Sybase Informix DB2
y Dedicated backup agent – – – –
EMC EDM IBM Tivoli EMC Networker Veritas Netbackup
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 53
ControlCenter can monitor several types of database and backup applications. Most are managed by dedicated agents that can manage only one type of application. However, the Common Mapping Agent can also manage several types of databases by proxy.
Monitoring and Managing the Data Center - 53
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Discovery of Managed Objects by Agents y Automatic Discovery: Many agents discover data objects automatically y Assisted Discovery: These agents must discover their objects by administrator action – Common Mapping Agent – Database Agent for Oracle – Fibre Channel Connectivity Agent – Storage Agents for CLARiiON, Centera, Invista, NAS, SMI, HP StorageWorks, HDS and ESS
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 54
Many ControlCenter agents automatically discover the object they monitor as soon as they are installed and started (e.g. Symmetrix, SDM, and Host Agents). The agent collects information about the object and forwards it to the Store. The Store then populates the Repository with this information. Some Objects, listed under Assisted Discovery, must be manually discovered. Typically, this happens when the agent must monitor the object via a network connection. An administrator can issue a manual discover command through the Console (the appropriate agent must be installed and running first). The dialogs available under this menu allow the administrator to choose the discovered object type, enter information (network address, version, etc.), and monitor the results of the discovery.
Monitoring and Managing the Data Center - 54
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Collection Policies (DCP) y Formal set of statements used to manage the data collected by ControlCenter agents y Policies specify the data to collect and the frequency of collection y ControlCenter agents have predefined collection policy definitions and templates – Default definitions can be easily modified, or new definitions can be created from the templates provided
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 55
Data Collection Policies are a formal set of statements that define how ControlCenter Agents gather and report configuration information about the objects that they manage. Data Collection Policies define which objects should be monitored and with what frequency they should be polled. By default, all agents of the same type discover their information at the same time every day. The defaults can be changed very easily to suit the environment.
Monitoring and Managing the Data Center - 55
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Console View of the Storage Environment
SAN Switch Server Dual HBAs
WWN of HBAs
Storage Array
Storage Array Front-end Directors and Ports © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 56
After discovery, the managed objects can be displayed in a number of ways in the ControlCenter Console. Shown here is the Topology View. This view shows the Server with its two HBAs (and their WWN), connected to a Storage Array via a SAN switch.
Monitoring and Managing the Data Center - 56
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Alerts - Overview y Why Alert? - Data availability – Monitor and report on events that could lead to application outages – Every ControlCenter agent can monitor a number of metrics ¾30 agents and 700+ alerts
y Alert categories – Health ¾Examples - Database instance up/down, Symmetrix service processor down, Connectivity device port status
– Capacity ¾Examples - File System Space, File/Directory Size Change
– Performance ¾Examples – Symmetrix Total Hit %, Host CPU Usage © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 57
Alerts are categorized by ControlCenter as being related to health, capacity, or performance. Alerts can be configured and customized by the administrator. Customization includes: y Setting threshold values to trigger alerts y Assigning severity (from Information to Fatal) based on threshold values y Specifying different means of Notification of Alerts y Include/exclude objects to be monitored for Alert conditions
Monitoring and Managing the Data Center - 57
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Alert Notification Notification capabilities y Messages are directed to the ControlCenter console by default y Messages can be directed to a Management Framework via Integration Gateway (SNMP) – governed by Management Policy associated with the Alert y E-mail notification as specified in the Management Policy
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 58
Listed are the notification capabilities of ControlCenter. In addition to these, one can specify custom scripts to be executed upon encountering different alert conditions.
Monitoring and Managing the Data Center - 58
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
EMC ControlCenter Console View of Alerts Message
Object Name
Alert state
© 2006 EMC Corporation. All rights reserved.
Severity Alert severity
Monitoring and Managing the Data Center - 59
This is the Alerts View. It is the main functional view for alerts, enabling you can quickly find the cause of the alert, which object is affected, and what the level of the alert is. The columns in this view include: y Alert State: Yellow Bell – New alert (Text will be in Bold font). Gray Bell – Acknowledged or Assigned Alert (Text is in normal font). y Severity: Ranges from 1 to 5. 1 = Fatal, 2= Critical, 3 = Warning, 4 = Minor, 5 = Information y Object Name: Host, storage array, network component (such as a switch), or other managed object for which the alert triggered. y Message: A description of the condition that caused the alert. Look here for information about the specific resources affected.
Monitoring and Managing the Data Center - 59
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Managing in the Data Center After completing this module, you will be able to: y Describe individual component tasks that would have to be performed in order to achieve overall data center management objectives y Explain the concept of Information Lifecycle Management
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 60
The objectives for this module are shown here. Please take a moment to review them.
Monitoring and Managing the Data Center - 60
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Managing Key Data Center Components Client
HBA Port HBA Keep Alive
Port IP
IP
Network
SAN Storage Arrays Availability
Performance Cluster
Hosts/Servers with Applications © 2006 EMC Corporation. All rights reserved.
Reporting
Capacity
Security Monitoring and Managing the Data Center - 61
In the module on Monitoring, we learned about the importance of monitoring the various data center components for Health, Capacity, Performance, and Security. In this section, we will focus on the various management tasks that need to be performed in order to ensure that Capacity, Availability, Performance, and Security requirements are met. The major components within the data center to be managed are: y IP Networks y Servers and all applications and databases running on the servers y Storage Area Network (SAN) y Storage Arrays Data Center Management can be broadly categorized as Capacity Management, Availability Management, Security Management, Performance Management and Reporting. Specific management tasks could address one or more of the categories. E.g. A LUN Masking task, addresses Capacity (storage capacity is provided to a specific host), Availability (if a device is masked via more than one path then single point of failure is eliminated), Security (masking prevents other hosts from accessing a given device) and Performance (if a device is accessible via multiple paths then host based multipathing software can improve performance by load balancing).
Monitoring and Managing the Data Center - 61
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management y Capacity Management – Allocation of adequate resources
y Availability Management – Business Continuity ¾ Eliminate single points of failure ¾ Backup & Restore ¾ Local & Remote Replication
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 62
Capacity Management ensures that there is adequate allocation of resources for all applications at all times. Capacity Management involves tasks that need to be performed on all data center components in order to achieve this goal. Let us take the example of allocating storage to a new applications that will be deployed on a new server from an intelligent storage array (we will explore this specific example in much more detail later in this module). To achieve this objective the following tasks would have to be performed on the storage array, the SAN and on the server: y Storage Array: Device configuration, LUN Masking y SAN: Unused Ports, Zoning y Server: HBA Configuration, host reconfiguration, file system management, application/database management Availability Management ensures business continuity by eliminating single points of failure in the environment and ensuring data availability though the use of backups, local replication and remote replication. Backup, local and remote replication have been discussed in Section 4 – Business Continuity. Availability management applies to all data center components. In this example, of a new application/server, availability is achieved as follows: y Server: At least two HBAs, multi-pathing software with path failover capability, Cluster, Backup. y SAN: Server is connected to the storage array via two independent SAN Fabrics, SAN switches themselves have built-in redundancy of various components. y Storage Array: Devices have some RAID protection, Array devices are made available to the host via at least two front-end ports (via independent SAN fabrics), Array has built-in redundancy for various components, local and remote replication, backup.
Monitoring and Managing the Data Center - 62
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management, continued y Security Management – Prevent unauthorized activities or access
y Performance Management – Configure/Design for optimal operational efficiency – Performance analysis ¾ Identify bottlenecks ¾ Recommend changes to improve performance
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 63
Security Management prevents unauthorized access to, and configuration tasks on, the data center components. Unauthorized access to data is prevented as well. In the new application/server deployment example, security management is addressed as follows: y Server: Creation of user logins, application/database logins, user privileges. Volume/Application/Database management can only be performed by authorized users. y SAN: Zoning (restricts access to front-end ports by specific HBAs). Administrative/Configuration operations can only be performed by authorized users. y Storage Array: LUN Masking (restrict access to specific devices by specific HBAs). Administrative/Configuration operations can only be performed by authorized users. Replication operations are restricted to authorized users as well. Performance Management ensures optimal operational efficiency of all data center components. Performance analysis of metrics collected is an important part of performance management and can be complicated because data center components are all inter-related. The performance of one component will have an impact on other components. In the new application/server deployment example performance management will involve: y Server: Volume Management, Database/Application layout, writing efficient applications, multiple HBAs and multi-pathing software with intelligent load balancing. y SAN: Design sufficient ISLs in a multi-switch fabric. Fabric design – core-edge, full mesh partial mesh … y Storage Array: Choice of RAID type and layout of the devices (LUNs) on the back-end of the array, choice of front-end ports (are the front-end ports being shared by multiple servers, are the ports maxed out), LUN Masking devices on multiple ports for multi-pathing. Monitoring and Managing the Data Center - 63
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Data Center Management, continued y Reporting – Encompasses all data center components is used to provide information for Capacity, Availability, Security and Performance Management – Examples ¾ Capacity Planning Storage Utilization File System/Database Tablespace Utilzation Port usage
¾ Configuration/Asset Management Device Allocation Local/Remote Replica Fabric configuration – Zone and Zonesets Equipment on lease/rotation/refresh
¾ Chargeback Based on Allocation or Utilization
¾ Performance reports © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 64
Reports can be generated for all data center components. Data center reports can be used for trend analysis, capacity planning, chargeback, basic configuration information, etc.
Monitoring and Managing the Data Center - 64
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 1 – Storage Allocation to a New Server SAN
Host Storage Allocation Tasks
File / Database Mgmt
File System Mgmt
File System Host / Database Used Used
Volume Mgmt
Host Allocated
SAN Zoning
Array Assign Volumes Ports
Allocate Volumes Hosts
Reserved
Mapped
Config New Volumes
Unconfigured
Configured Volume Group Allocated
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 65
Let us explore the various management tasks with the help of an example. Let us assume that a new server has to be deployed in an existing SAN environment and has to be allocated storage from a storage array. The allocated storage is to be used by an application which uses a relational database. The database uses file systems. The picture breaks down the individual allocation tasks. We will explore the individual tasks in the next few slides. Storage Array Management y Configure new volumes on the array for use by the new server y Assign new volumes to the array front end ports SAN Management y Perform SAN Zoning – Zone the new servers HBAs via redundant fabrics to the front end ports of the storage Array y Perform LUN Masking on the storage array – Give the new server access to the new volumes via the array front end ports Host Storage Management y Configure HBAs on new server y Configure server to see new devices after zoning and LUN Masking is done y Volume Management (LVM tasks) y File System Management y Database/Application Management Monitoring and Managing the Data Center - 65
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Management – Allocation Tasks y Configure new volumes (LUNs) – Choose RAID type, size and number of volumes – Physical disks must have the required space available
y Assign volumes to array front end ports – This is automatic on some arrays while on others this step must be explicitly performed Intelligent Storage System Front End Host
Connectivity
Back End Cache
Physical Disks LUN 0
LUN 1 RAID 0 RAID 1 RAID 5 … © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 66
As we learned previously, the physical disks at the backend of the storage array are not directly presented as LUNs to a Host. Typically, a RAID Group or RAID set would be created and then LUNs could be created within the RAID set. These LUNs are then eventually presented to a host. These LUNs appear as physical disks from a host point of view. The space on the array physical disks that has not been configured for use as a host LUN is considered un-configured space and can be used to create more LUNs. Based on the storage requirements configure enough LUNs of the required size and RAID type. On many arrays, when the LUN is created, it is automatically assigned to the Front End ports of the array. On some arrays, the LUNs have to be explicitly assigned to array front end ports – this operation is called Mapping.
Monitoring and Managing the Data Center - 66
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Server Management – HBA Configuration y Server must have HBA hardware installed and configured
HBA Driver
New Server
Multi-path
– Install the HBA hardware and the software (device driver) and configure
HBA
HBA
y Optionally install multi-pathing software – Path failover and load balancing © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 67
The installation of the HBA hardware, software, and HBA configuration has to be performed before the server can be connected to the SAN. Multi-pathing software can be optionally installed. Most enterprises would opt to use multi-pathing because of availability requirements. Multi-pathing software can also perform load balancing, which will help performance.
Monitoring and Managing the Data Center - 67
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
SAN Management – Allocation Tasks y Perform Zoning – Zone the HBAs of the new server to the designated array front end ports via redundant fabrics ¾ Are there enough free ports on the switch? ¾ Did you check the array port utilization? Storage Array SW1 HBA
Port Port
HBA
New Server
Port
SW2
Port
y Perform LUN Masking – Grant the HBAs on the new server access to the LUNs on the array © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 68
Zoning and LUN Masking operations have been discussed in detail in the section on FC SAN. Zoning tasks are performed on the SAN Fabric. LUN Masking operations are typically performed on the storage array. The switches should have free ports available for the new server. Check the array port utilization if the port is shared between many servers.
Monitoring and Managing the Data Center - 68
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Server Management – Allocation y Reconfigure Server to see new devices y Perform Volume Management tasks y Perform Database/Application tasks
VG LV
HBA
FS HBA
DB App © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 69
Reconfigure Server to see new devices y Bus rescan or a reboot Perform Volume Management tasks y Create Volume Groups/Logical Volumes/File Systems − # of Logical Volumes/File Systems depends on how the database/application is to be laid out Database/Application tasks y Install database/application on the Logical Volumes/File Systems that were created y Startup database/application
Monitoring and Managing the Data Center - 69
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 2 – Running out of File System Space Solutions y Offload non-critical data File System
– Delete non-essential data – Move older/seldom used data to other media ¾ ILM/HSM strategy ¾ Easy retrieval if needed
y Extend File System
Warning: FS is 66% Full
– Operating System and Logical Volume Manager dependent – Management tasks seen in Scenario 1 will apply here as well
Critical: FS is 80% Full © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 70
In this scenario, we will explore data center management tasks that you would possibly have to execute to prevent a file system from getting 100% full. When a file system is running out of space, either: y Actively perform tasks which off load data from the existing file system (keep file system the same size) − Delete unwanted files − Offload files that have not been accessed for a long time to tape or to some other media from which it can be easily retrieved if necessary y Extend the file system to make it bigger − Considerations for extending file systems ¾ Dynamic extension of file systems is dependent on the specific operating system or logical volume manager (LVM) in use − The possible tasks to extend file systems is discussed in more detain in the next slide In reality, a good data center administrator should constantly monitor file systems and offload non-critical data and also be ready to extend the file system, if necessary.
Monitoring and Managing the Data Center - 70
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 2 – Running out of File System Space, continued Correlate File System with Volume Group or Disk Group.
Done No
Is there free space available in the VG?
Yes
No
Execute Command to extend File System.
Is the File System being replicated? Yes
Does the server have additional devices available?
Yes
Execute Command to extend VG.
No Does the Array have configured LUNs that can be allocated?
Yes
Perform tasks to ensure that the larger File System and Volume Group are replicated correctly
Allocate LUNs to server
No Does the array have unconfigured capacity? No © 2006 EMC Corporation. All rights reserved.
Yes
Configure new LUNs
Identify/Procure another array Monitoring and Managing the Data Center - 71
The steps/considerations prior to the extension of a file system have been illustrated in the flow chart. The goal is to increase the size of the file system to avoid application outage. Other considerations revolve around local/remote replication/protection employed for the application. For instance, if the application is protected via remote/local replication and a new device is added to the Volume Group, ensure that this new device is replicated as well. The steps include: y Correlate the file system to the logical volume and volume group if an LVM is in use y If there is enough space in the volume group – extend the file system y If the volume group does not have space – does the server have access to other devices which can be use to extend the volume group – extend the volume group – extend the file system y If the server does not have access to additional devices – allocate additional devices to the server – many or all of the steps discussed in scenario 1 will have to be used to do this (configure new LUNs on array, LUN mask, reconfigure server to recognize new devices – extend volume group – extend file system)
Monitoring and Managing the Data Center - 71
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report Storage Arrays VG
Production (Green)
LV FS VG
Remote Replica (Red)
DB App LV FS
Port DB AppVG
SW1
LV
Port
FS
Port
DB App
SW2 Hosts/Servers with Applications
© 2006 EMC Corporation. All rights reserved.
Port
Local Replica (Blue)
Monitoring and Managing the Data Center - 72
Scenario 3: In this scenario, we will explore the various data center tasks that will be necessary to create a specific report. A number of servers (50 – only 3 shown in picture) with 2 HBAs each and are connected to a Storage Array via two switches SW1 and SW2. Each server has independent paths (2 HBAs) to the storage array via switch SW1 and switch SW2. Applications are running on each of the servers, array replication technology is used to create local and remote replicas. The Production devices are represented by the green devices, local replica by the blue devices and the remote replicas by the red devices. A report documenting the exact amount of storage used by each application (including that used for local and remote replication) has to be created. The amount of raw storage used must be reported as well. The cost of the raw storage consumed by each application must be billed to the application owners. A sample report is shown in the picture. The report shows the information for two applications. Application Payroll_1 has been allocated 100 GB of storage. Production volumes are RAID 1 volumes hence the raw space used by the production volumes is 200 GB. Local replicas are on unprotected (no fault tolerance) volumes, hence raw space used by local replicas is 100 GB. The remote replicas are on RAID5 (5 disk group) volumes, hence raw space used for remote replicas is 125 GB. What are the various data center management steps to perform in order to create such a report?
Monitoring and Managing the Data Center - 72
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report – Tasks – Correlate Application Æ File Systems Æ Logical Volumes Æ Volume Groups Æ Host Physical Devices Æ Array Devices (Production) – Determine Array Devices used for Local Replication – Determine Array Devices used for Remote Replication – Determine storage allocated to application based on the size of the array devices
Example: VG
DB App
Array 1
Remote Array
LV
Source Vol 1
Local Replica Vol 1
Remote Replica Vol 1
FS
Source Vol 2
Local Replica Vol 2
Remote Replica Vol 2
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 73
The first step in determining the chargeback costs associated with an application is to correlate the application with the array devices that are in use. As indicated in the picture, trace the application to the file systems, logical volumes, volume groups, and eventually to the array devices. Since the applications are being replicated, determine the array devices used for local replication and the array devices used for remote replication. In the example shown, the application is using “Source Vol 1&2” (in Array 1). The replication devices are “Local Replica Vol 1&2” (in Array 1) and “Remote Replica Vol 1&2” (in the Remote Array). Keep in mind that this can change over time. As the application grows, more file systems and devices may be used. Thus, before a new report is generated, the correlation of application to the array devices should be done to ensure that the most current information is used. After the array devices are identified, the amount of storage allocated to the application can be easily computed. In this case “Source Vol 1&2” are each 10GB in size. Thus the storage allocated to the application is 20GB (10+10). The allocated storage for replication would be 20GB for local and 20GB for remote. The allocated storage is the actual storage that can be used, it does not represent the actual raw storage used by the application. To determine the raw space, determine the RAID protection that is used to the various array devices.
Monitoring and Managing the Data Center - 73
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Scenario 3 – Chargeback Report – Tasks, continued – Determine RAID type for Production/Local Replica/Remote Replica devices – Determine the total raw space allocated to application for production/local replication/remote replication – Compute the chargeback amount based of price/raw GB of storage – Repeat steps for each application and create report – Repeat the steps each time the report is to be created (weekly/monthly)
Example: 2 Source Vols = 2*10GB RAID 1 = 2* 20GB raw = 40GB 2 Local Replica Vols = 2*10GB = 2*10GB raw = 20GB 2 Remote Replica Vols = 2*10 GB RAID 5 = 2*12.5 GB raw = 25GB Total raw storage = 40+20+25 = 85GB Chargeback cost = 85*0.25/GB = 21.25 © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 74
To determine the raw space, review the steps displayed on the slide using the example listed. Determine RAID type for Production/Local Replica/Remote Replica devices. In the example shown, production devices are 10GB RAID 1, Local replica devices are 10GB with no protection, and remote replica devices are 10GB RAID 5 (5 disk group) devices. Determine the total raw space allocated to application for production, local replication, and remote replication. Based on the values from step 1, you can determine that the total raw space used by the application is 85GB. (Total raw storage = 40+20+25 = 85GB). Compute the chargeback amount based on price per raw GB of storage. Based on the cost per GB of storage (for the example this equals .25/GB), the chargeback cost can be computed. (Chargeback cost = 85*0.25/GB = 21.25). Repeat these steps for each application and create a report. Repeat the steps each time the report is to be created (weekly/monthly). The exercise would have to repeated for every single application in the enterprise in order to generate the require report. These tasks can be done manually. Manual creation of the report may be acceptable if only one or two applications exist. The process can become extremely tedious if many applications exist. The best way to create this report would be to automate these various tasks.
Monitoring and Managing the Data Center - 74
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management y Information Management Challenges y Information Lifecycle y Information Lifecycle Management – Definition – Process – Benefits – Implementation
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 75
Information Lifecycle Management (ILM) is a key approach for assuring availability, capacity, and performance. Let’s look at some of the aspects of ILM.
Monitoring and Managing the Data Center - 75
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Key Challenges of Information Management CHALLENGE
Scaling infrastructure within budget constraints
Information growth is relentless
CHALLENGE
Scaling resources to manage complexity
CHALLENGE
Information is more strategic than ever
Access, availability, and protection of critical information assets at optimal cost
Information changes in value over time
Ability to prioritize information management based on data value
CHALLENGE
Reducing risk of non-compliance CHALLENGE
© 2006 EMC Corporation. All rights reserved.
1 2 3 4 5
Monitoring and Managing the Data Center - 76
Companies face three key challenges related to information management: Strong growth of information: y Post-dot com rate of growth is around 50%, driven by digitization, increased use of e-mail, etc. y Just planning for growth can take up to 50% of storage resources y Meeting growth needs has increased the complexity of a customer environment Information is playing a more important role in determining business success: y New business applications provide more ways to extract a competitive advantage in the marketplace, e.g., companies like Dell, WalMart, and Amazon, where, at the heart of their respective business models, is the strategic use of information. Finally, information changes in value, and many times not necessarily in a linear fashion. y For example, customers become inactive, reducing the need for account information; pending litigation makes certain information more valuable, etc. y Understanding the value of information should be at the heart of managing information in general
Monitoring and Managing the Data Center - 76
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
The Information Lifecycle Sales Order Application Example
New Order Record
Order Processing
Warranty Claim
TIME
VALUE Orders Fulfilled
Protect Create
© 2006 EMC Corporation. All rights reserved.
Warranty Voided
Migrate Access
Dispose Archive
Monitoring and Managing the Data Center - 77
Information that is stored on a computer has a different value to a company, depending on how long it is stored on the network. In the above example, this sales order goes through differing value to the company from the time that it is created to the time that the warrantee is eventually voided. In a typical sales example as this one, the value of information is highest when a new order is created and processed. After order fulfillment, there is potentially less need to have real-time access to customer/order data, unless a warranty claim or other event triggers that need. Similarly, after the product has entered EOL, or after the account is closed, there is little value in the information and it can be disposed.
Monitoring and Managing the Data Center - 77
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Definition
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 78
Information Lifecycle Management is a strategy, not a product or service in itself; further, this strategy is proactive and dynamic in helping plan for IT growth as it relates to business needs, and reflects the value of information in a company. A successful information lifecycle management strategy must be: y Business-centric by tying closely with key processes, applications, and initiatives of the business y Centrally managed, providing an integrated view into all information assets of the business, both structured and unstructured y Policy-based, anchored in enterprise-wide information management policies that span all processes, applications, and resources y Heterogeneous, encompassing all types of platforms and operating systems y Aligned with the value of data, matching storage resources to the value of the data to the business at any given point in time
Monitoring and Managing the Data Center - 78
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Process Policy-based Alignment of Storage Infrastructure with Data Value
AUTOMATED Classify data / applications based on business rules
Implement policies with information management tools
Manage storage environment
Tier storage resources to align with data classes
FLEXIBLE Storage infrastructure that is
Application and Lifecycle Aware © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 79
The process of implementing the ongoing modification of an Information Lifecycle Management strategy consists of four activities: y Classify data and applications on the basis of business rules and policies to enable differentiated treatment of information y Implement policies with information management tools—from creation to disposal of data y Manage the environment with integrated tools that interface with multi-vendor platforms, and reduce operational complexity y Tier storage resources to align with data classes - storing information in the right type of infrastructure based on its current value.
Monitoring and Managing the Data Center - 79
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Information Lifecycle Management Benefits Information growth is relentless
Information is more strategic than ever Information changes in value over time
© 2006 EMC Corporation. All rights reserved.
1. Improve utilization of assets through tiered storage platforms 2. Simplify and automate management of information and storage infrastructure 3. Provide more cost-effective options for access, business continuity and protection 4. Ensure easy compliance through policy-based management 5. Deliver maximum value at lowest TCO by aligning storage infrastructure and management with information value
Monitoring and Managing the Data Center - 80
Implementing an ILM strategy delivers key benefits that directly address the challenges of information management. y Improved utilization, by the use of tiered platforms, and increased visibility into all enterprise information y Simplified management by integration of process steps and interfaces to individual tools in place today, and by increased automation y A wider range of options backup, protection, and recovery to balance the need for continuity with the cost of losing specific information y Painless compliance by having better control upfront in knowing what data needs to be protected and for how long y Lower TCO while meeting required service levels through aligning the infrastructure and management costs with information value so resources are not wasted or complexity introduced by managing low-value data at the cost of high-value data
Monitoring and Managing the Data Center - 80
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Path to Enterprise – Wide ILM App
App
App
Data
Data
Data
App
Automated Networked Storage
Step 1 Networked Tiered Storage y y y
Enable networked storage Automate environment Classify applications / data
App
App
Data
Data
App
ILM for Specific Applications
Step 2 Application-specific ILM y y
Define business policies for various information types Deploy ILM components into principal applications
App
App
Cross-Application ILM
Step 3 Enterprise-wide ILM y y y
Implement ILM across applications Policy-based automation Full visibility into all information
Lower cost through increased automation © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 81
Implementing ILM enterprise wide will take time, and no one believes it can be done instantaneously. A three step roadmap to enterprise-wide ILM is illustrated. y Step 1 and 2 are tuned to products and solutions available today, with the goal to be “ILMenabled” across a few enterprise-critical applications. In step 1, the goal is to get the environment to an automated networked storage environment. This is the basis for any policy-based information management. The value of tiered storage platforms can be exploited manually. In fact, many enterprises are already in this state. y Step 2 takes ILM to the next level with detailed application/data classification and linkage to business policies. While done in a manual way, the resultant policies can be automatically executed with tools for one or more applications, resulting in better management and optimal allocation of storage resources. y Step 3 of the vision is to automate more of the “front-end” or classification and policy management activities so as to scale to a wider set of enterprise applications. It is consistent with the need for more automation and greater simplicity of operations.
Monitoring and Managing the Data Center - 81
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Module Summary Key points covered in this module: y Individual component tasks that would have to be performed in order to achieve overall data center management objectives were illustrated – Allocation of storage to a new application server – Running out of file system space – Creating a chargeback report
y Concept of Information Lifecycle Management
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 82
These are the key points covered in this module. Please take a moment to review them.
Monitoring and Managing the Data Center - 82
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Apply Your Knowledge Upon completion of this topic, you will be able to: y Describe how EMC ControlCenter can be used to manage the Data Center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 83
The next set of slides illustrate a small subset out of a vast set of management tasks that can be performed using EMC ControlCenter. Representative examples of EMC Symmetrix and CLARiiON array configurations, as well as reporting are presented.
Monitoring and Managing the Data Center - 83
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Configuration – Symmetrix
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 84
The drop-down menus shown allow users to perform an exhaustive set of array configuration tasks on the Symmetrix. The two tasks listed in the module – namely Configure new volumes and Assign volumes to array front end ports can be accomplished via the menu choices Logical Device Configuration and SDR Device Mapping respectively.
Monitoring and Managing the Data Center - 84
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Configuration - CLARiiON Some configuration options are only available when you select a specific RAID Group, Storage Group or LUN
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 85
Shown in this slide are the array configuration options for the EMC CLARiiON array. These include creation of new RAID groups, Binding LUNs within a RAID group, as well as certain operations on a RAID group.
Monitoring and Managing the Data Center - 85
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
SAN Management – Zoning Operations
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 86
As seen previously in this module, zoning and LUN masking are the two key SAN management allocation tasks. Zoning tasks such as creating/modifying/deleting zones and zone sets can be performed from EMC ControlCenter. Fabric level management tasks such as activating/deactivating zoning and importing zone sets can also be performed.
Monitoring and Managing the Data Center - 86
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Device Masking - Symmetrix
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 87
Shown here is an example of LUN masking on the EMC Symmetrix array from EMC ControlCenter. The menu driven user interface lets the administrator grant access to devices to different hosts connected to the Symmetrix via the same front-end director and port.
Monitoring and Managing the Data Center - 87
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Array Device Masking - CLARiiON
1
2 © 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 88
This slide shows the LUN masking operation for a CLARiiON array. A host connected to a CLARiiON array is given access to LUNs by placing the host and the desired LUNs in the same Storage Group.
Monitoring and Managing the Data Center - 88
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Allocation Reports How much storage is allocated, and where is it?
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 89
EMC ControlCenter StorageScope is a part of the EMC ControlCenter software. StorageScope provides extensive reports on various components of the data center. Shown here are examples of two such reports – one from the hosts perspective (top) and one from the arrays perspective (bottom). Host-based allocation data shows how much storage has been allocated to a host, breaking the measures down by type: y Primary or Replica (mirror or copy storage) y Logical structure allocation like volume groups, logical volumes, filesystems, and databases (not shown in this example) y Array-based allocation data shows how much storage has been allocated for use. Many categories of allocated storage are reported: − Unconfigured (raw) or Configured (configured for use by hosts or the array itself) − Allocated (presented to hosts or used by the array itself) − Primary, Local Replica (same-array mirror or copy storage), or Remote Replica (cross array mirror) The array based report shows corresponding information for each of the arrays managed by EMC ControlCenter.
Monitoring and Managing the Data Center - 89
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Hosts Chargeback Report
y Chargeback summarizes host accessible storage devices, filesystems, and databases, independently dragged to the group
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 90
As seen previously in this module, Chargeback reporting is a key task in managing the data center. EMC ControlCenter StorageScope can generate summary reports on the amount of storage allocated to hosts/groups of hosts/functional organizations which own many groups of hosts etc.
Monitoring and Managing the Data Center - 90
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Summary This topic introduced a very small subset of data center management tasks that can be performed using EMC ControlCenter.
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 91
Monitoring and Managing the Data Center - 91
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Section Summary Key points covered in this section: y Areas of the data center to monitor y Considerations for monitoring the data center y Techniques for managing the data center
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 92
This completes Section 5 – Monitoring and Managing the Data Center. Please take a moment to review the key points covered in this section.
Monitoring and Managing the Data Center - 92
Copyright © 2006 EMC Corporation. Do not Copy - All Rights Reserved.
Course Summary Key points covered in this course: y Storage concepts and architecture y Evolution of storage and storage environments y Logical and physical components of storage systems y Storage technologies and solutions y Core data center infrastructure elements and activities for monitoring and managing the data center y Options for Business Continuity
© 2006 EMC Corporation. All rights reserved.
Monitoring and Managing the Data Center - 93
This completes the Storage Technology Foundations training. Please take a moment to review the key points covered in this course.
Monitoring and Managing the Data Center - 93
Related Documents
Monitoring Managing Data Centre
November 2019 4
Data Centre Monitoring
April 2020 4
Data Centre Automation
May 2020 2
Data Centre Asset Planning
May 2020 4
Managing Filestream Data Win32
June 2020 3