Split Block

  • October 2019
  • PDF

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 Split Block as PDF for free.

More details

  • Words: 412
  • Pages: 1
during a hot backup, the data file is copied to another location by os. at this point of time, when oracle updates any block that is currently being copied by os, a fractured/split block occurs. an oracle block consists of one or more os blocks. by default, the size of a os block for windows is 2k. if the block size of oracle is 8k, then each oracle block will occupy 4 os blocks. if an oracle block contains 100 records, these 100 records will be stored in the 4 os blocks. when a file is copied by os, it follows an internal mechanism to copy the file. if the os file contains 1000 os blocks, we do not know which blocks will be copied first and which will be copied last. it may copy the os blocks in any order. when a tablespace is in backup mode and the file is being copied in os. 1) the os blocks could be updated before they are copied. 2) the os blocks could be updated after they are copied. both these conditions does not produce any defect. 3) suppose oracle is updating an oracle block that consists of 4 os blocks. in this case, after oracle updated the two os blocks and before updating the other 2 os blocks, any of the two os blocks are copied by os, the oracle block becomes fractured in the backup. there will be a change in the change vector # of the blocks. this is called split block or fractured block. create a table with char datatype char(2000) so that each oracle block stores 5-8 rows. insert as many as 2 lakh records. perform a logfile switch. try to update one row in each oracle block. perform a logfile switch. now, check the size of the generated logfile. perform a logfile switch. start a hot backup for the datafile in which the table is stored. update the same records that were updated before and end the backup. perform a logfile switch. compare the sizes of the two logfiles. this shows the difference in sizes of the two logfiles. when a datafile is in backup mode, and when a row is updated, the image of the whole block is stored in the logfile followed by the changes to the row. during recovery, whether the image of the blocks are applied followed by redo entries or only the redo is applied to the datafile without applying the image??????

Related Documents

Split Block
October 2019 37
Block
December 2019 64
Block
November 2019 57
Split System
May 2020 10
Split System
June 2020 10
Split Casing
December 2019 25