DB2 FOR LUW / DB2FORLUW / 970 - DB2

Some partial writes/reads may occur during rebalance/reduce max
or lower high water mark operations causing SQLB_BADHDR errors,
FODC dumps and bringing down db2.

With this defect fix, a new registry variable is being
introduced so that when set, it will verify extents in the
following manner to prevent any future bad pages and/or dump
detailed diagnostic data into db2diag.log and into a binary dump
for further root cause analysis:

-New registry variable DB2_EM_PAGE_VERIFICATION (db2set
DB2_EM_PAGE_VERIFICATION=TRUE)
   -Extent verification during rebalance operations
      -Verifies that one full extent is read successfully, it
does not verify data consistency of extent's pages
      -If extent read was not successful, retry read 2 more
times
      -Verbose logging that dumps diagnostic data on failed
reads, and states if re-reads were successful
      -If read fails both re-reads, release latches and wait for
1 minute then retry extent move operation (retry forever)
   -Extent verification during reclaim storage (i.e. LOWER HIGH
WATER MARK, REDUCE MAX)
      -Verifies that each page of extent has correct checksum
pending the page header is not corrupted
      -If bad page identified, re-read extent up to 2 more times
      -If all re-reads fail, stop extent movement and return
error

****************************************************************
* USERS AFFECTED:                                              *
* ALL                                                          *
****************************************************************
* PROBLEM DESCRIPTION:                                         *
* See Error Description                                        *
****************************************************************
* RECOMMENDATION:                                              *
* This is currently a limitation                               *
****************************************************************

No Local Fix

The corrupted tables need to be re-created and re-loaded with
data. Some data loss is expected.

The data from corrupted tables can be extracted via db2dart
/DDEL command to be imported into newly created tables.

Link to db2dart command for more details:

https://www.ibm.com/support/knowledgecenter/SSEPGG_11.1.0/com.ib
m.db2.luw.admin.cmd.doc/doc/r0003477.html

not known / see Local fix

forerunner  : 
follow-up : IT20367 IT20368 

Date  - problem reported    :
Date  - problem closed      :
Date  - last modified       :

24.04.2017
09.05.2017
09.05.2017