PM61424: DB2 RESTART FAILED WITH STORAGE ABEND 04E RC00E2000C IN DSNSVSVB OFFSET0AC2 & THE DBRT_POOL WAS FOUND TO BE FULL OF PDA BLOCKS

Obtain the fix for this APAR.

APAR status

• Closed as program error.

Error description

• V10 DB2 RESTART failed with storage ABEND04E 00E2000C DSNSVSVB
  OFFSET0AC2 as the DBRT_POOL was full of PDA blocks and DB2 start
  terminated with DSNV086E RC00D96001 . This was a SAP system and
  the problem seen during RESTART when rebuilding pageset
  information from checkpoint log records as the DBRT_POOL became
  full of PDA blocks for NPIs . In V9 ABENDs 04E 00E20004 in
  LOC DSNIRCSB DSNSVBK OFFSET05F8 and 04E 00C90101 in LOC DSNIDM
  DSNIRCSB ERQUAL5006 were also seen during RESTART, with
  MSGDSNV086E 00D96001 as the 1GB DBRT_POOL was full of PDA blocks
  for NPIs . From the dumps, a PHB pointed to the DBRT_POOL and
  3186 SHBs were chained for PDA NPI blocks in V9, but for V10
  3623 SHBs were chained for these PDA NPI blocks before DB2
  RESTART failed short on storage. A ZAP was produced in V9 to
  increase the DBRT_POOL's max allocation from 1G to 2G but on
  RESTART DB2 still failed ABEND04E RC00E20003 LOC DSNIRCSB
  DSNSVBK OFFSET05EE as virtual storage below the bar became full
  when 1.3G was allocated to the DBRT_POOL
  DB2STGLK/K
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED: All DB2 for z/OS users.                      *
  ****************************************************************
  * PROBLEM DESCRIPTION: DB2 RESTART FAILED WITH STORAGE ABEND   *
  *                      04E RC00E2000C IN DSNSVSVB OFFSET0AC2   *
  *                      AND ALL STORAGE USED FOR RESTART TABLE  *
  *                      STORAGE POOL WAS EXHAUSTED.             *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  During DB2 restart, Data Manager (DM) builds the Data Base
  Restart Table (DBRT) entry for each active (R/W state) page set.
  These entries are later used to open the pageset if necessary.
  However, each DBRT entry gets very large if there are
  Non-Partition Index (NPI) involved, especially, the NPIs on
  Partition By Growth (PBG) table spaces. If the DB2 restart
  involves several thousands of such PBGs and several thousands
  NPIs then the storage pool (ACOMPH12 or ACOMPH20) used to build
  DBRT entries gets exhausted which results in DB2 abnormally
  terminated with reason code RC00E2000C. To avoid this storage
  constraint during DB2 restart, Data Manager Restart (DMRR) logic
  will acquire the minimum amount of storage for control blocks
  (PDA/PDP) to open the data set. This will significantly reduce
  the storage needed to build all the necessary control blocks
  (DBRT entry, PDA, PDP), hence avoid the global out of storage
  condition during DB2 restart.
  

Problem conclusion

• DB2 restart logic is modified to reduce the storage usage to
  avoid global storage shortage which prevents DB2 from
  restarting.
  

Temporary fix

• *********
  * HIPER *
  *********
  

Comments

APAR Information

• APAR is sysrouted FROM one or more of the following:

• APAR is sysrouted TO one or more of the following:

  UK77963 UK77964

Modules/Macros

• DSNB1OPP DSNIERST DSNIRCSB
  

Fix information

• Fixed component name

  DB2 OS/390 & Z/

• Fixed component ID

  5740XYR00

Applicable component levels

• RA10 PSY UK77963

     UP12/05/01 P F204

• R910 PSY UK77964

     UP12/05/01 P F204

Fix is available

• Select the PTF appropriate for your component level. You will be required to sign in. Distribution on physical media is not available in all countries.