IT32607: CONCURRENT STARTING AND EXITING OF DBWORKER THREADS COULD LEAD TO INCORRECT CE_USECOUNT/REF_CNT

APAR status

• Closed as program error.

Error description

• There is a timing problem when starting and stopping dbWorker
  threads concurrently across cpu vps/processors which can lead to
  the ref_cnt value for the db_sch_worker function (from onstat -g
  pr) to become incorrect.  If this value does get out of sync, it
  might be possible for the cache entry to be removed while
  threads may still be referencing it.
  
  If that were to happen it could lead to assertion failures.  One
  possible stack could be:
  
  13:10:11  Stack for thread: 26375823 dbWorker3
  
  0x000000010009b9dc (oninit)affail_interface
  0x000000010009f314 (oninit)mt_ex_throw_sig@AF11_4
  0x0000000100061f54 (oninit)afsig_handler
  0x0000000100a35d80 (oninit)db_sch_worker
  0x0000000100e917d8 (oninit)udrlm_clang_execute_internal
  0x0000000100e91518 (oninit)udrlm_clang_execute
  0x00000001015dcd5c (oninit)udrlm_exec_routine
  0x00000001003e2920 (oninit)udr_execute
  0x0000000100a44f8c (oninit)dbsched_start_udr
  0x0000000100b14f44 (oninit)th_init_initgls
  0x000000010161f7fc (oninit)startup
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * Users of Informix 11.70.xC9 and earlier versions.            *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * There is a timing problem when starting and stopping         *
  * dbWorker                                                     *
  * threads concurrently across cpu vps/processors which can     *
  * lead to                                                      *
  * the ref_cnt value for the db_sch_worker function (from       *
  * onstat -g                                                    *
  * pr) to become incorrect.  If this value does get out of      *
  * sync, it                                                     *
  * might be possible for the cache entry to be removed while    *
  * threads may still be referencing it.                         *
  *                                                              *
  * If that were to happen it could lead to assertion failures.  *
  * One                                                          *
  * possible stack could be:                                     *
  *                                                              *
  * 13:10:11  Stack for thread: 26375823 dbWorker3               *
  *                                                              *
  * (oninit)affail_interface                                     *
  * (oninit)mt_ex_throw_sig@AF11_4                               *
  * (oninit)afsig_handler                                        *
  * (oninit)db_sch_worker                                        *
  * (oninit)udrlm_clang_execute_internal                         *
  * (oninit)udrlm_clang_execute                                  *
  * (oninit)udrlm_exec_routine                                   *
  * (oninit)udr_execute                                          *
  * (oninit)dbsched_start_udr                                    *
  * (oninit)th_init_initgls                                      *
  * (oninit)startup                                              *
  ****************************************************************
  * RECOMMENDATION:                                              *
  ****************************************************************
  

Problem conclusion

• Upgrade to Informix 11.70.xC9W2.
  

Temporary fix

Comments

APAR Information

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

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

Fix information

• Fixed component name

  INFORMIX SERVER

• Fixed component ID

  5725A3900

Applicable component levels