IC67076: SQLEXEC THREAD WILL CRASH SERVER AFTER LOCK TABLE OVERFLOWS 100 TIMES

APAR status

• Closed as program error.

Error description

• The following conditions will crash the server:
  1. Lock table overflows 100 times, can not allocate more locks;
  2. "queuedjob" latch is held by someone else, so the current
  session will run into wait queue.
  
  19:06:32
  
  19:06:32  IBM Informix Dynamic Server Version 11.50.FC6W1X2
  Software
  Serial Number AAA#B000000
  
  19:06:32  Assert Failed: yield_processor: Conditional latch
  count non-
  zero in thread 144028
  
  
  
  19:06:32   Who: Session(140019, jfdba@icbcoa-126b3322, 3868,
  
  7000001d90a29e0)
  
    Thread(144028, sqlexec, 7000001d137dbb8, 4)
  
    File: mt.c Line: 16603
  
  19:06:32  Raw hex dump of stack located in
  /ifx_dump/af.36849fb7.rawstk
  19:06:32  Stack for thread: 144028 sqlexec
  
  
  
   base: 0x07000001dc83e000
  
    len:   528384
  
     pc: 0x00000001000dbf34
  
    tos: 0x07000001dc8bb7d0
  
  state: ready
  
     vp: 4
  
  
  
  0x00000001000dc168 (oninit)afstack
  
  0x00000001000de27c (oninit)afhandler
  
  0x000000010005b5c4 (oninit)yield_processor_mvp
  
  0x000000010005e248 (oninit)mt_lock_wait
  
  0x000000010006adf8 (oninit)mt_lock
  
  0x000000010026a5e8 (oninit)ph_submit_task
  
  0x000000010026931c (oninit)ph_alarm_post
  
  0x0000000100100454 (oninit)invoke_alarmf
  
  0x000000010034fde8 (oninit)lkgrow
  
  0x00000001003577bc (oninit)lkalloc
  
  0x0000000100355234 (oninit)lkmanagex
  
  0x000000010035fe74 (oninit)rsdbopen
  
  0x00000001004818e8 (oninit)isdbopen
  
  0x0000000100240efc (oninit)sqisdbopen
  
  0x00000001004e1a30 (oninit)ddopen
  
  0x00000001008fef8c (oninit)dbopen
  
  0x00000001008ff27c (oninit)find_db
  
  0x00000001008fff5c (oninit)sqopendb
  
  0x000000010042e9f0 (oninit)sqdbopen
  
  0x00000001004480bc (oninit)sq_dbopen
  
  0x0000000100267474 (oninit)sqmain
  
  0x00000001003b2de8 (oninit)listen_verify
  
  0x00000001003b12b8 (oninit)spawn_thread
  
  0x0000000100d51b0c (oninit)startup
  
  This problem can be found in:
  AIX 5.3 IDS 11.50.FC6W1X2
  HP 11.11 IDS 11.50.FC6W2
  

Local fix

Problem summary

• ****************************************************************
  * USERS AFFECTED:                                              *
  * All Users                                                    *
  ****************************************************************
  * PROBLEM DESCRIPTION:                                         *
  * If IDS runs out of locks , it dynamically allocate more      *
  * locks until a max limit is reached. And when we hit the max  *
  * limit there is possibility that IDS server would crash with  *
  * following error in online.log                                *
  * 19:06:32  Assert Failed: yield_processor: Conditional latch  *
  * count non-zero in thread 144028                              *
  ****************************************************************
  * RECOMMENDATION:                                              *
  * Please upgrade to IDS 11.50.xC7 and above.                   *
  ****************************************************************
  

Problem conclusion

• Problem is first fixed in IDS 11.50.xC7
  

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

  IBM IDS ENTRP E

• Fixed component ID

  5724L2304

Applicable component levels

• RB15 PSY

     UP