Applies to:Siebel System Software - Version: 8.1  and later [Release: V8 and later ]
Information in this document applies to any platform.
Error Message Area:Server Request Broker - SRB
- SRBoker is still running.
- Batch mode components are still active.
- Try restarting the component to re-establish the connection.
Applies to:Product Release: V7 (Enterprise)
Version: 7.8.2  Pub Sect
Database: IBM DB2 for zOS and OS/390 v8
Application Server OS: IBM AIX 5L 5.2
Database Server OS: IBM zOS
This document was previously published as Siebel SR 38-3307425821.
SBL-SRB-00051: no connection for the component
I have attached the logs for you info.
Numerous hangs were encountered when doing performance/stress testing.
This only occured when the webserver was using SSL.
Using the "pstack.sh" script(available Troubleshooting Steps 30) against the hung process
it could be seen that a thread was actually crashing, but had caused a hang.
Following stack was seen:
in pth_spinlock._global_lock_common [/usr/lib/libpthread.a] at 0xd011029c ($t94)
0xd011029c (_global_lock_common+0x35c) 80410014 lwz r2,0x14(r1)
pth_spinlock._global_lock_common(??, ??, ??) at 0xd011029c
malloc_common.__malloc_prefork_lock_common() at 0xd02f7df0
malloc_common.__malloc_prefork_lock() at 0xd02f7208
fork.f_fork() at 0xd033ec90
mwsystem__FPCc() at 0xd934a940
system_append__FPcP4FILET1() at 0xd934b184
MwPrintProcOSInfo__Fi() at 0xd934afc4
MwPrintCurProcOsInfo__Fv() at 0xd934af0c
MwPrintProcInfo__FiPc() at 0xd934ab64
MwAbort() at 0xd928ae98
This turned out to be the behaviour described in "Alert 1198: Object Manager Hang Behaviors on UNIX Platforms Due to Calls to Malloc Memory Management Function in the Crash Handler Code "
The actual "fork" statment is generating the malloc, CR #10642640 was raised to track this.
When setting the MWNO_SIGNAL_CATCHING=TRUE, the processes started to crash.
The callstacks were random , but always ended in a memory management statement such as malloc.
This was reproduced in Oracle Technical Support and Change Request 10521329
was logged to address this product defect.