2793139 - Random proc cache corruption leading to many signal at kboalloc - SAP ASE

• The ASE errorlog reports (very) many signal 11 messages at address kboalloc
• Which can render ASE unusable for user applications
• Or results in ASE crash if affected spid is holding a spinlock (rgheapblock_spin)
  
  
• Prior to the signal 11 at kboalloc
• other (random) signal 11s can be seen
• and/ or errors is the 700 range, for example 706, 707 or 719
• There is no fixed pattern, but these are examples of kind of messages seen in ASE log:

Current process (0x106903ad) infected with signal 11 (SIGSEGV)
Address 0x0x0000000000863570 (LeSQLRep::LeSQLRep(ExeCtxt&) ...

... no stack trace printed

**** end of signal context ****
task 275317677 is zombie

Error: 719, Severity: 20, State: 1
The last prochdr value is not available for the prochdr '0x0x2aac3c59e000'. This is an Adaptive Server internal error. Please contact Sybase Tech Support.

Error: 707, Severity: 20, State: 5
System error detected during attempt to free memory at address 0x0. Please consult the ASE error log for more details.

• Followed by signal in memory allocation routine kboalloc:

Current process (0x59002d) infected with signal 11 (SIGSEGV)
Address 0x0x000000000164b09f (kboalloc+0x2ff) ...

• The signal 11 at kboalloc repeats itself many times and may end with ASE shutting down with a signal 11 for a spid holding spinlock:

Current process (0x1101008d) infected with signal 11 (SIGSEGV)
Address 0x0x000000000163e1b8 (kbgethashent+0x38) ...
...
Spinlocks held by kpid 285278349
Spinlock Resource->rgheapblock_spin at address ...

SAP Adaptive Server Enterprise (ASE) 15.7 SP140

kboalloc , KBA , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Problem