2511621 - Signal 11 in buf__unhash_and_keep_grabbed_mass() or bufnewpage() - SAP ASE

• ASE reports an "Infected with signal 11" error in the module buf__unhash_and_keep_grabbed_mass() or bufnewpage():

Current process (0x2cd10229) infected with signal 11 (SIGSEGV)
Address 0x0x0000000000d88bef (buf__unhash_and_keep_grabbed_mass+0x54f)
 
or
 
Current process (0x3fd01ff) infected with signal 11 (SIGSEGV)
Address 0x0x0000000000d8eb4d (bufnewpage+0xd4d)

• Issue is reported with different stack traces. Patterns reported show functions including:

buf__unhash_and_keep_grabbed_mass
bufnewpage
pg__getpgbuf
pg__ea_initpg
pg__ea_nextpg
pg__allocate

or

buf__unhash_and_keep_grabbed_mass
bufgrab
getpage_with_validation
pg_get_alloc_page
pg_logdevusage
pg_getpagecounts
ptngetpages
LeRun

  Note: Two examples of complete stack traces are available under Attachments.

• ASE server goes down because the terminated process (spid) was holding cache spinlock on a (named) cache:

Spinlocks held by  kpid 12345678
Spinlock tempdb_cache at address ....

or

Spinlock default data cache at address ...

SAP Adaptive Server Enterprise (ASE) 15.7

instext, CR811140, CR#811140, 811140, session, descriptor, buffer, memory, corruption, SEGV, segmentation, storage, access, violation, fault, segfault, kisignal, stacktrace,  registers , KBA , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Bug Filed