3662868 - Signal 11 in buffer of HK function followed by ASE crash due to spinlock held on a data cache - SAP ASE

• An ASE server reports a signal 11 in a buffer or HouseKeeper (HK) related functions for example
  • hk__wash_mass
  • bufsearch_cache_getlatch
  • bufdunlink
    
    
• ASE crashes and shuts down due to the spid holding the 'default data cache' spinlock or a named cache spinlock which is likely to be a tempdb cache
• Sometimes the issue starts with an error 696
  
  
• The stack function where this is hit and further stack patterns can vary but these are typical, shortened examples of ASE errorlog sequences:
  
  1.
  Current process (0x850043) infected with signal 11 (SIGSEGV)
  Address 0x0x0000000001b1be68 (hk__wash_mass+0xf8), siginfo (code, address) = (128, 0x(nil))

  Spinlocks held by  kpid 12345678
  Spinlock default data cache at address 0x00002aaac477d900 owned by 12345678

  -- relevant stack functions
  hk__wash_mass
  hk__pool_wash
  hk__wash_main

  
  2.
  Current process (0x4c25072e) infected with signal 11 (SIGSEGV)
  Address 0x0x0000000001afe667 (bufsearch_cache_getlatch+0x497), siginfo (code, address) = (1, 0x0x0000000000000018)

  Spinlocks held by kpid 1277495086
  Spinlock tempdb_cache1 at address 0x00002aaac0ba0480 owned by 1277495086

  bufsearch_cache_getlatch
  getpage_with_validation
  bt__latchrootpage

SAP Adaptive Server Enterprise (ASE) 16.0 & 16.1

spinlock, CR 831281, 831281, house keeper, LRUUNLINK , cm_mass_destroystate , buf_check_and_destroy , bufpagedestroy , bwfinish , sdfinish_normal , sddone_normal , blkioComplete ,
kdcProcessCompletions , udcPoll , Run to completion task iocRunController(47c788b0) exception 11 , KBA , BC-SYB-ASE , Sybase ASE Database Platform (non Business Suite) , Known Error