RAC Database Hang Due To Row Cache Objects Latch Contention While LCK0 Waiting For Row Cache Cleanup (Doc ID 2491737.1)

Last updated on MARCH 18, 2022

Applies to:

Symptoms

• RAC database hung on latch: row cache objects from 11.2.0.2
• Huge waits for latch: row cache objects and latch: shared pool by the processes
• Not able to login to RAC instance by user processes
• LCK0 waits on "row cache cleanup" for long time
• AWR snapshot not generated during the problem period
• Call stack shows similar to:
  LCK0:
  ksedsts <- ksdxfstk <- ksdxcb <- sspuser <- __sighndlr
  <- call_user_handler <- sigacthandler <- _syscall6 <- sskgpwwait <- skgpwwait
  <- kslges <- kghfrunp <- kghfnd <- kghalo <- kghgex
  <- kghfnd <- kghalo <- kghgex <- kghfnd <- kghprmalo
  <- kghalp <- parchk <- pkaidl <- pkamh2 <- pdybF00_Init
  <- pdy1F01_Driver <- pdw0F82_Run_Code_Gen <- pdw0F01_Code_Gen <- phpcmp <- pcicog

  Foreground:
  ksedsts <- ksdxfstk <- ksdxcb <- sspuser <- __sighndlr
  <- call_user_handler <- sigacthandler <- _syscall6 <- sskgpwwait <- skgpwwait
  <- kslges <- kqrbip <- kqrbfr <- kqrbgl <- ksbabs
  <- ksbrdp <- opirip <- opidrv <- sou2o <- opimai_real
  <- ssthrdmain <- main

    
  
  

Changes

 RAC Database 11.2.0.2 and above.

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.