ANBOB™

Oracle 19.1 or 19.4  on AIX ,There are cases where this error has occurred, An instance terminated because a DBWR raised ORA-600[kjcsmpav:1] during  a buffer allocation.

ORA-600: internal error code, arguments: [kjcsmpav:1], [0x7000xxxxxxx], [4], [0], [211], [220], [], [], [], [], [],

kjcsmpav ==>  KJC message pool free list  allocate a msg buffer,which is a RAC ralated latch dealing with the Kerel Job Control message pool(Global Enqueue Services)，

It tried to allocate a message from the message pool mp. Before it sets the OPID of the process 220, to the KJCSMP_OPID(mp) to exclude
the other processes, someone else, the process 211 already had the message pool mp.

Call stack

dbgexPopulateIncCtx <-dbgexProcessError <-dbgeExecuteForError <-
dbgePostErrorKGE <-dbkePostKGE_kgsf
<-kgeadse <-kgerinv_internal <-kgerinv <-kgeasnmierr <-kjcsmpav
<-kjcssba_prealloc <-kjccgmb_prealloc_wLoc <-kjbenterdlmnw <-KCLENTERNAME_NW <-kclenterblk_nw
<-kclenterb_nw <-kcbbic1_int <-kcbbic1 <-kcbbiop <-kcbbdrv
<-ksb_act_run_int <-ksb_act_run <-ksbabs <-ksbrdp <-opirip
<-opidrv <-sou2o <-opimai_real <-ssthrdmain <-main
<-start

kgeadse                    kernel generic error KGE ADd String Internal Error onto the error stack  [partial hit for: kgeadse ] 
kgerinv_internal           kernel generic error record internal named error with va_list or soft internal named error  [partial hit for: kgerinv_internal ] 
kgerinv                    kernel generic error record internal named error with va_list  [partial hit for: kgerinv ] 
kgeasnmierr                kernel generic error assert named internal error  [partial hit for: kgeasnmierr ] 
kjcsmpav                   kernel lock management communication  [partial hit for: kjc ] 
kjcssba_prealloc           kernel lock management communication  [partial hit for: kjc ] 
kjccgmb_prealloc_wLoc      kernel lock management communication  [partial hit for: kjc ] 
kjbenterdlmnw              kernel lock management global cache service  [partial hit for: kjb ] 
kclenterNAME_NW            kernel cache lock manager/buffer cache  [partial hit for: kcl ] 
kclenterblk_nw             kernel cache lock manager/buffer cache  [partial hit for: kcl ] 
kclenterb_nw               kernel cache lock manager/buffer cache  [partial hit for: kcl ] 
kcbbic1_int                kernel cache buffers databasewriter  [partial hit for: kcbb ] 
kcbbic1                    kernel cache buffers databasewriter  [partial hit for: kcbb ] 
kcbbiop                    kernel cache buffers databasewriter  [partial hit for: kcbb ] 
kcbbdrv                    kernel cache buffers databasewriter central write driver  [partial hit for: kcbbdrv ] 
ksb_act_run_int            kernel service  background processes  [partial hit for: ksb ] 
ksb_act_run                kernel service  background processes  [partial hit for: ksb ] 
ksbabs                     kernel service  background processes action based server  [partial hit for: ksbabs ] 
ksbrdp                     kernel service  background processes run a detached background process  [partial hit for: ksbrdp ] 

-- define in rdbms/src/server/ccl/dlm/kjcs.c
 /* get message pool allocation latch */
 /* get from the kgh pool if needed and allowed */
if (allocm < total_alloc && /* mb still needed? */
       KJCSMP_MFLT(mp)) /* Free lists allocated (ie not reserve pool) */
 alloc_new = TRUE;
 KJASSERTF5(KJCSMP_OPID(mp) == 0,
     OERINM("kjcsmpav:1"),

This should be a logic problem of the program code. At present, the final bug number has not been confirmed. Only Bug 30486436 records the problem.

打赏

微信扫一扫，打赏作者吧～