Reader core dump due to SIGBUS.
"Error: process sp_ordr killed due to SIGBUS [sp_cop/9467]
Notice: process sp_ordr killed due to SIGBUS; (core file = /usr/ssw/splex1/splex/var/
Process launched: sp_ordr (for o.CCON) [pid = 11635]
Notice: Oracle env - CCON:/usr/ssw/app/oracle/product/9207 [sp_ordr(pdb)/11635]"
Capture queue build up 14G messages
Name: o.CCON (Capture queue)
Number of messages: 33732560 (Age 1103 min; Size 13765 mb)
Backlog (messages): 28044153 (Age 1053 min)
It was determined that there are resource limitations in the system, hence it is not able to start and maintain reader activities in loading all the cache files to memory.
Due to large transaction size, it was suggested that the amount of memory available for process heap be increased from current value of 1GB to 4GB to handle the current situation. This needs kernel parameter change, and a reboot.
1. Shutdown Shareplex.
2. Increase hard limit and soft limit for ulimit data size.
ulimit -aH
shows 1GB allocated for data(kbytes), it needs to be changed to 4GB. This is the max vallue for a 32-bit OS.
Reboot server after making change.
3. Restart Shareplex.
ulimit -aH
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 1048576
stack(kbytes) 262144
memory(kbytes) unlimited
coredump(blocks) unlimited
ulimit -aS
time(seconds) unlimited
file(blocks) unlimited
data(kbytes) 1048576
stack(kbytes) 262144
memory(kbytes) unlimited
coredump(blocks) 4194303
