-
标题
"show post" or "show post detail" commands fail with semaphore error. -
说明
"show post" or "show post detail" commands fail with semaphore error.
sp_ctrl (dc2db03-pri:2550)> show post
cnc_post_shs_status() Error in sp_cnc
rpc_post_shs_status() internal error
cnc_post_shs_status() Error in sp_cnc
rpc_post_shs_status() internal error
No process found for poster service or invalid datasrc-datadst
or invalid quename
The following error messages appear in the event_log:
03/05/08 15:39 Notice: Result (28) No space left on device calling semop to get semaphore (3008) [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Granting request would exceed system limit on number of processes waiting on a semaphore [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Result (28) No space left on device calling semop to get semaphore (1008) [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Granting request would exceed system limit on number of processes waiting on a semaphore [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: shs_lock: Error getting global semaphore [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Result (28) No space left on device calling semop to get semaphore (3008) [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Granting request would exceed system limit on number of processes waiting on a semaphore [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Result (28) No space left on device calling semop to get semaphore (1008) [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: Granting request would exceed system limit on number of processes waiting on a semaphore [sp_cnc(shs)/10956]
03/05/08 15:39 Notice: shs_lock: Error getting global semaphore [sp_cnc(shs)/10956] -
原因
The system ran out of semaphores.
In this particular case the OS was Solaris and the resolution was to set semmnu and semume to 3000 and bounce the server. -
解决办法
Edit your kernel settings to allow more semaphores for the user who runs shareplex. You will need to be root to do this. You system administrator can help to set this up.