When compare is issued at time post is stopped due to error then the compare messages are stuck in the queue.The compare clients time-out and exit. When post is restarted , it will wait for each ODR_DEQ (compare) message for 300 sec (5 minutes) before it removes the sp_de_lock file from the state directory and proceed.
So if there are a lot of compare messages it may appear that post is hung or slow.
Event logs shows this message :
38 Notice 2011-05-24 15:47:00.488251 19353 1 Poster: create lock file=/splex/kwong/splex7/7.5.2-PremierFarnell/run/var100/state/sp_de_lock-KCWTAB2-18768-6_ processDataEquatorMsg_create_75plus_semaphore_file,L3725) (posting from ora102, queue irvspxu15, to ora102) [module opo]
39 Notice 2011-05-24 15:47:02.497197 19353 1 Poster: wait_count=2 repairWaitCnt=300 (processDataEquatorMsg,L938) (posting from ora102, queue irvspxu15, to ora102) [module opo]
40 Notice 2011-05-24 15:47:04.507217 19353 1 Poster: wait_count=4 repairWaitCnt=300 (processDataEquatorMsg,L938) (posting from ora102, queue irvspxu15, to ora102) [module opo]
.
General functionality
Workaround:
Resolution 1
Wait for the all the compare messages to process.
Resolution 2
Check the SP_SYS_VARDIR/state directory and remove the lock file (sp_de_lock file )
Resolution 3
Read the post queue and discard the compare messages.
Lock file info:
go to SP_SYS_VARDIR/state looks for sp_de_lock-KCWTAB2-18768-6_ processDataEquatorMsg_create_75plus_semaphore_file,L3725
