No compare server is running and compare client exists on target with event log message “Error, lock file not found” and “Error: compare server is not responding; pipe broken” as the following event log entries on target show:
sp_ctrl (server_name:2100)> show log reverse
Info 2017-04-11 13:45:48.563182 20373 1058592608 Compare client exited normally, pid = 9675 [sp_declt]
Error 2017-04-11 13:45:38.560123 9675 1556535136 Compare client: Error, lock file not found see (pid=9675)(main,L1402) [sp_declt] [module deq]
Error 2017-04-11 13:45:38.558513 9675 1556535136 Compare client: Error: compare server is not responding; pipe broken. (sp_declt_wait_for_lockfile,L1035) [sp_declt] [module deq]
Notice 2017-04-11 13:34:36.122026 9675 1556535136 Compare client: Opened DataEquator client session log file /u01/splex/vardir2100/log/declt_130-1_BOSSPRD_10.86.40.42_p9675.log [sp_declt] [module deq]
Info 2017-04-11 13:34:36.032268 9675 1058592608 Compare client launched, pid = 9675 [sp_declt]
The “status” on target show that two Compare clients are running though there is no corresponding Compare server seen on source:
sp_ctrl (server_name:2100)> status
Brief Status for server_name
Process State PID Running Since
--------------- ------------------------------ -------- --------------------
Cop Running 20373 10-Apr-17 12:43:27
Import Running 20407 10-Apr-17 12:43:33
Post Running 20476 10-Apr-17 12:43:49
Compare Clnt Running 31398 11-Apr-17 08:49:46
Compare Clnt Running 23630 11-Apr-17 10:44:56
The Post queue is not draining anymore as Post is not processing any transactions right now.
Carry out the following steps on target:
1. Shutdown SharePlex and if it does not shutdown normally use the shutdown force option as needed.
2. If there is any orphaned compare client process or Post process showing up when the “ps –ef|grep sp_” is issued after shutdown, kill it manually at OS level by “kill -9 <PID>”.
3. Restart SharePlex on target and verify that no orphaned compare process re-spawns by issuing “status” at target sp_ctrl.
Such orphaned compare clients can remain due to several reasons, among others are some of the following:
A very large table was being compared but subsequently cancelled by “cancel <jobid>” which resulted in the compare server exiting gracefully but the compare client lingered.
Multiple compare were issued for the same table.
