We have an agent that transfers data between other databases and Oracle. It is scheduled to run twice a day. It sits on a server in a 2 server cluster which is mostly used for mail routing.Sometimes the agent runs longer than the server doc allows. And then trouble starts because on the next run there will be even more data to process.
Analyzing the log I noticed that is last nights 3 hour run there were three gaps of no processing.
The agent started at 22:30. One gap was between 12 PM and 01 AM; the housekeeping probably. One between 01 AM and 01:30 AM; more housekeeping.
But there was also a 50 min gap between 22:40 and 11:30 !!! Al I can see in the log is replication of this DB (only 300 records, 5MB) to the other server in the cluster. Nothing else.
Does anyone have an idea why this happens ?
Is there some kind of ‘put agent on hold’ mechanism on cluster machines ?
Thanks,
Peter