We have a bpel process that runs at night. The process consumes message from a queue and then writes to a file. I can see that there are some junk instances being generated even after the process is completed. I mean there is no new instance being created but in the dashboard of em console a lot of instances are failing.
Am unable to open the instance as as it says: Unable to open flowtrace due to exception while initializing object
I had shut down the interface but could still see those faulted instances.
I had run a query to check on the count of no:of such junk instances and could see that there are around 7200+ junk instances that are non-recoverable being generated everyday.
Any idea how to identify what the issue is and how to prevent junk instances being generated.
Is your SOA server clustered? If yes, have you applied singleton property on your process? Which SOA version you're using?
Also can you test your process on-demand (instead of nightly run) so that you can collect some diagnostic logs and paste it here?