Server Tables Cleanup component failing - deadlock trying to delete records from s_srm_request (Doc ID 496314.1)

Last updated on SEPTEMBER 08, 2023

Applies to:

Symptoms

On the morning of 10/3 around 7:30 am est, one of our production Object manager servers failed. Research found that the 'Server Tables Cleanup' component on one of the object manager servers was failing when trying to delete a record from the S_SRM_REQUEST table. The failure message states that the record that is being deleted has been updated by another user. The process continued to try to delete the record and failed for many hours creating a 25 gig log file.

The server became unavailable and had to be rebooted. Stopping and starting the Siebel server did not work prior to the reboot. Following the reboot, the server came back up and the 'Server Tables Cleanup' process ran successfully several times for the first 1/2 hour the server was up. After that, the same error started occuring again.

Our production environment has 4 siebel servers. Each of them is configured to run the component "Server Tables Cleanup". This has been our configuration since upgrading to Siebel 7.7 last December. It seems like the server tables cleanup process started having conflicts for the first time yesterday 10/3. Is it possible that different servers are trying to delete the same record?

Should we only have "Server Tables Cleanup" enabled on 1 of the 4 servers to prevent the deadlock condition? If so, I'm not sure why this error just started yesterday and was not present since we went live with our current architecture?

Changes

Cause

My Oracle Support provides customers with access to over a million knowledge articles and a vibrant support community of peers and Oracle experts.