This started the other day and it was first only on xogging, then it started to affect the UI (Creating & saving projects). It is an error on the audit trail trigger (I think).
I opened a case with CA, but with the impact of this issue, I am looking for any feedback I can get.
Original error when I opened the case yesterday:
<Description>Project Object insert failed</Description>
ORA-06512: at "SC51339T.ODF_AUD_5001168_FCT", line 5
ORA-06512: at "SC51339T.T_PROJECT_OODF_CA_PROJECT", line 289
ORA-04088: error during execution of trigger 'SC51339T.T_PROJECT_OODF_CA_PROJECT'
ERROR 2016-06-07 10:31:20,352 [http-nio-14001-exec-22] niku.xql2 (clarity:admin:25559636__D8A5EE38-2DBB-40F7-A461-B3C17920ED13:projmgr.projectNewPost) Internal Processing exception
SQL error code: 4098
Error message: [CA Clarity][Oracle JDBC Driver][Oracle]ORA-04098: trigger 'SC51339T.T_PROJECT_OODF_CA_PROJECT' is invalid and failed re-validation
I turned off auditing on ALL attributes from project object and all subobjects.
Turned on auditing on one attribute on project object
Removed auditing on one attribute on project object
Nothing has helped. Support recommended an app server restart which I am waiting on. Anyone seen this before, that has more feedback then the KB article on the error which advised to turn off auditing.
Removing auditing on the project object attributes should normally remove the audit triggers. If this didn't happen, then we may need to look on database side, and probably deactivate and drop the trigger manually. Then you should be able to add your audited attributes back, and this will readd a validate trigger.
Please let us know how it goes.
Kind Regards -Nika
Thanks Nika. Supported confirmed that should fix it also, but we had the DBA try to restore the trigger from previous versions etc, but that did not resolve it.
We ended up solving it internally.
From the error in the app log, we found a lookup that was being referenced. When we queried the DB with the lookup, it was returning two rows (one was expected).
We xogged out the instance and deleted the instance record that should not have been there We then rexogged the same instance and PPM assigned it a new 5million ID. That solved the issue.