Open nikmagini opened 8 years ago
This alert resolved itself automatically after 16 hours; probably the transfer task that was blocking the deletion expired.
Already seen in testbed in 2011 BTW:
https://github.com/dmwm/PHEDEX/issues/651#issuecomment-26934517
On Nov 14th the central FilePump agent started alerting this:
2016-11-14 16:37:55: FilePump[14613]: alert: database error: DBD::Oracle::st execute failed: ORA-02292: integrity constraint (CMS_TRANSFERMGMT.FK_XFER_TASK_REPLICA) violated - child record found (DBD ERROR: OCIStmtExecute) [for Statement "delete from t_xfer_replica where (node, fileid) in (select xr.node, xr.fileid from t_xfer_delete xd join t_xfer_replica xr on xr.node = xd.node and xr.fileid = xd.fileid where xd.time_complete is not null and xd.time_complete >= xr.time_create)"] at /data/ProdNodes/PHEDEX/perl_lib/PHEDEX/Core/DB.pm line 322
It won't stop transfers from progressing, but it stops the registration of completed deletions and the statistics updates.
Looks like some deletion was executed while there was some other task (transfer? verify?) of the same replica in progress.
Needs further investigation
N.