Dear Sir / Madam,
We recently encounter a occasion, that many users in SAP system seems hang up.
During analysis (From ST04), we soon find out that during the time, many UPDATE process (accross various server) encounter DB level blocking.
However, the overall recovery time is still rather long (near 1 hour), as we are trying to use least disruptive way to kill that process.
(i.e. the update process holding the DB connection, which being at the head of chain of DB connection encountering blocking)
We have tried in this sequence :
1) Contact user by phone, seeking them to kill his own by log out (Not work, user only sign out, update process still there)
2) On the server where the concerned UPDATE process exist, use SM04 to log off the concerned user (Not work)
3) On the server where the concerned UPDATE process exist, use SM50, cancelled the update process "without core" (That Works)
After kill that UPDATE process, from SM13, we see the concern update shown RED Alert [Error (no retry)],
but the hang up situation did resolved.
Retrospectively there seems additional options
4) On DB engine side, KILL the DB connection session with Database tools
(which I think may cause similar effect, just like DB Cluster switching occurs during SAP program (with update) running )
5) Kill the update from SM13 (Is it at all possible ?)
So what would be the preferred way for Netweaver Administrator, when trying to do this recovery ?
(i.e. cause least effort for functional team to recover application level problem)
Any advice would be much appreciated.
Regards,
Eric
FYI.
We are using Netweaver 7.01, SQL Server 2008, Windows 2008