I'd start with running sybdiag + collecting all possible metrics on the SPID (all MDAs that have any reference to SPID), collect multiple stacktraces (dbcc stacktrace(spid)) and send all this stuff to SAP support. Dbcc stacktrace may show you which function the code executes at the moment - if you run it several times you may see if the process is in a hung state or loping on anything. You might also attach to ASE's kernel memory and fish out information on the culprit spid from there - again dispatching it to support for investigation. But that's for support guys to advise you... You might chance using dbcc dbreboot instead of full ASE recycle - but I don't trust dbreboot 100%...
↧
