I've been using SA since the Watcom days and never looked back. But once in a while I'll upgrade to a new EBF (for a reason), run some tests, so far so good, will then push the upgrade to the production server only to find out that without any apparent reason the database server crashes silently on one halfway complex query or stored procedure that's been working without a glitch for the past couple of years or longer... This is actually a question, not gratuitous lamentations. I'd like to know what the recommended practice is for troubleshooting what's causing the crash. There are times when rolling back the EBF is not really an option. Here's what I've done:
Now, unless I'm not reading it correctly, the log file does not provide any information on why the server crashes. All I can see is a trace of various blocks in a stored procedure being executed, then nothing more when the proc should be doing a Short of trying to make sense of the dump file which Sybase Engineering are getting, is there anything else I can do to try and isolate what causes the crash? I understand this is situation I could report to Sybase directly, if I had a support plan that is ( not an option for budgetary reasons ). I'm reporting the matter using Case Express. Is there anything else I can do to:
Edit:To follow up on Thomas' comment below, here's a copy of the proc: |
Cross-posted from sqlanywhere-forum.sap.com. Danke Volker für den Hinweis.
If you can show some source code perhaps it is possible to rewrite the select. Is it possible to get the plan of the Query before and after the EBF ?
We'll try to diagnose the crash as quickly as we can with the dbsupport submission. If you can provide the dbsupport submission numbers here (dbsupport -ls), we might be able to look into the issue a bit further for you.
If you wanted to track the issue through engineering, you should open a technical support case with us, so you that can be notified regarding the resolution and request an EBF with this fix: http://www.sybase.com/contactus/support/
For now, downgrading to 11.0.1.2550 would be your only real option. Were you just testing with 2558 or were you looking for a specific fix?