|
Hey everybody, Sybase 10 32 Bitis running as a Service on Server 2008x64. Last 4 weeks the server crashes 2 times with following error with not entry in logfile Error in application dbsrv10.exe, version 10.0.1.3415, time stamp 0x46094b94, faulty module dbserv10.dll, version 10.0.1.3415, time stamp 0x46096fa0, exception code 0xc0000005, fault offset 0x002f3be5, process id 0x1394, application start time 01cfd8d5a795cc57. Is it possible to activate an extended log to see for the next time? I have heard from the verbose log, but solves that my problem? |
Error in application dbsrv10.exe, version 10.0.1.3415, time stamp 0x46094b94, faulty module dbserv10.dll, version 10.0.1.3415, time stamp 0x46096fa0, exception code 0xc0000005, fault offset 0x002f3be5, process id 0x1394, application start time 01cfd8d5a795cc57. That's a fatal internal server crash; the "exception code 0xc0000005" gives it away. That is a C0000005 access violation... a bug inside SQL Anywhere that should not happen, and certainly should not be caused by anything you do... and no amount of bug hunting on your part is going to find the cause. SQL Anywhere 10.0.1.3415 is pretty old. 10.0.1.4310 exists; it may or may not be the latest EBF, and it may or may not fix the problem, but it is worth trying a recent EBF. However, it may not be easy to obtain that EBF since "the gates have been closed" on downloading recently. Also, Version 10 Technical support is still available, however, so if you can get a crash dump you might get someone to look at it; i.e., fixes might not be available, but diagnostic services are. They may suggest a workaround. Even better: Upgrade to Version 12 or 16. 1
Well, the user certainly is not responsible for that crash but I remember (rare) situations when I could reproduce a db server crash by re-doing a particular SQL statement, so a user still might cause such a crash - and might be able to prevent that particular command... That being said, I certainly agree that there's not much sense in bug hunting here - it just may be that the bug finds you:)
(08 Oct '14, 09:12)
Volker Barth
|
What do you mean with logfile? Do you already use a console log file, such as via "dbsrv10 -o MyConsole.log"? (If not, that would be highly recommended...)
Yes, sorry i mean the console log. There is no entry at the crashtime. So at the moment i do some tests with
CALL sa_server_option( 'DebuggingInformation','YES' );
I will see if this show me more informations about the service when go to taskmanager an kill the service (in a testmaschine)...
FWIW, that option (aka dbsrv10 -z) will usually have an impact on performance so it may not be used in general for production systems... - OTOH, it's not that easy to set that option immediately before a crash - there's a lack of a "about to crash" system event:)