I´m running a SA188.8.131.5296 with HA in a Windows 2008R2x64.
My problem is the database server is shutting down with no indication of why. I´ve got just an "Internal Server Error". There is nothing on the console log file and the request logging file does not show nothing that explains the crash.
I´ve looked to the crash log file and the only readable information is "EXCEPTION_CODE=3221225477".
I´ve changed "PUBLIC.MAX_QUERY_TASKS" to "1" and nothing happened.
As far as my production enviroment is a 24/7, I´ve returned it to SA184.108.40.20687.
I´m going to send the DUMP file to Sybase Brasil support team.
Until there, any clue?
asked 05 Jun '11, 18:13
I´ve found a way to reproduce the error on my development environment. It´s just a sub-query that in a few seconds uses ALL server memory...almost 7GBytes. This query crashes the primary, and then crashes the mirror, because my application keeps trying to run it again until the mirror becomes primary.
This is the query with false table names and false column names:
SELECT c_seq, b_name FROM schema.a A JOIN schema.b B ON a_b_seq = b_seq JOIN schema.c C1 ON user_seq = c_user_seq WHERE a_type_seq = 12 AND c_st = 's' AND (SELECT MIN(d_dt) FROM schema.c C2 JOIN schema.d D ON d_c_seq = C1.c_seq WHERE C1.c_seq = C2.c_seq ) >= date(current timestamp) ORDER BY b_name;
The mini-dump file you've sent into the Sybase Brasil support team is the key to this problem.
If you used dbsupport to submit the file, you can run "dbsupport -ls" to show the submission IDs of the mini-dump files you've sent to Sybase. If you post the numbers here, somebody might be able to have a quick look (no guarantees).
There is no answer that can be provided without looking at the mini-dump file, so the answer will need to be provided from the Sybase Brasil support team if you are e-mailing the dump files to them.