problem with mysql-crashes
Hey folks, before posting to the bugs-list I would like to try it here to get some help :). Every now and then MySQL-4.0.13 crashes on my Dual-Athlon-MP-machine. Its not reproducible nor I know a query which does it. (The log shows that at the end, so no query :-( : Trying to get some variables. Some pointers may be invalid and cause the dump to abort... thd-query at 0x5e13a478 is invalid pointer thd-thread_id=2223573 ) The server is the standalone db-server for one accessing webserver, its running RedHat-Linux 7.2 with all their latest patches. Kernel 2.4.19-XFS with MySQL running on an XFS-filesystem. No replication but query-cache enabled. Ok but now the worse thing started for 2 times I already got this problem in the logs: Number of processes running now: 16 mysqld-max process hanging, pid 17193 - killed mysqld-max process hanging, pid 17191 - killed mysqld-max process hanging, pid 17190 - killed mysqld-max process hanging, pid 17145 - killed mysqld-max process hanging, pid 17144 - killed mysqld-max process hanging, pid 17098 - killed mysqld-max process hanging, pid 17025 - killed mysqld-max process hanging, pid 16987 - killed mysqld-max process hanging, pid 28393 - killed mysqld-max process hanging, pid 28392 - killed mysqld-max process hanging, pid 28391 - killed mysqld-max process hanging, pid 28389 - killed mysqld-max process hanging, pid 28388 - killed mysqld-max process hanging, pid 28387 - killed mysqld-max process hanging, pid 28386 - killed mysqld-max process hanging, pid 28385 - killed 030722 20:46:26 mysqld restarted Afterwards there is a init-process taking 99% from one CPU. The first time I tried to reboot the machine, which didn't work as the shutdown process was hanging then. Any ideas whats wrong and how to fix it? Didn't see a mention of a fix for that problem in the 4.0.14-changelog, will it still fix it? Thanks in advance, Thomas -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: problem with mysql-crashes
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Hi, On Wed, 23 Jul 2003 [EMAIL PROTECTED] wrote: before posting to the bugs-list I would like to try it here to get some help :). Every now and then MySQL-4.0.13 crashes on my Dual-Athlon-MP-machine. Is that a self-compiled binary or one of ours? Its not reproducible nor I know a query which does it. (The log shows that at the end, so no query :-( : Trying to get some variables. Some pointers may be invalid and cause the dump to abort... thd-query at 0x5e13a478 is invalid pointer thd-thread_id=2223573 ) Not very helpful. Does this occur under high load only? The server is the standalone db-server for one accessing webserver, its running RedHat-Linux 7.2 with all their latest patches. Kernel 2.4.19-XFS with MySQL running on an XFS-filesystem. No replication but query-cache enabled. What version of glibc is this system running on? Is it 2.2.3 by chance? Ok but now the worse thing started for 2 times I already got this problem in the logs: Number of processes running now: 16 mysqld-max process hanging, pid 17193 - killed mysqld-max process hanging, pid 17191 - killed mysqld-max process hanging, pid 17190 - killed mysqld-max process hanging, pid 17145 - killed mysqld-max process hanging, pid 17144 - killed mysqld-max process hanging, pid 17098 - killed mysqld-max process hanging, pid 17025 - killed mysqld-max process hanging, pid 16987 - killed mysqld-max process hanging, pid 28393 - killed mysqld-max process hanging, pid 28392 - killed mysqld-max process hanging, pid 28391 - killed mysqld-max process hanging, pid 28389 - killed mysqld-max process hanging, pid 28388 - killed mysqld-max process hanging, pid 28387 - killed mysqld-max process hanging, pid 28386 - killed mysqld-max process hanging, pid 28385 - killed 030722 20:46:26 mysqld restarted Afterwards there is a init-process taking 99% from one CPU. The first time I tried to reboot the machine, which didn't work as the shutdown process was hanging then. Any ideas whats wrong and how to fix it? Didn't see a mention of a fix for that problem in the 4.0.14-changelog, will it still fix it? What version of glibc is being used on that system? I noticed that you use the Max binary, which is dynamically linked - so it will use the locally installed libc.so. Is there any special reason for using Max? Or could you try to use our statically linked Standard binary instead? The problem above looks very similar to the problems we had when we linked against glibc 2.2.3 instead of 2.2.5: under high load, a mysqld thread starts eating up all CPU resources on an SMP system. Unfortunately we never found the reason for that - it did not seem to happen when using glibc 2.2.5... Bye, LenZ - -- Lenz Grimmer [EMAIL PROTECTED] Senior Production Engineer MySQL GmbH, http://www.mysql.de/ Hamburg, Germany For technical support contracts, visit https://order.mysql.com/?ref=mlgr -BEGIN PGP SIGNATURE- Version: GnuPG v1.2.2-rc1-SuSE (GNU/Linux) Comment: For info see http://quantumlab.net/pine_privacy_guard/ iD8DBQE/HoQOSVDhKrJykfIRAh+MAJ91+kr7H92xcK4JtFhz5HQ3r6LGgwCggT/v OCHlqAabKjd6FnL6PMOmKRg= =U2ag -END PGP SIGNATURE- -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]
Re: problem with mysql-crashes
Hi Lenz, thanks a lot for your reply! Is that a self-compiled binary or one of ours? It are the MySQL-4.0.13-MAX-RPMs from your site. Not very helpful. Does this occur under high load only? Yeah I know, otherwise I would have tried to generate a reproducible testcase with a given query or such. Any way to have valid pointers there? Nope, the load on that system is mostly around 0.5 with peaks up to 1.5, not something I would call a high load for a dual-system. Also the crashes are at many different times (even in the deep night) so I don't think its related to load :-(. What version of glibc is this system running on? Is it 2.2.3 by chance? Nope, sorry. Its glibc-2.2.4-32. What version of glibc is being used on that system? I noticed that you use the Max binary, which is dynamically linked - so it will use the locally installed libc.so. Is there any special reason for using Max? Or could you try to use our statically linked Standard binary instead? Yeah I could try that instead, maybe I should try with 4.0.13 to see if its a problem of that version. Or should I better try 4.0.14 already? I was hosting a couple of bdb-tables for some time, but no longer now. Currently only myisam and innodb-tables so I could try the standard-built too. The problem above looks very similar to the problems we had when we linked against glibc 2.2.3 instead of 2.2.5: under high load, a mysqld thread starts eating up all CPU resources on an SMP system. Unfortunately we never found the reason for that - it did not seem to happen when using glibc 2.2.5... Hmm, maybe 2.2.4 has the same problem? The problem is that the init-process is taking all the cpu, didn't see a mysql-process doing the same before. Thanks, Thomas -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe:http://lists.mysql.com/[EMAIL PROTECTED]