Paul Stearns wrote: >As reported under the subject "Random Database Slowdowns..." on the >win32 list, our database still hangs on an average of 1-2 times per >day. > >I can find no error messages or logs associated to the problem. It >affects both IIS ADO connections as well as local connections from >tools such as mysqlcc, mysqladmin and command line tools such as mysql. > >I see no CPU activity associated with the "hangs". I cannot stop and >restart the service, but most of the time a reboot will resolve the >problem. Sometimes the problem will reoccur within a few minutes of a >reboot, other times it takes hours.
I can't swear it's the same problem, but we had very similar symptoms some time ago (version was around 4.0.8 or so). This was also on Mac OS X Server 10.2.x. Seemingly random queries would just not finish. They were queries we had run before and could even run at the same time from a different client. But this process would just not finish. It would sit in show processlist forever. If it was the only query running the mysqld cpu load would drop to around zero. There was no i/o activity if it was the only query running. Any temp files associated with the query wouldn't grow. Sometimes it was a small query, sometimes a big one. If you tried to kill the process from within the command line client or mysqladmin it would show up as killed in the process list but never die. Issuing mysqladmin shutdown wouldn't shut down the mysqld server b/c it couldn't kill off the queries either. Even kill -9 <mysqld pid> would hang the machine. The only solution was rebooting while mysqld was running b/c we couldn't shut it down. We tried moving to InnoDB and got the same situation (and show innodb status\G revealed no work was being done within InnoDB). We tried swapping RAM, swapping hard drives, changing drive formats (HFS+ to UFS), installing Yellow Dog Linux on the machine, etc.... I was pretty certain it was just a MySQL on the Mac issue, but then I pulled the drives and RAM and put them in an identical Mac. Same drives, same RAM, same data, same OS, same MySQL...the problem disappeared. We sent the machine back to Apple (it was new) and they replaced the mother board. That machine would show random errors even in just desktop use, so though the solution seems onerous we really didn't have much choice. If you can, I'd try running the queries on a different machine. Good luck, Ware -- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
