tag:blogger.com,1999:blog-31421954.post1770278078585734948..comments2023-10-30T08:23:12.960-07:00Comments on mySQL DBA, Architecture, Dev, Scale, HA, Code : MAJOR Problems in mysql-5.0.51Dathan Pattishallhttp://www.blogger.com/profile/00356367514107959723noreply@blogger.comBlogger6125tag:blogger.com,1999:blog-31421954.post-91972428363544119092008-02-07T09:31:00.000-08:002008-02-07T09:31:00.000-08:00Dathan, there is 1 known memory leak, which is to ...Dathan, there is 1 known memory leak, which is to do with index_merge queries. Please check if this bug is relevant: http://bugs.mysql.com/bug.php?id=27732Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-31421954.post-34402685658015568962008-02-03T15:30:00.000-08:002008-02-03T15:30:00.000-08:00tryinnodb_adaptive_hash_index = 0(5.0.52 or higher...try<BR/>innodb_adaptive_hash_index = 0<BR/>(5.0.52 or higher)<BR/>This should fix the locking issues. The slave / relay log corruption is a different issue, not sure what version the fix is making it into, but likely not in 5.0.54.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-31421954.post-10628887144974073782008-01-25T00:07:00.000-08:002008-01-25T00:07:00.000-08:00The relay log corruption bug is still open in 5.0....The relay log corruption bug is still open in 5.0.54. See http://bugs.mysql.com/bug.php?id=26489. Most of the time, doing 'stop slave, reset slave, change master to the SQL thread position' is sufficient to fix the corruption.<BR/><BR/>I don't understand the blocking problem that you mention. We have intermittent long lock waits on the adaptive hash index latch (the lock holder doesn't release it while doing something that take a long time -- like sorting). This can be fixed by adding calls in a few places to release the latch. Is that your problem?<BR/><BR/>I am using 5.0.37 now and hope to upgrade to 5.0.54.Mark Callaghanhttps://www.blogger.com/profile/09590445221922043181noreply@blogger.comtag:blogger.com,1999:blog-31421954.post-84015864532626672992008-01-16T13:16:00.000-08:002008-01-16T13:16:00.000-08:00We also experienced the relay log corruptions with...We also experienced the relay log corruptions with high I/O and load, and also that it had the tendency to crash MySQL. One good thing about 5.0.54 is that this bug is fixed:<BR/><BR/>http://bugs.mysql.com/bug.php?id=31793<BR/><BR/>However, we still have a script to automatically reset replication and start the slave threads anew.veganloveburgerhttps://www.blogger.com/profile/11935117470776753328noreply@blogger.comtag:blogger.com,1999:blog-31421954.post-68841183210986440612008-01-14T13:06:00.000-08:002008-01-14T13:06:00.000-08:00So these locking problems seem to be InnoDB-specif...So these locking problems seem to be InnoDB-specifc. Thanks!Steffen Weberhttps://www.blogger.com/profile/16399610489555689619noreply@blogger.comtag:blogger.com,1999:blog-31421954.post-50100712787583618952008-01-14T12:12:00.000-08:002008-01-14T12:12:00.000-08:00Could you please be a little bit more specific abo...Could you please be a little bit more specific about the "huge blocking locks"?Steffen Weberhttps://www.blogger.com/profile/16399610489555689619noreply@blogger.com