gabriel rosenkoetter on Tue, 22 Oct 2002 11:30:05 -0400 |
[Those of you not using Red Hat and their up2date can just stop reading now, probably. Though if you're having mysterious MySQL problems, this may be related.] There's no mention of it in: http://www.redhat.com/mailing-lists/redhat-watch-list/msg00445.html This is a security patch to glibc. I'm sure it's a good idea. But it utterly breaks MySQL's daemon. The specific problem I saw (last night, at 6:30 pm) was that safe_mysqld would catch a mysqld "hanging", kill it, and restart it every time a Win2k IIS ColdFusion machine tried to use ODBC to connect to the MySQL DB. Running mysqld in debug mode on the command line showed that the daemon was, in fact, seg faulting (with no core dump, but maybe I just had ulimit set to 0). I found someone with the same problem on a RH 7.1 mailing list: http://www.redhat.com/mailing-lists/seawolf-list/msg16780.html This response isn't too promising (note that the link provided to bugzilla is obviously broken; I can't find anything in bugzilla listing a *problem* with this update, just listing the update itself, but I didn't look too hard): http://www.redhat.com/mailing-lists/seawolf-list/msg16784.html But there's a quick, easy fix that doesn't involve backdating your glibc RPM: http://www.redhat.com/mailing-lists/seawolf-list/msg16794.html Presumably, an updated MySQL RPM will fix this, and the (performance-limiting) variables can be pulled out of my.cnf. -- gabriel rosenkoetter gr@eclipsed.net Attachment:
pgpeiysG5F1ET.pgp
|
|