I realize this is not a new issue. I've read quite a bit about it over the
last couple days. However, I've seen very little to suggest a reason and/or
fix for when this happens frequently and randomly. Fortunately we have not
lost any data (yet?), however we have sites that go down temporarily until
we repair the offending table.
We didn't notice this problem until we replaced our main DB server with a
newly rebuilt one running RH Linux with the latest patches and MySQL
4.0.13. The only significant job of the server is to run MySQL and it does
not replicate in any direction outside of the binary logs.
The tables that are most often effected are our session tables of our web
application that are frequently accessed with selects, inserts, updates, and
deletes, and on every web page. We have several sites that run on different
db's, but with the same table structure.
I've tried dumping, dropping, and importing the whole table several times.
I've run repair and optimize while the server is up and myisamchk while the
server was shutdown. All the tables MyISAM and the results of the repair
process tells me there was a problem with the indexes.
Does anyone know why this is an ongoing problem? I've started looking into
configuration variables for the MySQL server, but our my.cnf is virtually
the same as it always has been, stock my-medium.cnf, and the same for just
about every version we've ever used.
Thanks,
Gary
_________________________________
Gary Cornelisse
Senior Applications Programmer
Conduit Internet Technologies, Inc.