Brandon Metcalf wrote:
> We are running 3.23.58 on Solaris 8 and seeing the following problem.
> Periodically, mysqld gets into a state where we can't query just one
> table in all of our DBs--queries just hang. Restarting mysqld always
> fixes the problem, but a SIGKILL is required to stop it.
>
> Could this be something is corrupt in this particular table? Would
> dumping this DB and reloading it help?
this happens to me without fail on Linux if i set max connections or
table cache too large. i get "Error in accept: too many open files" in
the mysql error log and then the db hangs as you describe, requiring a
SIGKILL to get it back.
make sure you run myisamchk after killing though - you never know how
messed up your tables are.
-jsd-
