>>>>> "Timothy" == Timothy Chi <tchi@stripped> writes:
Timothy> I've actually been experiencing the same things on my server.
Timothy> Basically, when I start things up, everything if fine for awhile. Slowly
Timothy> but surely, overall performance begins to degrade as MySQL continues to eat
Timothy> up all the cpu. Eventually, the system basically comes to a grinding halt,
Timothy> with load average hovering around 17.65, Mysql using about 78%.
Timothy> Sun E450
Timothy> 4 400Mhz Processors
Timothy> 3.5 Gig RAM
Timothy> Solaris 2.6 w/ all latest patches (Y2k cluster + latest patch)
Timothy> I'm running Mysql 3.22.23b
Timothy> (Also, Apache 1.34, mod_perl 1.16)
Timothy> Why do all the processes just enter into a Query mode after a short amount
Timothy> of time and never return? My queries are extremely simple, usually straight
Timothy> SELECT col FROM tablename WHERE col1=val1 AND col2=val2.
Timothy> Any ideas on why this might be happening? I think its very similar to
Timothy> Brian's problem. I can probably deliver any type of debugging info you
Timothy> need. Upgrading to 3.22.26 did not help.
Timothy> Any insight or suggestions you could offer would be greatly appreciated.
Timothy> Here's a processlist and ext (after performance is nil):
It would be extremely useful if you can send me the output from
pstack when this happens again! This will give us a back trace for each
thread so we can find out what it's doing!
Another option is to run mysqld under gdb 2.8 and when this happens
again provide a back trace for each tread!