>>>>> "David" == David Johnson <davej@stripped> writes:
David> We had our production database go bezerk and take 100% of the CPU for about 5
> minutes before we had to kill it. There is speculation that we had some table locking
> contention, but it seems to me that a bad lock wouldn't cause the locked threads to sleep
> and wait instead of eating all the CPU. Am I right?
Normally bad locking will put all threads to sleep, but there are
always exceptions to this rule.
Did you manage to do a 'mysqladmin proc' when this happen?
|• Quickie||David Johnson||11 May|
| • Quickie||Michael Widenius||11 May|
| • Re: Quickie||David Johnson||11 May|