In the last episode (Feb 19), Adam Powell said:
>
> Watching top on our database machine, I see the odd mysqld process that
> claims 99% CPU for about 3-4 seconds... which is bad. I run a large site,
> and its impossible to keep track of every single mysql command within the
> site. One of them is a badly written query that is slowing down all the
> rest of the queries.
>
> Is it possible to work out exactly what query is causing this? Can I
> somehow get this information by looking at the PID of the process? Is there
> a way I can automatically log any queries that take over a certain time to
> execute?
3.23.10 has a "slow-query" logfile, which logs a timestamp, execution
time, and the query text for any queries exceeding the
"long_query_time" value (10 seconds by default, but adjustable).
If you don't want to run 3.23, your best bet would be to have a shell
open with "mysqladmin proc" all typed in, and hit ENTER when you see
mysql start hogging the CPU :)
--
Dan Nelson
dnelson@stripped
