I think I had some table corruption.
After running isamck on my databases, I haven't
seen the problem in the mysqladmin proc anymore.
I also recompiled my kernel 2.0.36 again the new glibc 2.1
and things seem to be a lot better. It is the weekend, so
my web/mysql isn't exactly running at production speed
right now but I will report back during the week.
Another side effect is that now my linux mysql server is running
like my sgi mysql, i.e. very low amount of slow queries (actually
the sgi has never reported a slow query, ever) and more
tmp tables are being created. Hmmm.
Michael Widenius wrote:
> >>>>> "Vinny" == Vinny Stoessel <xaymaca@stripped>
> Vinny> I recently upgraded to mysql 3.22.22
> Vinny> on redHat 5.0 kernel Intel linux 2.0.36.
> Vinny> I also upgraded to glibc 2.1 and compiled
> Vinny> mysql against that. I have not recompiled the
> Vinny> kernel yet. MJy question is about updates with low_priority.
> Vinny> when I do a proc I sometime see a a table
> Vinny> with a lock even thought the query is updata low_priority.
> Vinny> Things work as expected with the inserts. I.e. they
> Vinny> wait for selects to finish before inserting.
> Vinny> Do update low_priority queries work differently?
> No; UPDATE should behave exactly like a INSERT.
> Next time you see this, could you do a cut and paste of it and post it
> to me?
Xaymaca Studios http://www.blendermania.com