Re: sort latch usage in MTS mode

It is interesting that after upgrade to solaris 10, the latch: shared pool contention is gone.
so 920/solaris8 works fine.
10.2.0 /solaris 10 works fine.

10.2.0 /solaris 8 MTS sucks. (especially when there are lots of sorts/hash going on)

oracle has admit there is bug inside and working on it. On 8/7/07, Nigel Thomas <nigel_cl_thomas_at_yahoo.com> wrote:
>
>
>
>
>
> Zu Chao wrote: maybe not many people use MTS recently due to nearly-free
> RAM. But when connection# goes real high, you still have to, and we are such
> a user.
>
> Kevin Closson replied: just out of curiosity, what is "real high" in terms
> of connect count?
>
> And another (maybe irrelevant) point: when there is a mid-tier shared
> connection pool, that can also act as a concentrator/multiplexer (eg I have
> LoadRunner tested 2000 "users" executing a realistic high volume workload
> with (iirc) only 100-200 real Oracle connections). Yes, I know 2000 is not
> "real high" :) but you can scale out this approach.
>
> MTS and app server shared pool have different strengths and weaknesses (for
> the benefit of the BAAG party - or just to stimulate more informed replies -
> I'll refrain from proving my ignorance by guessing what they are :-).
>
> Regards Nigel

-- 
Regards
Zhu Chao
www.cnoug.org
--
http://www.freelists.org/webpage/oracle-l