Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: large pool always filling up
"Howard J. Rogers" <hjr_at_dizwell.com> wrote in message
news:41acf645$0$12650$afc38c87_at_news.optusnet.com.au...
> Alan wrote:
> > "Howard J. Rogers" <hjr_at_dizwell.com> wrote in message
> > news:41abc00d$0$17542$afc38c87_at_news.optusnet.com.au...
> >
> >>Gerry Sinkiewicz wrote:
> >>
> >>>"Mr Boombastic" <shaggy_at_popstar.com> wrote in message
> >>>news:76dee8e3.0411291210.5b03b342_at_posting.google.com...
> >>>
> >>>
> >>>>Hello
> >>>>
> >>>>We are running an Oracle 8.1.7.4 database in production and we have a
> >>>>problem on one of our databases where the large pool is always filling
> >>>>up and on an almost daily basis gets to more than 90 - 100 % full. How
> >>>>do we identify any problem queries? Statspack info on the large pool
> >>>>is limited, as is anything on Metalink.
> >>>>
> >>>>And yes, we use MTS on the database. I'm not sure where to start the
> >>>>tuning process really as previous investigations have all yielded a
> >>>>blank (where we mostly used Statspack for analysis).
> >>>>
> >>>>One possible avenue i have in mind is to investigate the dispatcher
> >>>>process (v$dispatcher, v$dispatcher_rate) and also identifying any
> >>>>shared server contention. Would reducing any MTS help our issue?
> >>>>
> >>>>Im not sure where else to look really. Maybe an experienced DBA out
> >>>>there can help and point me in the right direction?
> >>>
> >>>
> >>>How big is you large_pool?
> >>>
> >>>If it is less than 300MB, think larger.
> >>
> >>
> >>
> >>Without more detail from the OP, I think that such advice takes us
> >>nowhere. If we read his post literally, then his large pool is actually
> >>too large (he reports it only being between 90% and 100% full, meaning
> >>potentially 10% is just wasted memory), and he should therefore actually
> >>reduce its size.
> >>
> >>He doesn't report getting ORA-04031 error messages, for example. He
> >>merely seeks to know which queries are consuming the large pool.
> >>
> >>Generic advice on how to size caches is not wise at the best of times (I
> >>very rarely have ever needed a large pool bigger than 48M, for example,
> >>but I wouldn't want to read too much into that number either). But when
> >>it doesn't even answer the actual question asked...
> >>
> >>Regards
> >>HJR
> >
> >
> > Actually, if you read all of his posts,
>
>> > large pool usage is the problem. His symptom is TNS errors, and he has
> > he is making an assumption that the
>
>
>