Re: kproc

In article <94464a$pcg$1_at_nnrp1.deja.com>,   Snogfest Hosebeast <snogfest_hosebeast_at_my-deja.com> wrote:
> In article <942200$16a$1_at_nnrp1.deja.com>,
> Paul Bennett <spacey845_at_my-deja.com> wrote:
> > In article <3a6431cd.6595073_at_news.freeserve.net>,
> > alex_at_alenda.NOSPAM!.freeserve.co.uk (Alex Hudghton) wrote:
> > > On Mon, 15 Jan 2001 15:51:38 GMT, Paul Bennett <spacey845_at_my-
 deja.com>
> > > wrote:
> > >

 [snip]
> > > >
> > > >Should they disappear all by themselves? Are there any "common
> > > >pitfalls" that might kill them without any blatantly obvious
> warning?

[snip -- in summary, they're Oracle AIO processes]

> > I have min==1 and max==10
> >
> > With only one client connected to the server, I have 9 oracle-owned
> > kprocs running, all started at boot time. After disconnecting that
> > client, I have ... erm hold on ... the same 9 kprocs running.
> >
> > Does this sound a bit fishy to anyone else???
> >

 [snip]
> >
> I have our MAX set to 250.
> running "pstat -a | grep aio | wc" comes up with 250, so I need to
> increase it.

Thanks.

My concern was that even after all clients disconnect, the number of kprocs running does NOT diminish, even though the minimum required is set to one. That does not sound right to me, based on my limited knowledge. I'd've thought that with no activity, they'd drop to the minimum reasonably quickly (i.e. certainly in less than 24 hours!).

I've just thought: Would there be one running for each (or some) of the Oracle processes themselves? (e.g. ora_pmon_sid, ora_lgwr_sid, etc).

If so, I'll probably have to bump up the maximum a touch, wont I?

---
Pb
My opinions are my own and should not be taken as
representing any other group or entity.


Sent via Deja.com
http://www.deja.com/