Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.server -> Re: memory process use is 10M is that normal?mlml
(markp7832_at_my-deja.com) writes:
> In article <379591b1.8033000_at_news.cis.dfn.de>,
> stojentin_at_orgdv.din.de wrote:
>> Hi Mike, >> >> please keep in mind that you see the whole memory >> consumption, regardless if it is shared memory or not. >> The shadow process for a client (wich keep memory >> for all connections actually open from this maschine) >> uses different memory areas in different modes. >> Maybe you see a lot of shared memory, which counts >> only once over all processes but can be seen when >> looking at each of these processes >>
The error message is from the HPUX os:
This message comes up when trying to run a program that uses oracle
on the local server:
Pid 15264 received a SIGSEGV for stack growth failure.
Possible causes: insufficient memory or swap space,
or stack size exceeded maxssiz.
/usr/visibility/v63/jds/v63ln/menu.sh[25]: 15264 Memory fault(coredump)
Pid 15265 received a SIGSEGV for stack growth failure.
Possible causes: insufficient memory or swap space,
or stack size exceeded maxssiz.
/usr/visibility/v63/jds/v63ln/menu.sh[25]: 15265 Memory fault(coredump)
Pid 15266 received a SIGSEGV for stack growth failure.
Possible causes: insufficient memory or swap space,
or stack size exceeded maxssiz.
--
> 7.2 down to 7.0 if an application executed a large package and the
> shared pool did not have a contiguous memory area large enough to hold
> the entire package you would get an error. I can not remember if this
> was fixed in 7.3 or not. We have pined all our large packages, 48 -
> 100K, since ver. 7.0 so if you are using packages you can use the
> dbms_shared_pool package to pin them at instance start-up and see if
> this helps.
>
> We were told before we bought Oracle to plan on 5M per user for our
> purchased MRP package which was written in Forms 3.0. Newer software
> uses even more memory so I do not think 10M per user process is
> excessive. How much real memory do you have? Did you configure your
> Unix swap to be at least 2 to 4 times real memory?
>
> Also if you have a large SGA you may find that the overall system runs
> better with less memory in the SGA and more advailable to the OS. If
> you are short of memory you could use MTS to reduce OS requirement for
> the user processes, but you may have to increase the shared_pool size
> to support MTS. That brings up another possibility, if you are using
> packages you could just need a larger shared pool to begin with.
>
> I hope these ideas help. Please post the actual error messages and
> someone may be able to give you a specific answer.
>
>
>
>
> --
> Mark D. Powell -- The only advice that counts is the advice that
> you follow so follow your own advice --
>
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.
--
/-----------------------------------------------/ \--\--\ \-- \-- \-- / www.ncf.carleton.ca/~ae299 /---\ \ \ \\ \ \ / ae299_at_freenet.carleton.ca /-----\ \ \\-- \-- \-- /-----------------------------------------------/Received on Thu Jul 22 1999 - 07:44:29 CDT