Re: swap space declines on ultrasparc1/Solaris2.5 with Oracle Workgroup Server

From: Mark Rosenbaum <mjr_at_netcom.com>
Date: 1997/01/03
Message-ID: <mjrE3FvwM.Eu6@netcom.com>#1/1

In article <32CC0B01.3533_at_nfesc.navy.mil>, Geoffrey Dann <geoff_at_nfesc.navy.mil> wrote:
>I have a new Ultra-1 with 64 MB of memory, two 2.1GB disks,
>and Solaris 2.51. I installed Oracle Workgroup Server 7.3.
>The swap partition is 192 MB.

64 MB for Oracle 7.3 strikes me as small. You may want to increase memory.

>Swap space goes downhill over a few days, and the only
>solution I know is to reboot. I've tried using the
>default setup, where /tmp is mounted on the swap partition,
>and using /tmp separately.

Sounds like a memory leak in some software. Is Oracle the only thing running? Can you look at the memory for each process that runs just after reboot and just before the next one. If you are almost ready for an other reboot it might help to look before and after.

I don't remember there being any differences on Solaris so a PS command should give the virtual and real memory for each process running.

The TEE command will let you see the output and save it to a file.

Once you have the two (before and after) views of the system compare.

The easist way might be to use the CAT command to concatenate the two files together and then use the SORT command to sort the file on the command name field. That way the before and after for each command will be next to each other.

All of these commands should be explained by the MAN command.

Hope this helps

Mark Rosenbaum			Otey-Rosenbaum & Frazier, Inc.
mjr_at_netcom.com			Consultants in High Performance and
(303) 727-7956			Scalable Computing and Applications
POB 1397			ftp://ftp.netcom.com/pub/mj/mjr/resume/

Boulder CO 80306 Received on Fri Jan 03 1997 - 00:00:00 CST