SUMMARY: out of swap space

From: Tim Winders <twinders_at_spc.cc.tx.us>
Date: Fri, 8 Nov 1996 22:10:50 -0600 (CST)

Thanks to the following people for responding and pointing me in the right
direction!

Brian Sheehan <sheehan_at_scripps.edu>
Kent R Arnott <karnott_at_falcon.tamucc.edu>
alan_at_nabeth.cxo.dec.com
"Mark E. Glidden" <glidden_at_poseidon.harvard.edu>
"WHITTAKER, Bruce" <bjw_at_ansto.gov.au>
Vince LaCava <vlacava_at_magicnet.net>
CHRIS RUFFIEUX <ruffieux_at_infi.net>
Thomas Leitner <tom_at_finwds01.tu-graz.ac.at>
Andrew Leahy <A.Leahy_at_st.nepean.uws.edu.au>
16ES0001..DFS016#c#FOEHA016.GROUPWISE.DFS_at_NOTES.compuserve.com
Mireille BOF <mbof_at_presn2.u-nancy.fr>
Becki S Kain <bkain1_at_ford.com>
Tom Rioux <rioux_at_ip6480nl.ce.utexas.edu>
Louis Bouchard - Bouygues Telecom <bouchard_l_at_decus.fr>
Pirie Hart <pirie_at_u.washington.edu>

My original question:
===========================
I am starting to get console messages saying I have less than 10MB free
swap space and don't know why/what to do about it (other than reboot).

Configuration.

Alpha 1000 4/233, 128MB RAM, RZ28M (2GB) RZ29M (4GB). I am running ADVFS
on all partitions. The RZ28M has the root domain on partition a and has
the swap space on partition b. I changed the default disklabel to combine
partition g and h into one area (so h was zero in size) and that was set
as my usr domain. I added the RZ29M and configured partition c (the whole
thing) as ADVFS and combined it with the other usr domain to have about
5GB or /usr space. (But didn't add a secondary swap space).

I thought the swap I had would be enough with the 128MB RAM. This system
runs as my mail server, it is running Netscape Enterprise Server 2.0,
NS Proxy 2.5b, DHCP, CAP, Samba, POP3, IMAP4, named. It is my primary
(only) "internet" box.

Are there any suggestions to reconfigure the system, monitor activity etc.
I am not really sure where to proceed. Thanks so much!
=======================

Some additional information. The swap partiontion (rz0b) is 196MB in
size. This can easily be determined by running /sbin/swapon -s. They
system was running in lazy mode already. The general concensus was that
you should have between 2-4 times the swap space as RAM depending on
conditions. I have 1.5 times...

Many people suggested changing to lazy mode if I was in eager swap mode.
You can tell which mode you are running by checking the file
/etc/swapdefault. It is exists, you are running eager swapping. If not,
you are running lazy swapping. This is explained in the swapon man pages.

In my case I had to either add swap space or kill some processes. I will
eventually add more swap space, but for now, I killed the NS-Proxy server
and my swap utilization went from 90+% (after 3 hours from reboot) to 5%
(after 8 hours from reboot). I don't know if the ns-proxy 2.5beta has a
memory problem or just a much bigger hog than 2.0, but that fixed my
problem.

Thanks to all again!

=== Tim
Received on Sat Nov 09 1996 - 05:16:36 NZDT

This archive was generated by hypermail 2.4.0 : Wed Nov 08 2023 - 11:53:47 NZDT