solaris 7 & squid was Re: Possible memory leak ?

hi alexander,
 
> > >I'm have some problems with my 2.2.STABLE5 squid.
> > >It became very slow after some amount of requests. With small traffic, it
> > >going down not fast, but at some time it became down too.

we're running Version 2.2.STABLE5-hno.19991127 under solaris 7 with no hassles.
the machine is an ultra-1, 512M memory, 20G disk for the cache filesystems.
we're supporting about 650 users with this config. the system is dedicated to
squid & is not under any stress at all.

we're running squid under the SOCKS5 shared library - http://www.socks.nec.com
so we can run squid on an internal host & have it handle internet requests
via a socks5 server running on a firewall. intranet requests are handled
via direct connects. routing decisions are made from the socks5 config.

the changes made to the default solaris 7 config are:

in /etc/system:
set priority_paging = 1

solaris7 recommended + security patches installed - you can fetch these
from http://sunsolve.sun.com
also installed is:
106980-05 SunOS 5.7: libthread patch

the cache filesystems are mounted with the "noatime" option. /etc/vfstab
entries for cache filesystems look like so:
/dev/md/dsk/d3 /dev/md/rdsk/d3 /var/cache1 ufs 2 yes noatime

in /etc/nscd.conf:
        suggested-size hosts 1001
this figure was found after numerous looks at the output of 'nscd -g'.
nscd hit rate is generally around 99%. there is also a caching bind-8.2.2-P5
named running on the host. some sources of info say this is not an ideal
config for squid, but it's necessary in this situation to handle both
internet & intranet dns lookups.

changes made to squid config are as follows:

in src/aiops.c
#define NR_ASYNC_IO_THREADS 32

squid was compiled with:
./configure --prefix=/usr/local/pkgs/squid-2.2STABLE5 --enable-delay-pools
--enable-async-io
using CFLAGS env variable of:
CFLAGS=-g -O2 -Wall -mcpu=ultrasparc
to gain support for the ultrasparc processor.

within squid.conf changes made are:
cache_mem 24 MB
# bandwidth in australia is expensive
maximum_object_size 64 MB
dns_children 10
# increase to handle large POST requests for internal uploads
request_size 1024 KB
# turn off as squid doesn't seem to release the fd's associated
# with 1/2 closed connects
half_closed_clients off

these changes are mostly culled from recommendations found in the squid-users
archive & the FAQ. searchable archives of lots of useful mailing lists,
including squid-users are at:
http://www.progressive-comp.com/Lists/
hopefully someone will find this mini-summary useful.

cheers,
pauline
Received on Thu Dec 09 1999 - 17:01:31 MST