Hi,
On Wed, Aug 04, 2004 at 11:28:09AM +0200, Henrik Nordstrom wrote:
> On Wed, 4 Aug 2004, A. Sajjad Zaidi wrote:
>
> > > 2) set half_closed_clients off in squid.conf
> >
> > Worth a try though I don't see what this has to do with the stability
> > issue.
>
> Depends on what you see in cache.log.. Setting this off makes Squid pretty
> much reactive to failed client requests.
I just reinstalled the whole system with Slackware 10.0 and Squid
2.5.STABLE6 with all the patches. Squid was compiled with the following
options:
--prefix=/usr/local/squid --enable-snmp --enable-linux-netfilter \
--enable-cache-digests --enable-truncate --enable-async-io=160 \
--enable-storeio=ufs,aufs --enable-removal-policies=lru,heap \
--enable-poll --enable-underscores \
--enable-cachemgr-hostname=pxy.xxxx.xxx
Also changed the partitioning so there are 2 16GB caches, one on each
disk, instead of one large cache like before. half_closed_clients is set
to off.
Seems to have survived the weekend without any problems or degraded
service. Will check again after the cache is full (should take about 2-3
more days).
Thanks for the help.
> collecting the information required for filing a Squid bug report is a
> good start in narrowing down your problem. See the FAQ on how to report a
> bug.
It didn't really look like a Squid problem, but I'll find out soon
enough and will file a bug report if it is.
-- A. Sajjad Zaidi GnuPG Key ID: 0xD7AD0E13 "What's blue and square? An orange in disguise..."Received on Mon Aug 09 2004 - 01:02:05 MDT
This archive was generated by hypermail pre-2.1.9 : Wed Sep 01 2004 - 12:00:02 MDT