squid-1.1-alpha15/16 FATAL storeAbort's?

From: Andrew Leahy <A.Leahy@dont-contact.us>
Date: Wed, 4 Sep 1996 22:23:23 +1000 (EST)

G'day Squider's,

I'm just a newbie to this list and to squid caches, so I'm not sure if
this is a "known" problem/issue with Squid, if it is I'm sorry to fill
your mailboxes!

We've been experimenting lately at replacing our CERN proxies with the
Squid software, early indications looked fantastic! We're getting a much
higher hit-rate, and the load on the serving machines is negligible,
basically it's bloody brilliant!

But... we're now experiencing problems with squid-1.1alpha15 on two boxes,
one host is a SparcServer (Solaris 2.5) and the other is an AlphaStation
(Digital Unix 4.0), both exhibit the same symptoms. These hosts are
configured as cache neighbours and also peer with a SunOS 4.1.3 server
running Squid-1.0.9 (which runs remarkably stable). During periods of what
I call "heavy use" for us (say 6000-8000 connections per hour), the 1.1
squids will crash and restart every 10-30minutes. With an error message
like this in the cache.log:

        FATAL: storeAbort: bad store_status

Is this a symptom of a configuration/compilation problem on my end? Or is
it a feature of the alpha squids! Any other pointers?

Andrew "Alf" Leahy, phone: 047 360385
Unix Administrator, mailto:alf@nepean.uws.edu.au
Systems & Ops, TASS, CCD, UWS Nepean http://www.st.nepean.uws.edu.au/~alf/
Received on Wed Sep 04 1996 - 05:25:21 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:32:56 MST