Re: [squid-users] cache level "fsck" for SQUID ?

Remove the swaplog (defaults to the top level of the cachedir) while
squid is offline. Remove the -clean one as well. That will force squid
to rebuild the cache state data.

Downside: you will lose all freshness data about the cache. So your hit
rates will be lower than usual for a short while.

Rob

----- Original Message -----
From: "Marc Elsen" <elsen@imec.be>
To: <squid-users@squid-cache.org>
Cc: <elsen@imec.be>
Sent: Tuesday, June 05, 2001 10:01 PM
Subject: [squid-users] cache level "fsck" for SQUID ?

>
>
> Hi, I am using 2.4.STABLE1 on Linux (Redhat 6.2).
>
> I am faced with an increasing number of TCP_SWAPFAIL_MISS
> errors in access.log which corresponds (e.g.) with :
>
> 2001/06/05 13:53:43| storeAufsOpenDone: (2) No such file or directory
> 2001/06/05 13:53:43| /imec/comp/squid/cache2/02/19E/000B3CCF
>
> (for instance).
>
> I have 4 cache dirs of 10Gb and I am using aufs.
> I now have a rate of 2 incidents/minute. The problem
> seems to be deteriorating.
>
> I was wondering whether there would be a tool in Squid, such
> as "fsck" at the Unix file system level, which could for instance
> after squid was shutdown, make a synced-state of what is cache.swap
> and the files in the cache dir(s) ?
>
> Or does squid handle these anomalies too, by just simply restarting
> it ?
>
> Marc.
>
Received on Tue Jun 05 2001 - 06:04:02 MDT