Re: [SQU] Cache Digests not Building

Paul,

        If you have administrative control over the siblings, it
should be fairly easy to debug the problem. See FAQ for tips on how to
enable digest debugging.

Alex.

On Tue, 31 Oct 2000, Paul Kincaid wrote:

> Running Squid 2.3.STABLE4 with IPF Transparent and Cache Digests enabled
> at compile time -- I am trying to get Cache Digests to work...
>
> When accessing
> http://localhost:<squid-port>/squid-internal-periodic/store_digest, the
> only thing that is returned is "The cache is currently building its
> digest". I also just noticed that when trying that address from Lynx, it
> first gives a 404 - Not Found error, then goes to the "The cache is
> currently...."
>
> The mgr:store_digest shows:
>
> store_digest: size: 4262 bytes
> entries: count: 4836 capacity: 6818 util: 71%
> deletion attempts: 0
> bits: per entry: 5 on: 14745 capacity: 34096 util: 43%
> bit-seq: count: 16805 avg. len: 2.03
> added: 4836 rejected: 1983 ( 29.08% ) del-ed: 0
> collisions: on add: 0.87 % on rej: 1.06 %
>
> There are two halves to this network (ie 2 distinct paths for users to get
> out), each half has 2 Squid Boxes (running BSDi 4.1 BTW). All 4 boxes are
> acting as siblings and they are all set up to forward requests to a
> firewall suite acting as a default, no-query, no-digest parent. ICP is
> disabled via "icp_port 0"
>
> The transparency and actual Squid functions are working like a champ, no
> problems there. And if ICP is enabled, that too works like a champ, but
> we want to get Digesting working.
>
> Now for the questions.... Obviously, what does the message I am receiving
> via the web page mean, other than the obvious? Where is the digest stored
> after it is built in memory and saved to disk (thinking in terms of the
> digest-url=URL flag for the cache_peers)?
>
> Is there a good location to look for help/explanations for the variables
> used in the client mgr:menu diags? (Or Cachemgr.cgi for that matter)
>
> I found an email explaning the store_digest output that was very helpful
> in understanding that output - more like those would be great. (Thank you
> Alex Rousskov).
>
> Thanks,
> Paul
>
>
>
>
> --
> To unsubscribe, see http://www.squid-cache.org/mailing-lists.html
>
>

--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html