Re: [SQU] Cache Digests not Building

From: Paul Kincaid <pkincaid@dont-contact.us>
Date: Tue, 31 Oct 2000 17:17:18 -0500 (EST)

I found within the FAQ regarding just general Squid debugging (ie squid -k
debug), but could not find anything about Digest debuging -- section 16.13
really just deals with debugging after the digests are up and have been
swapped between each other.

I see within my cache.log (with full Squid debugging enabled) that it is
trying to go out and get the Digest from its peers (ie I see "GET
http://peer1.xxxxx.com/squid-internal-periodic/store_digest") but I am
assuming that I am getting back the same results as if I try to connect
locally via Lynx. There is nothing that I can decipher that makes it look
like it fails and any sort of explanation. A little further on it looks
like it is trying to determine its peer digest's status:
        peerDigestLookup: peer xxx.xxx.xxx.xxx
        peerDigestLookup: gone!
        peerDigestLookup: peer yyy.yyy.yyy.yyy
        peerDigestLookup: !ready && !requested
        peerDigestLookup: peer zzz.zzz.zzz.zzz
        peerDigestLookup: !ready && requested

The first request is going to the parents which are set up as no-query (ie
Firewalls), the second and subsequent ones are to its peers. All are
!ready (am I to assume that is "not ready") and its half and half whether
it is !requested or requested.

Then I get:
        neighborDigestSelect: choices 0 (0)
        peerNoteDigestLookup: peer <none>, lookup: LOOKUP_NONE

Am I to assume that this is telling me that it did not find the neighbor's
digests?

Having said all that, where in all of this mess is there anything that
would be helpful for debugging the local web cache. I see absolutely
nothing in the cache, access or store logs to the local cache's digest
creation.

I do have control over all the caches (siblings, parents, etc.)

Thanks for your quick response Alex.

On Tue, 31 Oct 2000, Alex Rousskov wrote:

> 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
Received on Tue Oct 31 2000 - 15:17:44 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:56:04 MST