Re: corrupt digest

From: Alex Rousskov <rousskov@dont-contact.us>
Date: Wed, 13 Jan 1999 17:30:35 -0700 (MST)

On Thu, 14 Jan 1999, Daniel Schneider wrote:

> we are using Squid 2.0 Patch 2 and we have enabled digest some days ago.

Upgrade to 2.1.P2! An important bug was fixed in that version that may
affect your digests.

> we have to caches: wwwproxy and wwwproxy2. Both use digest to ask each other.
> wwwproxy has been requesting many more digests from wwwproxy2 than the other
> way round, although wwwproxy is much larger than wwwproxy2.

Size does not (and should not) matter. All digests are refreshed every
hour. Ideally, digests should be refreshed when information becomes stale,
but that is harder to track.

Note that a proxy may ask for a digest and receive a [tiny] 304 reply. It
will then ask for a fresh digest again (after 304 expires). That may
happen when two proxies do not have synchronized clocks and perhaps for
other reasons. It is not a problem. Check access.log for reply status
(200 or 304). If you get digests with 200 status more often than once per
hour, you do have a problem.

> Well, cache.log of
> wwwproxy tells me now that he has got an corrupt digest from wwwproxy2 every 10
> mins. It requests a new digest.

There may be several reasons (and I am not sure which message you are
referring to). Consider increasing the debugging level for module 72 to,
say, 5. If that does not help, please send us the debugging info.

Thanks,

Alex.
Received on Wed Jan 13 1999 - 17:27:49 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:44:01 MST