After upgrading to squid-2.4-STABLE6, from 2.4-stable3, the problem seems
to have gone away. I'm going to test it over the next week or so to make
sure.
For historical purposes, when we restarted squid (after issuing the
reconfigure command) the stale pages would still be served. I don't
_believe_ that restarting squid would trigger the issue, but I didn't test it.
There wasn't anything interesting when "reload" squid, except a warning
about override-expire violating HTTP standards. Interestingly though, that
error message is not logged when squid starts.
Thanks for your help, and I'll e-mail again if the problem re-appears.
-Ryan Casey
At 11:38 PM 4/1/2002 +0200, Henrik Nordstrom wrote:
>"squid -k reconfigure" (what actually is done by RedHat when you
>"reload" the service) is not very much tested.. there may well be
>bugs there but it do not seem likely this problem could be caused by
>a reconfigure bug..
>
>Do the same thing happen if you restart Squid?
>
>Do you get anything interesting in cache.log then you "reload" Squid?
>
>Regards
>Henrik
>
>
>On Friday 29 March 2002 22:11, Ryan Casey wrote:
>
> > Hi Henrik --
> >
> > I just wanted to follow up with you on this. I've been able to
> > reproduce this. Basically squid will correctly expire a stale page
> > until I tell it to "reload" its configuration ("/sbin/service squid
> > reload" on my redhat system). After that, it will still serve a
> > stale page, and the file content itself is old not just the
> > headers. I did try switching from diskd to aufs just in case that
> > mattered, but it doesn't. I'm going to try upgrading to
> > 2.4-stable6 to see if that solves it.
> >
> > Do you have any suggestions for debug options? The cache server is
> > currently handling around 1/2 million hits per day, so I want to be
> > careful not to log too much.
> >
> > Thanks,
> > -Ryan Casey
Received on Thu Apr 04 2002 - 10:57:06 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:07:21 MST