[Gllug] Ubuntu Apache proxying problems

[- Tethys]

We've just migrated some of our web sites from a CentOS box to an
Ubuntu one (7.10). Apache is generally used as a front end, reverse
proxying requests through to something else (CherryPy, Django, etc.)
on the back end. However, after restarting the back ends, we're
getting the following error:

> Service Temporarily Unavailable
>
> The server is temporarily unable to service your request due to maintenance downtime or capacity problems. Please try again later.
> Apache/2.2.4 (Ubuntu) PHP/5.2.3-1ubuntu6.3 mod_ssl/2.2.4 OpenSSL/0.9.8e Server at www.example.com Port 80

This is despite the fact that the back end is indeed up and running,
and listening on the appropriate port. After waiting for a few minutes
(or on a restart of Apache), the site is accessible again. The error
log says:

[Tue May 27 16:52:30 2008] [error] (111)Connection refused: proxy:
HTTP: attempt to connect to 127.0.0.1:4400 (localhost) failed
[Tue May 27 16:52:30 2008] [error] ap_proxy_connect_backend disabling
worker for (localhost)
[Tue May 27 16:53:06 2008] [error] proxy: HTTP: disabled connection
for (localhost)
[Tue May 27 16:53:08 2008] [error] proxy: HTTP: disabled connection
for (localhost)
[Tue May 27 16:53:18 2008] [error] proxy: HTTP: disabled connection
for (localhost)

It seems like mod_proxy is detecting that the back end is down and
disabling access to it, but also caching that information somewhere.
Thus when the back end comes back up, mod_proxy still claims it's down
and won't pass traffic through.

Any ideas on what might be causing this, and what I can do to fix it
(other than revert to CentOS, which is looking remarkably tempting at
this point, although sadly not very feasible)?

Tet

-- 
Perl is like vise grips. You can do anything with it but it is the
wrong tool for every job. -- Bruce Eckel
-- 
Gllug mailing list  -  Gllug at gllug.org.uk
http://lists.gllug.org.uk/mailman/listinfo/gllug