Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@httpd.apache.org
Received-SPF: softfail (asf.osuosl.org: transitioning domain of
 a.d.stribblehill@durham.ac.uk does not designate 129.234.2.142 as permitted
 sender)
Date: Mon, 24 Jul 2006 10:13:34 +0100
From: Andrew Stribblehill <a.d.stribblehill@durham.ac.uk>
To: dev@httpd.apache.org
Subject: Balancer manager: bug and feature request
Message-ID: <20060724091334.GA10879@wompom>
Mail-Followup-To: dev@httpd.apache.org
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.5.11+cvs20060403
Sender: Andrew Stribblehill <a.d.stribblehill@durham.ac.uk>

That is, "bug" and "feature request", not "bug and feature" request.

As I reported in
http://issues.apache.org/bugzilla/show_bug.cgi?id=39907
I'm finding that with the worker MPM and a pristine httpd 2.2.2 on an
i386 Debian platform, changes I make to the configuration within
balancer-manager will periodically reset -- and it's faster under
load.

I can see the code where the config is being set, but I don't quite
know whether this is process-local or not. If so, this would explain
the symptoms.

If someone could point me in the right direction, I'll have a bash at
fixing this bug.


The feature request: sometimes we need to take down an origin server
on a planned basis. The app is session-based so we can't just
disable it completely. Is there any reason I shouldn't be able to set
the load factor for an origin server to 0? At present it's
constrained to the range 1..100 inclusive. If I had some form of
session replication, I could use route-redirect and then just disable
it, but this is not possible for me.

Thanks

-- 
FORTIES CROMARTY FORTH TYNE DOGGER
VARIABLE, MAINLY SOUTH OR SOUTHEAST 3 OR 4. MAINLY FAIR. MODERATE OR
GOOD, BUT FOG PATCHES AT FIRST IN FORTIES AND DOGGER