Mailing-List: contact dev-help@httpd.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@httpd.apache.org
Received-SPF: pass (asf.osuosl.org: domain of bjorn@exoweb.net designates
 221.122.43.98 as permitted sender)
Mime-Version: 1.0 (Apple Message framework v749.3)
Content-Transfer-Encoding: 7bit
Message-Id: <54938E40-EE83-4EF0-947A-D43BBD7B6EB8@exoweb.net>
Content-Type: text/plain; charset=US-ASCII; delsp=yes; format=flowed
To: dev@httpd.apache.org
From: =?ISO-8859-1?Q?Bj=F8rn_Stabell?= <bjorn@exoweb.net>
Subject: mod_proxy_balancer questions
Date: Sat, 6 May 2006 20:56:51 +0800

I've been playing replacing pound http://www.apsis.ch/pound/ with  
mod_proxy_balancer (mpb), and I've run into a few issues that I  
thought I'd share here, maybe confirm if I understood things correctly.

1) mpb logs when backend servers can't be contacted, but not when  
they are available again
It's important to be able to see when a backend server dies and when  
it's available again.
I was hoping for something that could be easily matched against, e.g.:
	backend B alive; taking online [info]
	backend B died; taking offline and moving clients B2 [error]
	backend B still dead [warning] -- regular check
	backend B has come alive again, taking online [error] -- error recovery

2) mpb doesn't warn/ log when switching backend server when  
stickysession=on
It's important to know when this happens as it'll probably mean some  
user lost his session.
Something like:
	backend B died; taking offline and moving clients to B2; waited S  
seconds because stickysessions was set [error] -- maybe even more  
serious?

I've been running with loglevel=debug, so I hope I haven't missed  
messages like the above amongst all the other messages.

Rgds,
Bjorn