You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Rob Kirkbride <ro...@gmail.com> on 2007/03/20 09:36:03 UTC

mod_jk "increase the backend idle connection timeout or the connection_pool_minsize"

Hi,

I've recently taken over support of a system using mod_jk. This is a big
website utilising 2 sites, 2 web & 3 app servers per site. each web server
talking to 3 app servers
The site takes millions of hits per day.

I'm getting the above info message regularly (twice a minute) even in the
night where the site is relatively lightly loaded. The full report is :-
[Tue Mar 20 08:27:46 2007][24590:14016] [info]
ajp_send_request::jk_ajp_common.c (1215): (app103) error sending request.
Will try another pooled connection
[Tue Mar 20 08:27:46 2007][24590:14016] [info]
ajp_send_request::jk_ajp_common.c (1241): (app103) all endpoints are
disconnected
[Tue Mar 20 08:27:46 2007][24590:14016] [info]
ajp_send_request::jk_ajp_common.c (1244): (app103) increase the backend idle
connection timeout or the connection_pool_minsize
[Tue Mar 20 08:27:46 2007][24590:14016] [info]  ajp_service::jk_ajp_common.c
(1930): (app103) sending request to tomcat failed,  recoverable operation
attempt=1

My workers.properties config is (per web server)
worker.app101.lbfactor=1
worker.app101.socket_timeout=6
worker.app101.socket_keepalive=true

My JBoss server.xml connector has got
   <Connector port="8009" address="${jboss.bind.address}"
         emptySessionPath="true" enableLookups="false" redirectPort="8443"
         maxThreads="1000" acceptCount="200"
         protocol="AJP/1.3"/>

Exact config :- apache 2.0.52 (stock RHEL4), mod_jk 1.2.21 (built from
source), JBoss 4.0.3

I note that the connection_pool_minsize should not be increased on Apache
2.0. I also understand the backend idle connection timeout defaults to 60
seconds so it does seem that increasing it doesn't sound sensible.

I can't say the above error is definitely causing problems though we're
getting occasional reports of long delays.

Any help would be much appreciated as I'm under a bit of pressure to try and
sort.


Thanks

Rob