Rainer, May be my problem is due to the fact I have the line :
#worker.izonetv.sticky_session_force=1 commented. The reason is that if I uncomment this line I have this error in Apache : [Thu Mar 12 17:31:37 2009][23548:6496] [error] service::jk_lb_worker.c (1144): All tomcat instances failed, no more workers left for recovery Any idea what can be the problem ? Toni. 2009/3/11 Toni Menendez Lopez <tonime...@gmail.com> > > > 2009/3/11 Rainer Jung <rainer.j...@kippdata.de> > >> On 11.03.2009 14:03, Toni Menendez Lopez wrote: >> >>> Hello everybody, >>> I have following architecture : >>> 2 Server with Apache and Tomcat >>> Versions : >>> APACHE >>> ------------- >>> httpd -v >>> Server version: Apache/2.0.52 >>> Server built: May 24 2006 11:45:06 >>> TOMCAT >>> ------------- >>> ./version.sh >>> Using JRE_HOME: /opt/jdk1.5.0_10 >>> Server version: Apache Tomcat/5.5.20 >>> Server built: Sep 12 2006 10:09:20 >>> Server number: 5.5.20.0 >>> OS Name: Linux >>> OS Version: 2.6.9-55.0.2.ELsmp >>> Architecture: i386 >>> JVM Version: 1.5.0_10-b03 >>> JVM Vendor: Sun Microsystems Inc. >>> >> >> What's your mod_jk version? > > > -->JK Version: mod_jk/1.2.23 > >> >> >> I am doing load balancing between both with JK, with an scenario of 50 >>> reques per session (aprox), and 500 reqxseg ( aprox ). >>> >> >> What is reqxseg? > > > --> HTTP request per second. > >> >> >> The thing is the following, when I shutdown the passive server, I have a >>> >> >> What is a "passive" server? I thought you do load balancing? >> What do you mean by "shutdown"? >> > > --> I have a red hat cluster in both servers which give to me an Virtual IP > for both servers, this virtual IP is of one of the servers and is the > entrace point for my architecture, so the requests are received only with > one apache, which delivers the request to the tomcats of both servers. > > --> The passive servers in the one which does not have the IP, si the > apache of this servers is just not receiving requests. > > --> Shutting down, is just swtch off the server ( just simulating a crash > of the server). > > download of my reqxseg, and the requests that where managed by the >>> passive server get stucked for long time, like 5 min. >>> New request afther shutting donw the passive server are well processed. >>> But the thing is if there is a way to reduce this time that requests are >>> stucked. >>> >> >> The answer depends on your answer to m<y above questions. >> >> This is my worker : >>> # izonetv LoadBalancer Definition >>> >>> worker.izonetv.balance_workers=izonetv-mifeas01_data,izonetv-mifeas02_data >>> worker.izonetv.method=Session >>> worker.izonetv.retries=1 >>> worker.izonetv.sticky_session=True >>> #worker.izonetv.sticky_session_force=1 >>> worker.izonetv.type=lb >>> # izonetv-mifeas01_data Node Definition >>> worker.izonetv-mifeas01_data.connect_timeout=10000 >>> worker.izonetv-mifeas01_data.fail_on_status=404 >>> >> >> You don't want to fail on 404. > > --> Sorry what does mean this parameter ? I just copy from another > configuration. > >> >> >> worker.izonetv-mifeas01_data.host=mifeas01_data >>> worker.izonetv-mifeas01_data.lbfactor=1 >>> worker.izonetv-mifeas01_data.port=8009 >>> worker.izonetv-mifeas01_data.reply_timeout=30000 >>> worker.izonetv-mifeas01_data.type=ajp13 >>> # izonetv-mifeas02_data Node Definition >>> worker.izonetv-mifeas02_data.connect_timeout=10000 >>> worker.izonetv-mifeas02_data.fail_on_status=404 >>> >> >> You don't want to fail on 404. >> >> worker.izonetv-mifeas02_data.host=mifeas02_data >>> worker.izonetv-mifeas02_data.lbfactor=1 >>> worker.izonetv-mifeas02_data.port=8009 >>> worker.izonetv-mifeas02_data.reply_timeout=30000 >>> worker.izonetv-mifeas02_data.type=ajp13 >>> >> >> Regards, >> >> Rainer >> >> --------------------------------------------------------------------- >> To unsubscribe, e-mail: users-unsubscr...@tomcat.apache.org >> For additional commands, e-mail: users-h...@tomcat.apache.org >> >> >