Re: tomcat 4.0.3 + mod_jk

GOMEZ Henri wrote: >>Up to now i Worked only on jk from JTC, but if we get the same >>functionality from jk2 I could have a look at it. I thought >>that jk2 was >>for apache2. But I may misunderstound this. > > > jk2 is the reworked jk, for ap 1.3/2.0, IIS/iPlanet to come > ok, for which sh

RE: tomcat 4.0.3 + mod_jk

>Up to now i Worked only on jk from JTC, but if we get the same >functionality from jk2 I could have a look at it. I thought >that jk2 was >for apache2. But I may misunderstound this. jk2 is the reworked jk, for ap 1.3/2.0, IIS/iPlanet to come -- To unsubscribe, e-mail:

Re: tomcat 4.0.3 + mod_jk

GOMEZ Henri wrote: >>Yes. Costin suggested to use a special lb-factor. I would use >>-1 for the >>lb-factor. But this wouldn't solve the problem if someone set >>more than >>one worker to this value. To use a new config tag for >>optionally setting >>a default worker would be a cleaner way.

RE: tomcat 4.0.3 + mod_jk

>Yes. Costin suggested to use a special lb-factor. I would use >-1 for the >lb-factor. But this wouldn't solve the problem if someone set >more than >one worker to this value. To use a new config tag for >optionally setting >a default worker would be a cleaner way. And it would be easier to

Re: tomcat 4.0.3 + mod_jk

GOMEZ Henri wrote: yes, I know. But when I come in without a session the function get_most_suitable_worker in jk_lb_worker.c will find a >>> >>worker by using >> the lb-factor and do weighted round robin. And at this >>> >>position I would >> like to switch off the load-balanc

RE: tomcat 4.0.3 + mod_jk

>My thinking for jk2 was that the session stickiness should go >in front of >everything, and be enabled by default, independent of lb. Agreed. May be we could build a more powerfull system later with a reworker LB supporting failure via JTC cluster. -- To unsubscribe, e-mail:

RE: tomcat 4.0.3 + mod_jk

>>>yes, I know. But when I come in without a session the function >>>get_most_suitable_worker in jk_lb_worker.c will find a >worker by using >>>the lb-factor and do weighted round robin. And at this >position I would >>>like to switch off the load-balancing and select the worker which is >>

Re: tomcat 4.0.3 + mod_jk

ok, I will patch worker_lb. But it will take a few days. To mutch work and to less time :). Bernd [EMAIL PROTECTED] wrote: > This also solves another problem - Apache2.0+JNI worker. It's exactly the > same, it acts similary with a load balancer ( multiple processes listening > on the same por

Re: tomcat 4.0.3 + mod_jk

This also solves another problem - Apache2.0+JNI worker. It's exactly the same, it acts similary with a load balancer ( multiple processes listening on the same port ), and we want the jni worker to take the load if no session. Wouldn't be simpler if you just patch worker_lb ? Say if a worke

Re: tomcat 4.0.3 + mod_jk

yes, I know. But when I come in without a session the function get_most_suitable_worker in jk_lb_worker.c will find a worker by using the lb-factor and do weighted round robin. And at this position I would like to switch off the load-balancing and select the worker which is connected to tomcat

RE: tomcat 4.0.3 + mod_jk

in mod_jk, when a session is created the following requests will allways go to the same tomcat. IBM call it Session Affinity ;) >Hi, > >is it possible to use only the session routing of mod_jk without load >balancing? > >Because we have a standalone loadbalancer in front of our cluster. In >the