You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Lionel Farbos <li...@free.fr> on 2005/04/20 18:41:20 UTC
Apache_MaxClients, AJP_maxThreads and jkWorker.cachesize
Hi,
I have 1 Linux Virtual Server (with round robin) + 2 apache 1.3 servers + 3 Tomcat servers load-balanced with mod_jk1.2.10.
1) If each Apache define : MaxClients=150
1.1) what must be the value for each Tomcat Connector AJP maxThread ?
1.2) is the value for each worker.cachesize have to be put ?
2) for my 2 apache 1.3 servers, I must have only 1 JkShmFile on an nfs mount RW ? Right ?
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Apache_MaxClients, AJP_maxThreads and jkWorker.cachesize
Posted by Lionel Farbos <li...@free.fr>.
On Wed, 20 Apr 2005 19:12:17 +0200
Mladen Turk <ml...@jboss.com> wrote:
> Lionel Farbos wrote:
> > Thank you again.
> >
> >
> > So it means than I can have 1 local JkShmFile on my web1 and other local JkShmFile on my web2.
> >
> > But, if, in web1_jkStatus, I disable 1 worker from my cluster (loadbalancer),
> > I also have to disable it from my web2_jkStatus ? Right ?
> >
>
> Well, you can try to use the single shm file.
> It won't hurt trying ;)
> Frankly never tested that on NSF and with multiple httpd's.
> So, tell us if it works.
Sorry Mladen, but when I put 1 shm file on a nfs mount, it doesn't work :-(
So, I use 1 local shm file/web
but I use a script to invoke, in one time, the disable or enable of worker on each web_jkStatus.
Regards.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Apache_MaxClients, AJP_maxThreads and jkWorker.cachesize
Posted by Mladen Turk <ml...@jboss.com>.
Lionel Farbos wrote:
> Thank you again.
>
>
> So it means than I can have 1 local JkShmFile on my web1 and other local JkShmFile on my web2.
>
> But, if, in web1_jkStatus, I disable 1 worker from my cluster (loadbalancer),
> I also have to disable it from my web2_jkStatus ? Right ?
>
Well, you can try to use the single shm file.
It won't hurt trying ;)
Frankly never tested that on NSF and with multiple httpd's.
So, tell us if it works.
Regards,
Mladen.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Apache_MaxClients, AJP_maxThreads and jkWorker.cachesize
Posted by Lionel Farbos <li...@free.fr>.
Thank you again.
On Wed, 20 Apr 2005 18:56:06 +0200
Mladen Turk <ml...@jboss.com> wrote:
> Each apache instance needs it's own shared memory.
>
So it means than I can have 1 local JkShmFile on my web1 and other local JkShmFile on my web2.
But, if, in web1_jkStatus, I disable 1 worker from my cluster (loadbalancer),
I also have to disable it from my web2_jkStatus ? Right ?
>
> Lionel Farbos wrote:
> > Hi,
> >
> > I have 1 Linux Virtual Server (with round robin) + 2 apache 1.3 servers + 3 Tomcat servers load-balanced with mod_jk1.2.10.
> >
> > 2) for my 2 apache 1.3 servers, I must have only 1 JkShmFile on an nfs mount RW ? Right ?
> >
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: Apache_MaxClients, AJP_maxThreads and jkWorker.cachesize
Posted by Mladen Turk <ml...@jboss.com>.
cachesize is used only for threaded apache's like
1.3 on windows or netware; apache2 with worker_mpm, etc...
The size must reflect the number of threads per child
process that for prefork mpms like 1.3 on unixes is
always 1, so there is no need to set it.
Each apache instance needs it's own shared memory.
2x150=300/3 in ideal case, so each TC should have 100
maxThreads.
If you expect that one of the Tomcats in the cluster can fail,
then the number should be 300/2 or 150.
Regards,
Mladen.
Lionel Farbos wrote:
> Hi,
>
> I have 1 Linux Virtual Server (with round robin) + 2 apache 1.3 servers + 3 Tomcat servers load-balanced with mod_jk1.2.10.
>
> 1) If each Apache define : MaxClients=150
> 1.1) what must be the value for each Tomcat Connector AJP maxThread ?
> 1.2) is the value for each worker.cachesize have to be put ?
>
> 2) for my 2 apache 1.3 servers, I must have only 1 JkShmFile on an nfs mount RW ? Right ?
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
>
>
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org