You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@tomcat.apache.org by Austin Gonyou <au...@coremetrics.com> on 2003/04/29 22:43:12 UTC
mod_jk stickyness....take two!
Recently I asked(pleaded?) for some information regarding mod_jk's
stickyness. In the context of that answer it seems that only Apache 2.0
and the worker MPM are capable of mod_jk stickyness.
I recently saw this in the CHANGES.txt for mod_jk and would like someone
to let me know if this applies to Apache 2.0 still, or is this global?
(i.e. prefork and worker MPMs on Apache 2.x, and works on Apache 1.x
also)
<cvslog>
Revision 1.9 / (view) - annotate - [select for diffs] , Thu Jan 2
17:12:55 2003 UTC (3 months, 3 weeks ago) by glenn
Branch: MAIN
CVS Tags: coyote_10, TOMCAT_4_1_24, TOMCAT_4_1_23, TOMCAT_4_1_22,
TOMCAT_4_1_21, TOMCAT_4_1_20, TOMCAT_4_1_19
Changes since 1.8: +4 -1 lines
Diff to previous 1.8 (colored)
Add new lb property sticky_session. If set to 0, requests with session
id's do not have to be routed back to the same Tomcat worker. This is
to support the new Tomcat Session Manager code which supports persistance
of session data across multiple Tomcat instances, such as JavaGroup's.
</cvslog>
If anyone could answer this pretty quickly, it'd help me know the path I
must take for our migration. Thanks.
--
Austin Gonyou <au...@coremetrics.com>
Coremetrics, Inc.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
Re: mod_jk stickyness....take two!
Posted by Charles So <ch...@mac.com>.
Hi Bill, I am investigating how to do true-balancing in tomcat/apache 1.
Here is what I used to do with IBM Websphere app server in this respect:
System config:
1) 2 s/w load balancer (2 machines, only 1 is active; the other standby)
2) 2 web servers/2 app servers (2 machines)
3) 2 db server (2 machines)
The load balancer calls a java program on both web/app servers and the
program calculates on-the-fly the loading of the system. The load
balancer (LB) would send the HTTP request to the less-loaded system.
The app server would then retrieves the session from the DB and perform
its work
I come across the persistent mgr in Tomcat which suppose to be able to
store sessions in files or via JDBC. It is still in beta stage.
I am wondering how you define true-loadbalancing.. can you share
information with me?
I hope to have similar config running under TC4, MySQL, Apache 1. Any
opinions?
On Thursday, May 1, 2003, at 01:20 PM, Bill Barker wrote:
> Sticky works (at least AFAIK). True load-balancing only works with
> Apache2
> under the "worker" MPM. This isn't really much of a change from 3.3.x:
> Tomcat 4.x should work at least as well as 3.3.x. They both have the
> same
> issues.
>
> "Austin Gonyou" <au...@coremetrics.com> wrote in message
> news:1051751553.3384.12.camel@portageek.digitalroadkill.net...
>> We just now got into tomcat4.x, and we're coming from 3.3.x. An
>> earlier
>> posting I asked about this I got a response which seemed like the
>> answer
>> was mod_jk with Apache2 was the answer. This makes a bit better sense.
>>
>> Just to clarify, from what you're saying, apache 1.x and 2.x will work
>> with the latest mod_jk to do sticky load-balancing. Right? TIA.
>>
>>
>>
>> On Tue, 2003-04-29 at 16:05, Filip Hanik wrote:
>>> I thought that the (server.xml) property jvmRoute did that
>>>
>>> Filip
>>>
>>>> -----Original Message-----
>>>> From: Austin Gonyou [mailto:austin@coremetrics.com]
>>>> Sent: Tuesday, April 29, 2003 1:43 PM
>>>> To: Tomcat Users List
>>>> Subject: mod_jk stickyness....take two!
>>>>
>>>>
>>>> Recently I asked(pleaded?) for some information regarding mod_jk's
>>>> stickyness. In the context of that answer it seems that only Apache
> 2.0
>>>> and the worker MPM are capable of mod_jk stickyness.
>>>>
>>>> I recently saw this in the CHANGES.txt for mod_jk and would like
> someone
>>>> to let me know if this applies to Apache 2.0 still, or is this
>>>> global?
>>>> (i.e. prefork and worker MPMs on Apache 2.x, and works on Apache 1.x
>>>> also)
>>>>
>>>> <cvslog>
>>>> Revision 1.9 / (view) - annotate - [select for diffs] , Thu Jan 2
>>>> 17:12:55 2003 UTC (3 months, 3 weeks ago) by glenn
>>>> Branch: MAIN
>>>> CVS Tags: coyote_10, TOMCAT_4_1_24, TOMCAT_4_1_23, TOMCAT_4_1_22,
>>>> TOMCAT_4_1_21, TOMCAT_4_1_20, TOMCAT_4_1_19
>>>> Changes since 1.8: +4 -1 lines
>>>> Diff to previous 1.8 (colored)
>>>> Add new lb property sticky_session. If set to 0, requests with
> session
>>>> id's do not have to be routed back to the same Tomcat worker. This
>>>> is
>>>> to support the new Tomcat Session Manager code which supports
> persistance
>>>> of session data across multiple Tomcat instances, such as
>>>> JavaGroup's.
>>>> </cvslog>
>>>>
>>>> If anyone could answer this pretty quickly, it'd help me know the
>>>> path
> I
>>>> must take for our migration. Thanks.
>>>>
>>>> --
>>>> Austin Gonyou <au...@coremetrics.com>
>>>> Coremetrics, Inc.
>>>>
>>>> --------------------------------------------------------------------
>>>> -
>>>> 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
>> --
>> Austin Gonyou <au...@coremetrics.com>
>> Coremetrics, Inc.
>
>
>
>
> ---------------------------------------------------------------------
> 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
Re: mod_jk stickyness....take two!
Posted by Bill Barker <wb...@wilshire.com>.
Sticky works (at least AFAIK). True load-balancing only works with Apache2
under the "worker" MPM. This isn't really much of a change from 3.3.x:
Tomcat 4.x should work at least as well as 3.3.x. They both have the same
issues.
"Austin Gonyou" <au...@coremetrics.com> wrote in message
news:1051751553.3384.12.camel@portageek.digitalroadkill.net...
> We just now got into tomcat4.x, and we're coming from 3.3.x. An earlier
> posting I asked about this I got a response which seemed like the answer
> was mod_jk with Apache2 was the answer. This makes a bit better sense.
>
> Just to clarify, from what you're saying, apache 1.x and 2.x will work
> with the latest mod_jk to do sticky load-balancing. Right? TIA.
>
>
>
> On Tue, 2003-04-29 at 16:05, Filip Hanik wrote:
> > I thought that the (server.xml) property jvmRoute did that
> >
> > Filip
> >
> > > -----Original Message-----
> > > From: Austin Gonyou [mailto:austin@coremetrics.com]
> > > Sent: Tuesday, April 29, 2003 1:43 PM
> > > To: Tomcat Users List
> > > Subject: mod_jk stickyness....take two!
> > >
> > >
> > > Recently I asked(pleaded?) for some information regarding mod_jk's
> > > stickyness. In the context of that answer it seems that only Apache
2.0
> > > and the worker MPM are capable of mod_jk stickyness.
> > >
> > > I recently saw this in the CHANGES.txt for mod_jk and would like
someone
> > > to let me know if this applies to Apache 2.0 still, or is this global?
> > > (i.e. prefork and worker MPMs on Apache 2.x, and works on Apache 1.x
> > > also)
> > >
> > > <cvslog>
> > > Revision 1.9 / (view) - annotate - [select for diffs] , Thu Jan 2
> > > 17:12:55 2003 UTC (3 months, 3 weeks ago) by glenn
> > > Branch: MAIN
> > > CVS Tags: coyote_10, TOMCAT_4_1_24, TOMCAT_4_1_23, TOMCAT_4_1_22,
> > > TOMCAT_4_1_21, TOMCAT_4_1_20, TOMCAT_4_1_19
> > > Changes since 1.8: +4 -1 lines
> > > Diff to previous 1.8 (colored)
> > > Add new lb property sticky_session. If set to 0, requests with
session
> > > id's do not have to be routed back to the same Tomcat worker. This is
> > > to support the new Tomcat Session Manager code which supports
persistance
> > > of session data across multiple Tomcat instances, such as JavaGroup's.
> > > </cvslog>
> > >
> > > If anyone could answer this pretty quickly, it'd help me know the path
I
> > > must take for our migration. Thanks.
> > >
> > > --
> > > Austin Gonyou <au...@coremetrics.com>
> > > Coremetrics, Inc.
> > >
> > > ---------------------------------------------------------------------
> > > 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
> --
> Austin Gonyou <au...@coremetrics.com>
> Coremetrics, Inc.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
RE: mod_jk stickyness....take two!
Posted by Austin Gonyou <au...@coremetrics.com>.
We just now got into tomcat4.x, and we're coming from 3.3.x. An earlier
posting I asked about this I got a response which seemed like the answer
was mod_jk with Apache2 was the answer. This makes a bit better sense.
Just to clarify, from what you're saying, apache 1.x and 2.x will work
with the latest mod_jk to do sticky load-balancing. Right? TIA.
On Tue, 2003-04-29 at 16:05, Filip Hanik wrote:
> I thought that the (server.xml) property jvmRoute did that
>
> Filip
>
> > -----Original Message-----
> > From: Austin Gonyou [mailto:austin@coremetrics.com]
> > Sent: Tuesday, April 29, 2003 1:43 PM
> > To: Tomcat Users List
> > Subject: mod_jk stickyness....take two!
> >
> >
> > Recently I asked(pleaded?) for some information regarding mod_jk's
> > stickyness. In the context of that answer it seems that only Apache 2.0
> > and the worker MPM are capable of mod_jk stickyness.
> >
> > I recently saw this in the CHANGES.txt for mod_jk and would like someone
> > to let me know if this applies to Apache 2.0 still, or is this global?
> > (i.e. prefork and worker MPMs on Apache 2.x, and works on Apache 1.x
> > also)
> >
> > <cvslog>
> > Revision 1.9 / (view) - annotate - [select for diffs] , Thu Jan 2
> > 17:12:55 2003 UTC (3 months, 3 weeks ago) by glenn
> > Branch: MAIN
> > CVS Tags: coyote_10, TOMCAT_4_1_24, TOMCAT_4_1_23, TOMCAT_4_1_22,
> > TOMCAT_4_1_21, TOMCAT_4_1_20, TOMCAT_4_1_19
> > Changes since 1.8: +4 -1 lines
> > Diff to previous 1.8 (colored)
> > Add new lb property sticky_session. If set to 0, requests with session
> > id's do not have to be routed back to the same Tomcat worker. This is
> > to support the new Tomcat Session Manager code which supports persistance
> > of session data across multiple Tomcat instances, such as JavaGroup's.
> > </cvslog>
> >
> > If anyone could answer this pretty quickly, it'd help me know the path I
> > must take for our migration. Thanks.
> >
> > --
> > Austin Gonyou <au...@coremetrics.com>
> > Coremetrics, Inc.
> >
> > ---------------------------------------------------------------------
> > 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
--
Austin Gonyou <au...@coremetrics.com>
Coremetrics, Inc.
---------------------------------------------------------------------
To unsubscribe, e-mail: tomcat-user-unsubscribe@jakarta.apache.org
For additional commands, e-mail: tomcat-user-help@jakarta.apache.org
RE: mod_jk stickyness....take two!
Posted by Filip Hanik <ma...@filip.net>.
I thought that the (server.xml) property jvmRoute did that
Filip
> -----Original Message-----
> From: Austin Gonyou [mailto:austin@coremetrics.com]
> Sent: Tuesday, April 29, 2003 1:43 PM
> To: Tomcat Users List
> Subject: mod_jk stickyness....take two!
>
>
> Recently I asked(pleaded?) for some information regarding mod_jk's
> stickyness. In the context of that answer it seems that only Apache 2.0
> and the worker MPM are capable of mod_jk stickyness.
>
> I recently saw this in the CHANGES.txt for mod_jk and would like someone
> to let me know if this applies to Apache 2.0 still, or is this global?
> (i.e. prefork and worker MPMs on Apache 2.x, and works on Apache 1.x
> also)
>
> <cvslog>
> Revision 1.9 / (view) - annotate - [select for diffs] , Thu Jan 2
> 17:12:55 2003 UTC (3 months, 3 weeks ago) by glenn
> Branch: MAIN
> CVS Tags: coyote_10, TOMCAT_4_1_24, TOMCAT_4_1_23, TOMCAT_4_1_22,
> TOMCAT_4_1_21, TOMCAT_4_1_20, TOMCAT_4_1_19
> Changes since 1.8: +4 -1 lines
> Diff to previous 1.8 (colored)
> Add new lb property sticky_session. If set to 0, requests with session
> id's do not have to be routed back to the same Tomcat worker. This is
> to support the new Tomcat Session Manager code which supports persistance
> of session data across multiple Tomcat instances, such as JavaGroup's.
> </cvslog>
>
> If anyone could answer this pretty quickly, it'd help me know the path I
> must take for our migration. Thanks.
>
> --
> Austin Gonyou <au...@coremetrics.com>
> Coremetrics, Inc.
>
> ---------------------------------------------------------------------
> 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