tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From André Warnier ...@ice-sa.com>
Subject Re: JK 1.2.28 - load balancer worker fails on startup with one worker down ?
Date Wed, 15 Apr 2009 22:20:08 GMT
If we just stick to the actual error message for a moment, and assume it 
means what it says :
 >> worker mygpgby06 can't resolve tomcat address mygpgby06.mycompany.com
the first question would be : why can the DNS name 
"mygpgby06.mycompany.com" not be /resolved/ to an IP address when host 
"mygpgby06" (?) is down for maintenance ?

Is there some kind of dynamic DNS system at work there ?

What happens if you replace
 >> worker.mygpgby06.host=mygpgby06.mycompany.com
by
 >> worker.mygpgby06.host=xxx.xxx.xxx.xxx
(xxx.xxx.xxx.xxx being the actual IP address of that host)

I'm just guessing here, but what if it is so that, at least at start, 
the load balancing members must at least be able to be resolved to an IP 
address, otherwise mod_jk determines that there's really something wrong 
with the configuration, and won't even start ?




Scott Bradshaw wrote:
> /portal/*=loadbalancerprod
> 
> The uriworkermap.properties file is correct - workers are correctly sent to
> it assuming all the workers are accessible.
> 
> The problem is when the workers in the load balancer are being initialized,
> if one worker is not available, the load balance worker is considered not
> valid. Because its not valid, requests will not be sent to it. This does not
> seem to be the desired behavior of a load balancer.
> 
> Scott
> 
> On Wed, Apr 15, 2009 at 5:32 PM, Jorge Medina <jmedina@e-dialog.com> wrote:
> 
>> Your workers.properties looks fine.
>>
>> What is the content of uriworkermap.proeprties ?
>>
>> -----Original Message-----
>> From: swbradshaw@gmail.com [mailto:swbradshaw@gmail.com] On Behalf Of
>> Scott Bradshaw
>> Sent: Wednesday, April 15, 2009 3:51 PM
>> To: users@tomcat.apache.org
>> Subject: JK 1.2.28 - load balancer worker fails on startup with one
>> worker down ?
>>
>> I apologize if this a silly question, but I can't figure it out! I've
>> looked over the documentation and I'm stumped.
>>
>> I have 5 load balanced workers defined. I have them setup and configured
>> correctly.
>>
>> workers.properties file (partial - not including all the individual
>> workers)
>> ---------------------------------------------------------------
>> worker.mygpgby02.type=ajp13
>> worker.mygpgby02.host=mygpgby02.mycompany.com
>> worker.mygpgby02.port=8009
>>
>> worker.loadbalancerprod.type=lb
>> worker.loadbalancerprod.balance_workers=mygpgby02,mygpgby03,mygpgby04,my
>> gpgby05,mygpgby06
>> worker.list=loadbalancerprod
>> ---------------------------------------------------------------
>> Now, here is the problem - one of those hosts(mygpgby06) is currently
>> down for maintenence. Whenever I startup IIS, the ISAPI proxy won't
>> work. The ISAPI log file shows this:  (please note mycompany.com is not
>> the actual url
>> - i changed it)
>>
>> [Wed Apr 15 14:22:00.463 2009] [4208:2848] [error] jk_ajp_common.c
>> (2526):
>> worker mygpgby06 can't resolve tomcat address mygpgby06.mycompany.com
>> [Wed Apr 15 14:22:00.463 2009] [4208:2848] [error] jk_worker.c (163):
>> validate failed for mygpgby06
>> [Wed Apr 15 14:22:00.463 2009] [4208:2848] [error] jk_lb_worker.c
>> (1599):
>> Failed creating worker mygpgby06
>> [Wed Apr 15 14:22:00.479 2009] [4208:2848] [error] jk_lb_worker.c
>> (1647):
>> NULL parameters
>> [Wed Apr 15 14:22:00.479 2009] [4208:2848] [error] jk_worker.c (163):
>> validate failed for loadbalancerprod
>> [Wed Apr 15 14:22:00.479 2009] [4208:2848] [error] jk_worker.c (262):
>> failed to create worker loadbalancerprod [Wed Apr 15 14:22:00.479 2009]
>> [4208:2848] [error] jk_uri_worker_map.c
>> (506): Could not find worker with name 'loadbalancerprod' in uri map
>> post processing.
>>
>>  If I take this worker out of the balance_workers list, everything
>> starts up fine. If I leave it in, my loadbalancerprod worker is
>> completely dead..
>>
>>  According to the documentation, "When starting up, the web server
>> plugin with instantiate the workers whose name appears in the
>> worker.list property..".
>>
>> So - one worker in the load balancer won't start so the whole load
>> balancer is considered a failed worker.
>>
>> Is there a property I'm missing to make this work ?
>>
>> Scott
>>
>> ---------------------------------------------------------------------
>> To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
>> For additional commands, e-mail: users-help@tomcat.apache.org
>>
>>
> 


---------------------------------------------------------------------
To unsubscribe, e-mail: users-unsubscribe@tomcat.apache.org
For additional commands, e-mail: users-help@tomcat.apache.org


Mime
View raw message