tomcat-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Rainer Jung <>
Subject Re: Mod_jk balancing, session problem
Date Wed, 09 Aug 2006 16:48:46 GMT

Artur schrieb:
>> - does it happen in times, when you get errors in the mod_jk-log? Does 
>> it happen at all times, or do you have peaks at special times?
>> I can't figure it out, when this happens, but there isn't any special
>> moment I can observe. 
>> It happens randomly. There isn't any error in mod_jk.log

Choose the suggested access log methods to improve observability (URLs
and Cookie and Set-Cookie Headers.

>> And here I need some help. In mod_jk.log I always see the same worker -
>> balancer worker, so I don't know which worker has been choosen. How can I
>> change it ? I assume by adding request log to mod_jk you mean something
>> like this:
>> # JkRequestLogFormat set the request format
>>   JkRequestLogFormat     "%w %V %T"

OK, unfortunately mod_jk does not yet allow to log the sub worker of an
lb. I'll check for improvement in 1.2.19...

>>> <IfModule mpm_worker_module>
>>>     StartServers         5
>>>     MaxClients           448
>>>     MinSpareThreads      5
>>>     MaxSpareThreads      25 
>>>     ThreadsPerChild      7
>>>     MaxRequestsPerChild   0
>>> </IfModule>
>> Interesting, very few ThreadsPerChild. Not really a problem, but I 
>> wonder if this is efficient.
>> We have decresed this value because we thought that maybe Apache is
>> opening too many threads to Tomcat. We had in Tomcat's log entries saying"
>> "All threads (500) are curently busy, increase maxThreds". 

Threads in Tomcat = Sum of numbers of apache threads used during peak
times. So this relates more to MaxClients and not to ThreadsPerChild.
If you want to connect many Apaches with a lot of threads to tomcat you
need to use tricks like the APR connector on the tomcat side. A few
hundred threads should be no problem, on a good OS even 1.000-2.000
should work.

>> Think about using connect_timeout and prepost_timeout. Maybe also set 
>> recovery_options to 3.
>> What values do you suggest for above parameters ?


This is in Milliseconds. Anything between 2000 and 20000 should be OK.
The right value depends on your trust in latency, reliability of your
network and the responsiveness of the target machine.


Values between 5000 and 30000 are reasonable. Arguments the same as before.

recovery_options=3 or 7 (here you'll need to experimnt on your own. The
added "4" is very new, but maybe helpful).



To start a new topic, e-mail:
To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message