tomcat-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Bernd Koecke>
Subject Re: [PATCH] jakarta-tomcat-connectors/jk/native/common/jk_lb_worker.c,jk_util.c
Date Fri, 05 Apr 2002 12:18:57 GMT
GOMEZ Henri wrote:
> Seems a good patch.
> Should be commited.
> BTW, Bernd could you make a little documentation which could
> be included in jk doc and add example ?
> Excellent works ;)

Thanks, I'll have a look at the docs in detail and add some example. Should I 
sent these files to this list?


> -
> Henri Gomez                 ___[_]____
> EMAIL :        (. .)                     
> PGP KEY : 697ECEDD    ...oOOo..(_)..oOOo...
> PGP Fingerprint : 9DF8 1EA8 ED53 2F39 DC9B 904A 364F 80E6 
>>-----Original Message-----
>>From: Bernd Koecke []
>>Sent: Friday, April 05, 2002 1:30 PM
>>Subject: [PATCH]
>>some days ago I sent a question for an only routing lb-worker. 
>>Costin asked for 
>>the patch and here it is.
>>I added two config-directives for the lb-worker:
>>- this is the name of the worker which should get the request 
>>if there is no 
>>session or no jvmRoute.
>>- the possible values are 'balance' and 'reject'. Explanation follows.
>>Behaviour of lb-worker:
>>a.) there is a jvmRoute and the corresponding node is ok:
>>    -> The request goes to the corresponding node
>>b.) there is no jvmRoute or the corresponding node didn't answered:
>>    -> if a local_worker was given. this one will get the request.
>>    -> if a local_worker was given, but it is in error state:
>>    fault_action == reject: the request gets an error.
>>    fault_action == balance: try to find another worker from 
>>the list of
>>                             balanced workers (c.).
>>    -> if no local_worker was given, go directly to c.).
>>c.) if no worker was found in a) and b) the normal balancing 
>>behaviour takes place.
>>I looked at cvs-repopsitory and it seems that jk_lb_worker.c 
>>and jk_util.c have 
>>the same version (1.8, 1.12) like mine from 
>>I need the fault_action, because our load balancer should not 
>>route a request to 
>>a node without a working local tomcat. So it is an error if an 
>>request arrives at a node without a working local tomcat. And 
>>it is also an 
>>error to route it to one of the other nodes, because sometimes 
>>we switch off one 
>>node only by telling the load balancer not to use it for 
>>requests. In this case 
>>the modules should use this node only for requests with a 
>>session on it and not 
>>for requests without a session. But for other use cases it 
>>might be useful if 
>>mod_jk tries to balance the request in case of a failure of 
>>the local tomcat.
>>I hope it is useful. Sorry i haven't looked into jk2 at this 
>>time, but may be i 
>>get time to do it soon.
>>Dipl.-Inform. Bernd Koecke
>>Schlund+Partner AG
>>Fon: +49-721-91374-0
> --
> To unsubscribe, e-mail:   <>
> For additional commands, e-mail: <>

Dipl.-Inform. Bernd Koecke
Schlund+Partner AG
Fon: +49-721-91374-0

To unsubscribe, e-mail:   <>
For additional commands, e-mail: <>

View raw message