[JBoss-dev] httpd and tomcatClustering gives 404 during redeploy.

Sacha Labourey sacha.labourey.ml at jboss.org
Thu Sep 21 08:37:13 EDT 2006


That is what I mean: the MainDeployer could "share" this knowledge that 
it is a redeploy with any other piece of code, including the 
tomcatdeployer, by setting a temporary thread local variable that could 
be checked by any deployer.

Jean-frederic Clere wrote:
> On Thu, 2006-09-21 at 09:36 +0200, Sacha Labourey wrote:
>   
>> What about having the MainDeployer set a thread context variable 
>> indicating whether we are inside a "Redeploy" scenario?
>>     
>
> MainDeployer knows that it is doing a redeploy(), the problem is just
> that the redeploy() should also be in the TomcatDeployer or that
> MainDeployer should tell to the TomcatDeployer that the undeploy() is
> called for a redeploy() (like undeploy(false) for a redeployement and
> undeploy(true) for an undeployement).
>
> Cheers
>
> Jean-Frederic
>
>   
>> Jean-frederic Clere wrote:
>>     
>>> On Wed, 2006-09-20 at 13:48 -0700, Scott M Stark wrote:
>>>   
>>>       
>>>> The TomcatDeployer knows that an undeploy is being done 
>>>> (performUndeploy) and can configure tomcat to return 503 until the 
>>>> completion of the subsequent deploy(performDeploy).
>>>>     
>>>>         
>>> My first goal is not to handle it in Tomcat but in httpd for this a
>>> ContainerListener in Tomcat would be enough. The only thing is to add a
>>> parameter to tell how long an application could stay undeployed
>>> returning 503 before returning 404.
>>>
>>> Cheers
>>>
>>> Jean-Frederic
>>>
>>>   
>>>       
>>>> Jean-frederic Clere wrote:
>>>>     
>>>>         
>>>>> Scott M Stark wrote:
>>>>>
>>>>>       
>>>>>           
>>>>>> The AbstractWebDeployer/TomcatDeployer can inform tomcat that an 
>>>>>> undeploy is begining. Are you looking for a distinction between an 
>>>>>> unavailable app and one that is being redeployed?
>>>>>>         
>>>>>>             
>>>>> Yes. I know that "there is no guarantee that the redeploy will 
>>>>> succeed" but in production that should not happend and even in this 
>>>>> case the customer probably prefer a 503 to a 404. (If the request 
>>>>> cannot be routed to another node).
>>>>>
>>>>>       
>>>>>           
>>>>>> I would think not as there is no guarantee that the redeploy will 
>>>>>> succeed. I don't see that the MainDeployer needs to change, just the 
>>>>>> TomcatDeployer.
>>>>>>         
>>>>>>             
>>>>> The actual TomcatDeployer doesn't know that a redeploy is taking place 
>>>>> how could it inform Tomcat?
>>>>>
>>>>> Cheers
>>>>>
>>>>> Jean-Frederic
>>>>>       
>>>>>           
>>>> _______________________________________________
>>>> jboss-development mailing list
>>>> jboss-development at lists.jboss.org
>>>> https://lists.jboss.org/mailman/listinfo/jboss-development
>>>>     
>>>>         
>>> _______________________________________________
>>> jboss-development mailing list
>>> jboss-development at lists.jboss.org
>>> https://lists.jboss.org/mailman/listinfo/jboss-development
>>>
>>>   
>>>       
>> _______________________________________________
>> jboss-development mailing list
>> jboss-development at lists.jboss.org
>> https://lists.jboss.org/mailman/listinfo/jboss-development
>>     
>
> _______________________________________________
> jboss-development mailing list
> jboss-development at lists.jboss.org
> https://lists.jboss.org/mailman/listinfo/jboss-development
>
>   




More information about the jboss-development mailing list