You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@geronimo.apache.org by Juergen Weber <we...@gmail.com> on 2009/07/15 17:34:36 UTC

State of EJB Clustering

Hi,

what is the state of Geronimo's EJB clustering?

Basically I wonder if it is possible to have n Java application clients that
address m servers that run stateless beans. A client should stick to a
server as long it is alive and transparently go to another one if it is
dead.

Is this information here still current for 2.2?
http://cwiki.apache.org/GMOxDEV/test-drive-geronimo-ejb-clustering.html

Or would GGache be the way to go?
http://cwiki.apache.org/GMOxDEV/geronimo-clustering-with-gcache.html

Thanks,
Juergen
-- 
View this message in context: http://www.nabble.com/State-of-EJB-Clustering-tp24500428s134p24500428.html
Sent from the Apache Geronimo - Users mailing list archive at Nabble.com.

Re: State of EJB Clustering

Posted by Juergen Weber <we...@gmail.com>.

David, thanks for your help.

Digging deeper I found this posting which is explaining things:
David Blevins Re: Client/Server Multicast Discovery and Failover
http://mail-archives.apache.org/mod_mbox/openejb-dev/200809.mbox/%3C0A2746FD-2820-4221-A223-60BDFED5F5D1@visi.com%3E

I think documentation could be based on this posting, missing is how to set
up the servers.

Looks like clustering needs multicast so I guess I have to persuade sysadms
to set up multicast ...

Cheers,
Juergen


djencks wrote:
> 
> Hi Juergen,
> 
> GCache never got finished.
> 
> The wadi clustering works but is for stateful beans and AFAIK it  
> provides only the state replication between cluster members and not  
> the client that supports failover.  I hope Gianny will tell me if I'm  
> wrong about this :-)
> 
> We do have the kind of clustering you want but I'm not sure if its  
> documented.  Basically the server cluster members maintain cluster  
> membership information using multicast hearbeats and the ejb client  
> maintains a list of servers and can change target server based on  
> availability.
> 
> There is an example of this at sandbox/failover and it used to work,  
> but AFAIK no one has tried running it for several months.
> 
> This would be a great thing to turn into a sample and make more public  
> and documented for 2.2.
> 
> thanks
> david jencks
> 
> 
> On Jul 15, 2009, at 8:34 AM, Juergen Weber wrote:
> 
>>
>> Hi,
>>
>> what is the state of Geronimo's EJB clustering?
>>
>> Basically I wonder if it is possible to have n Java application  
>> clients that
>> address m servers that run stateless beans. A client should stick to a
>> server as long it is alive and transparently go to another one if it  
>> is
>> dead.
>>
>> Is this information here still current for 2.2?
>> http://cwiki.apache.org/GMOxDEV/test-drive-geronimo-ejb- 
>> clustering.html
>>
>> Or would GGache be the way to go?
>> http://cwiki.apache.org/GMOxDEV/geronimo-clustering-with-gcache.html
>>
>> Thanks,
>> Juergen
>> -- 
>> View this message in context:
>> http://www.nabble.com/State-of-EJB-Clustering-tp24500428s134p24500428.html
>> Sent from the Apache Geronimo - Users mailing list archive at  
>> Nabble.com.
>>
> 
> 

-- 
View this message in context: http://www.nabble.com/State-of-EJB-Clustering-tp24500428s134p24513215.html
Sent from the Apache Geronimo - Users mailing list archive at Nabble.com.

Re: State of EJB Clustering

Posted by David Jencks <da...@yahoo.com>.

Hi Juergen,

GCache never got finished.

The wadi clustering works but is for stateful beans and AFAIK it  
provides only the state replication between cluster members and not  
the client that supports failover.  I hope Gianny will tell me if I'm  
wrong about this :-)

We do have the kind of clustering you want but I'm not sure if its  
documented.  Basically the server cluster members maintain cluster  
membership information using multicast hearbeats and the ejb client  
maintains a list of servers and can change target server based on  
availability.

There is an example of this at sandbox/failover and it used to work,  
but AFAIK no one has tried running it for several months.

This would be a great thing to turn into a sample and make more public  
and documented for 2.2.

thanks
david jencks

On Jul 15, 2009, at 8:34 AM, Juergen Weber wrote:

>
> Hi,
>
> what is the state of Geronimo's EJB clustering?
>
> Basically I wonder if it is possible to have n Java application  
> clients that
> address m servers that run stateless beans. A client should stick to a
> server as long it is alive and transparently go to another one if it  
> is
> dead.
>
> Is this information here still current for 2.2?
> http://cwiki.apache.org/GMOxDEV/test-drive-geronimo-ejb- 
> clustering.html
>
> Or would GGache be the way to go?
> http://cwiki.apache.org/GMOxDEV/geronimo-clustering-with-gcache.html
>
> Thanks,
> Juergen
> -- 
> View this message in context: http://www.nabble.com/State-of-EJB-Clustering-tp24500428s134p24500428.html
> Sent from the Apache Geronimo - Users mailing list archive at  
> Nabble.com.
>