[opencms-dev] How Run Opencms in Clustered Environment.?

Georg Raffer georg.nc at gmail.com
Tue Jun 6 12:28:15 CEST 2006


Hi Christian,

keep in mind, that OpenCms stores the user information NOT in the session.
they use the jsessionid as key in a hashmap to restore user information. 
so session replication doesn't help.

ciao, Georg

Christian Steinert wrote:
> Mike McGrath schrieb:
>>
>>>
>>> even with OCEE, you can only use one server for editing (but you can 
>>> display content on multiple servers).  if you just deploy the same 
>>> application on multiple servers pointed at the same database, it 
>>> should mostly work ok, though cached content will not automatically 
>>> be updated when it is edited.
>>>
>>> also, automatic failover will not work correctly, because OpenCms 
>>> stores data tied to the user session outside of the actual 
>>> HttpSession object.  so if you want to use the cluster to provide 
>>> high availability in the event of a server crash, i think you either 
>>> need to look into OCEE or make some modifications to the code yourself.
>>>
>>> jt
>>>
>> jt - Where does the session stuff reside?  At present I'm thinking 
>> about making a san cluster with Opencms using either NFS or GFS on 
>> the back end.  It *appears* to be working.  Basically I have the 
>> webapps directory on a SAN and the servers all read from it.  Anyone 
>> see any problems with this?
>>
>>    -Mike
>>
> Hi Mike,
>
> I have not looked into this yet, but probably you have to wrestle with 
> your servlet container in order to make it replicate your sessions.
> E.g. for tomcat: 
> http://tomcat.apache.org/tomcat-5.5-doc/cluster-howto.html
>
> regards
> c.
>
> _______________________________________________
> This mail is sent to you from the opencms-dev mailing list
> To change your list options, or to unsubscribe from the list, please 
> visit
> http://lists.opencms.org/mailman/listinfo/opencms-dev
>




More information about the opencms-dev mailing list