Alfresco and clustering

[nico]

Hi,

How can I use Alfresco in a clustering environnement.
It's easy to cluster database, but what about lucene and JBoss/Tomcat ?

Is it possible to put the lucene indexer on a different server ?

Thanks in advance for your help.

Regards,
Nico

[derek]

Hi,

It is easy to cluster most databases.  You'll have to read up on the particular one you are using.

JBoss/Tomcat have solutions for load balancing and so forth.  If you are running multiple servers for load balancing, then you need to configure that in JBoss or Tomcat.  Then the Alfresco apps within that need to be configured to talk to each other.  It is also possible to have a read-only Alfresco on the cluster or a replicated server for backup, etc.

In essence, the following components need to be replicated between Alfresco applications:
* caches
* content store
Indexes are maintained individually on each server and will therefore follow the database replication by some configurable time.  Each server must have access to its own Lucene indexes.

The functionality to replicate the caches is part of the OpenSource distribution.  The functionality to replicate the content is Enterprise, but you could use your own tools or 3rd party tools to do.  The functionality to keep the indexes synchronized is part of the Enterprise distribution.

Regards

[nico]

Hi,

Thanks for your answer.

Is there any plan on supporting a dedicated indexing server (maybe something like Solr - http://incubator.apache.org/solr/features.html) ?

Does the enterprise edition do someting like that ?

Thanx in advance,
Nico

[derek]

Hi,

There are no plans for this at the moment.  Perhaps you would like to put together some specifications and reasons for the requirement on JIRA.

Regards

[rsfeir]

derek,

We were about to deploy to production on 2 servers and hit some snags as described above.

We're on a JBoss/Tomcat install (JBoss 4.0.3 SP1) with Alfresco 1.1.  We added content to which ever instance we hit (we're on round robin), and we can see it in one instance of ALfresco but not the other.

We're using the open source version.

We have a shared drive on which we've configured alfresco to save the content.

I am presuming our issue is a Lucene indexing issue.

The question is:

Using ONLY the Open Source version of Alfresco, what are our options with the given setup to get alfresco running on 2 servers sharing the same DB and same content?

Would clustering Alfresco with JBoss cluster work?  Are there specific Alfresco config changes we need to make and if so where are the docs that we can follow?

Thanks much.

R