Re: [Exist-development] eXist replication & clustering

SourceForge Headquarters 225 Broadway Suite 1600 San Diego, CA 92101 +1 (858) 422-6466

Alexei,

great to read you have interest in the replication extension. This extension is just the start for all kind of ideas we have...

The branch you mention is quite out-of-date, the wiki page as well.
The most recent code is in trunk, please get a copy of that.

Included is a Readme file, and some XML fragments, these should be sufficient to get started.

Last time I checked the extension worked, but I did some refactoring after that, I might have overlooked some things making it fail. Unfortunately I can't have look at it right now for some time.

There are a few rough edges:
- the users must be existing on both instances
- on the consumer side the 'receiver' directory must exist and be writable
- there are memory limitations

There is sufficient logging when set on debug level.....

Thnx for providing your feedback. I consider the extension code stable, but it requires additional testing 'in real life'. Your input is appreciated and will result into better software!

Cheers

Dannes

--
Dannes Wessels

On 2 okt. 2012, at 06:47, Alexei Tarnakin <ale...@gm...> wrote:

> I made a small research on eXist clustering capabilities. According to Wikibooks article I checked out dizzzz's branch and tuned up ActiveMQ server and a couple of eXist servers.
> 
> First of all, I've discovered that article is totally out of date. collection.xconf should look like 
> 
> <collection xmlns="http://exist-db.org/collection-config/1.0">
> <triggers>
> <trigger class="org.exist.replication.jms.publish.ReplicationTrigger">
> <parameter name="java.naming.factory.initial" value="org.apache.activemq.jndi.ActiveMQInitialContextFactory"/>
> <parameter name="java.naming.provider.url" value="tcp://localhost:61616"/>
> <parameter name="connectionfactory" value="ConnectionFactory"/>
> <parameter name="destination" value="dynamicTopics/eXistdb"/>
> <parameter name="client-id" value="id1"/>
> </trigger>
> </triggers>
> </collection>
> 
> If  you dig svn log, you will notice that 
> org.exist.clustering.triggers.send.ClusterTrigger
> was renamed to
> org.exist.replication.jms,publish.ReplicationTrigger
> And also notice that client-id looks like mandatory, but not optional field and separated with a dash, not a dot (i.e. client-id). 
> 
> Then lets go check slave eXist server. Article is wrong again. Proper conf.xml should look like
> 
> <job type="startup" name="clustering"  class="org.exist.replication.jms.subscribe.MessageReceiverJob">
>         <parameter name="java.naming.factory.initial" value="org.apache.activemq.jndi.ActiveMQInitialContextFactory"/>
>         <parameter name="java.naming.provider.url" value="tcp://localhost:61616"/>
>         <parameter name="connectionfactory" value="ConnectionFactory"/>
>         <parameter name="destination" value="dynamicTopics/eXistdb"/>
>         <parameter name="client-id" value="id2"/>
>         <parameter name="subscriber-name" value="sub_name"/>
> </job>
> 
> OK, now looks like we configured everything properly, bot does cluster work now?... Nope! 
> Let's try discover why it does not work by means of xdebug. So simply add a couple of breakpoints to ReplicationTrigger and JMSMessageListener to check what we actually send to ActiveMQ and get on other side (yep, I actually skipped several steps like reconfiguring log4j.xml, restarting server & digging logs, cause it's boring). 
> So the problem is that master eXist sends eXistMessage with document type literally "XML" or "BINARY", while slave expects "DOCUMENT" or "COLLECTION". The problem as always was in indian style of code. Lets analyze it a little bit 
> 
>     @Override
>     public void afterCreateDocument(DBBroker broker, Txn transaction,
>             DocumentImpl document) throws TriggerException {
> 
>         ......
>         // Create Message
>         eXistMessage msg = new eXistMessage();
>         msg.setResourceType(eXistMessage.ResourceType.DOCUMENT);
>         msg.setResourceOperation(eXistMessage.ResourceOperation.CREATE_UPDATE);
>         msg.setResourcePath(document.getURI().toString());
> 
>         // Retrieve Metadata
>         Map<String, Object> md = msg.getMetadata();
>         MessageHelper.retrieveDocMetadata(md, document.getMetadata());
>         MessageHelper.retrieveFromDocument(md, document);
>         MessageHelper.retrievePermission(md, document.getPermissions());
>         .......
>         sendMessage(msg);
>     }
> 
> so resourcePath in eXistMessage was setted twice 
> msg.setResourceType(eXistMessage.ResourceType.DOCUMENT);
> setted to "DOCUMENT" and then one more time
> MessageHelper.retrieveFromDocument(md, document);
> setted to "XML" due to a bit of hardcode in MessageHelper. To fix it we need to patch it
> 
> Index: extensions/replication/src/org/exist/replication/shared/MessageHelper.java 
> =================================================================== 
> --- extensions/replication/src/org/exist/replication/shared/MessageHelper.java    (revision 17019) 
> +++ extensions/replication/src/org/exist/replication/shared/MessageHelper.java    (working copy) 
> @@ -159,7 +159,9 @@ 
>      public static void retrieveFromDocument(Map<String, Object> props, DocumentImpl document){ 
>           
>              props.put(EXIST_RESOURCE_TYPE,  
> -                    (document.getResourceType() == DocumentImpl.XML_FILE ? "XML" : "BINARY"));  
> +                      (document.getResourceType() == DocumentImpl.XML_FILE ?  eXistMessage.ResourceType.DOCUMENT : "BINARY")); 
>              props.put(EXIST_RESOURCE_DOCUMENTID, document.getDocId());  
>              props.put(EXIST_RESOURCE_CONTENTLENGTH, document.getContentLength()); 
> 
> now we got document updates enabled, but no collection operations yet. Let's try to fix it....
>                 ...well, I'm not going to test your patience, so just attached svn diff below and I'll keep on working. And I do hope that 2.1 release will be clustering/replication enabled, cause eXist really needs it due to lack of reliability under pressure.
> 
> 
> P.S. repo moved so proper link for checkout is http://svn.code.sf.net/p/exist/code/branches/dizzzz/clustering/
> 
> P.P.S. Sorry for my English, it's not my native
> 
> 
> <svn.diff>
> ------------------------------------------------------------------------------
> Don't let slow site performance ruin your business. Deploy New Relic APM
> Deploy New Relic app performance management and know exactly
> what is happening inside your Ruby, Python, PHP, Java, and .NET app
> Try New Relic at no cost today and get our sweet Data Nerd shirt too!
> http://p.sf.net/sfu/newrelic-dev2dev
> _______________________________________________
> Exist-development mailing list
> Exi...@li...
> https://lists.sourceforge.net/lists/listinfo/exist-development

Re: [Exist-development] eXist replication & clustering

eXist-db is a feature rich Open Source native XML database

Re: [Exist-development] eXist replication & clustering