-->
These old forums are deprecated now and set to read-only. We are waiting for you on our new forums!
More modern, Discourse-based and with GitHub/Google/Twitter authentication built-in.

All times are UTC - 5 hours [ DST ]



Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 3 posts ] 
Author Message
 Post subject: hibernate search clustering - JMS dynamic master selection
PostPosted: Fri May 25, 2012 12:54 pm 
Newbie

Joined: Fri Dec 09, 2011 1:09 pm
Posts: 2
Hi,

I'm wondering if it's possible for the to have the master be switched at runtime depending on the fact that the previously selected master has died. how can i do this via configuration???

Thank you.


Top
 Profile  
 
 Post subject: Re: hibernate search clustering - JMS dynamic master selection
PostPosted: Thu Jun 07, 2012 5:39 am 
Hibernate Team
Hibernate Team

Joined: Thu Apr 05, 2007 5:52 am
Posts: 1689
Location: Sweden
nellster wrote:
Hi,

I'm wondering if it's possible for the to have the master be switched at runtime depending on the fact that the previously selected master has died. how can i do this via configuration???

Thank you.


Hi,

this is a very advanced topic and outside of the scope of Hibernate Search. There is no option provided in Search to do such a thing.

In order to do that you probably need to look at several things. First, you need to configure your JMS setup to have a persistent setup. This means that messages gets stored so that in case the JMS setup fails or there is a power failure you are not loosing messages. Clustering JMS itself is a topic in itself.

The second part is clustering the master server. Given that the index is available over a shared file system you can have multiple masters. You just have to make sure that only one is active at the time. One way would be pacemaker (http://www.clusterlabs.org/). The idea is to setup a heartbeat and in case of the failure of the main master move the JMS queue consumer to one of the standby servers. You need to write some scripts/apps which allow you to start/stop message consumptions and configure these as "resources" in pacemaker.

As I said, this is outside of the scope of Search and very much on network/server level.

Depending on your needs it might be just enough to monitor the server (nagios etc). If the server goes down, you fix the problem and bring it back online. Once running the server will consume the indexing work in the queue. Given that the master/slave setup has some delay anyways in regards when new content becomes searchable this might be a workable solution.

I hope this gives you some ideas on what to look at to make your setup more failsafe. Feel free to share your experience and setup.

--Hardy


Top
 Profile  
 
 Post subject: Re: hibernate search clustering - JMS dynamic master selection
PostPosted: Thu Jun 07, 2012 11:54 am 
Hibernate Team
Hibernate Team

Joined: Fri Oct 05, 2007 4:47 pm
Posts: 2536
Location: Third rock from the Sun
I agree with Hardy so far you consider JMS only; also, applications servers might provide this service to you.

But we have plans for auto-election of the master node via JGroups channels and an Infinispan IndexManager which should do "the right thing" out of the box in the next release. There is a pull request open already for JGroups, but it still is to be considered highly experimental.

_________________
Sanne
http://in.relation.to/


Top
 Profile  
 
Display posts from previous:  Sort by  
Forum locked This topic is locked, you cannot edit posts or make further replies.  [ 3 posts ] 

All times are UTC - 5 hours [ DST ]


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum

Search for:
© Copyright 2014, Red Hat Inc. All rights reserved. JBoss and Hibernate are registered trademarks and servicemarks of Red Hat, Inc.