JBossDeveloper

We will create two nodes as live servers as well as backup servers for each other.

A live server cannot also be a backup server, however in the future we may change this

JBoss messaging isnt supported with AS6, HornetQ is the default JMS provider

you can either configure backup nodes or deploy with out failover (you may be able to colocate the backups in the same server as the live but you may need to upgrade the HQ version)

Hi Vineet,

I think you can raise a feature request on the HornetQ Jira. As Andy said JBoss Messaging is no longer supported within the community. Since AS 6.0, the default JMS provider is HornetQ. JBoss Messaging is not part of AS6 (and later) release bundle.

Howard

Hey Andy,

I was wondering if you can elaborate on the fail-over load balancing strategy a little bit more here. I am not clear what is the expected process. If we put into simple words here this is what I understand so far,  I configure a two node cluster in round-robin mode, each will get 50% of messages in general.

1) If one goes down does the other get 100% : No, according to this post and rest of the manual

2) To avoid loosing 50% of *New* messages we must configure a backup node for each server.

3) What about the messages which are on the failed node and were not delivered yet ??  They are lost unless the failed node comes back or the  backup is configured on shared storage (SAN) ? Is there a way to get active backups without SAN (double writing) etc.

4) How does the consumer get impacted by the node failure here ? Are they automatically redirected to the backups  ? What about the core-bridges etc are all automatically re-assigned etc

Best

Bhupesh

1) If one goes down does the other get 100% : No, according to this post and rest of the manual

correct, you need backup servers

2) To avoid loosing 50% of *New* messages we must configure a backup node for each server.

yes, as above

3) What about the messages which are on the failed node and were not delivered yet ??  They are lost unless the failed node comes back or the  backup is configured on shared storage (SAN) ? Is there a way to get active backups without SAN (double writing) etc.

yes, once the live node recovers or backup node becomes available messgaes are recovered

4) How does the consumer get impacted by the node failure here ? Are they automatically redirected to the backups  ? What about the core-bridges etc are all automatically re-assigned etc

consumers (including bridges etc) will reconnect automatically if configured to do so, i.e. reconnect-attempts etc

Hello,

Sorry for bumping this thread, but I'd like to know if there are any plans to support live to live failover. Failover using back-ups is ok, but I'd like to have this option too.

Thank you

not at the moment, but we may consider it in the future