[gaohoward]

Thanks. Looks like your code works for me. However I use 2.3.0.beta and the client needs a little change. I believe your version of hornetq is a little outdated and there is some bug you just hit. Can you upgrade your test to later versions?

Howard

[raymanf]

i am using HQ_2_2_14_FINAL. I afraid of using beta in our production servers.. I prefer final versions and this is the latest one.

I have added  my classpath. mybe some collision within my jars?

[gaohoward]

That doesn't seem right to me. When the live is killed and backup is activated, there should be something printed like:

16:13:20,749 INFO  [org.hornetq.core.server] HQ221010: Backup Server is now live

Make sure the live and backup point to the same journal store.

Howard

[raymanf]

Hi,

I have uploaded you my configuration files of both servers.

they both having the same values:

<paging-directory>${data.dir:./data}/paging</paging-directory>

<bindings-directory>${data.dir:./data}/bindings</bindings-directory>

<journal-directory>${data.dir:./data}/journal</journal-directory>

<journal-min-files>10</journal-min-files>

[raymanf]

Hi,

You were right there was problem with the same journal store and now they seem to work in case of failover.

but now I have another issue all of sudden my MDB doesnt consume the message in the second node cluster after i send the second message(and after killing the first node).

what could be the prob?

thanks.

[raymanf]

Thats the error I have noticed in my "backup" node:

Caused by: javax.naming.NameNotFoundException: topic not bound

but if I change the backup attribute from my backup node's hornetq-configuration.xml to false:

then node 2 will work fine as individual.

any idea why I get  topic not bound?

[gaohoward]

Can you post your MDB configuration?

[raymanf]

Ok it work. apperntly the second node must be identical to the master node within everything concerd to the busniess logic.

thanks for your help sir.

ray.