Hi. We're running CQ5 5.5 SP1 and I'm trying to setup an author instance on a machine at a disaster recovery site. I want the DR author instance to be clustered with our existing non-clustered author node on our staging server. If I get this to work, I'll redo it with the production author server, which will leave us with a hot backup server in the case that the primary data center goes dark.
I cloned the "master" author instance on our staging server using the CQ backup facility.
I restored that backup onto the DR server and proceeded to follow the directions found in the "Manual Slave Cloning" section of http://dev.day.com/docs/en/crx/current/administering/cluster.html
I then restarted the newly cloned slave isntance and saw that that it had automatically been added to the repository/cluster.properties file found on the staging master server. This implied that the two servers were now linked. You can also see this in the new slave's error.log file:
30.11.2012 08:06:56.801 *INFO* [FelixStartLevel] com.day.crx.core.cluster.ClusterController Trying to connect to a master, as the file clustered.txt exists.
30.11.2012 08:06:56.987 *INFO* [FelixStartLevel] com.day.crx.core.cluster.ClusterController Node author-dr started as: slave, connected to address: /XXX.XXX.XXX.237:8088
The problem is, the cloned slave author instance will not come up properly. I see the following error in the logs/error.log file:
30.11.2012 07:54:24.876 *ERROR* [FelixStartLevel] com.day.crx.core.data.ClusterDataStore Unable to download data with identifier: Unable to store record.
30.11.2012 07:56:08.669 *ERROR* [FelixStartLevel] com.day.crx.core.data.ClusterDataStore Unable to download data with identifier: Unable to store record.
They seem to continue as long as the instance is not shut down. While this is happening, attempting to connect to the CQ admin at http://dr-hostname:4502/ results in a simple "Not Found" error in the browser, which is what you see when you attempt to connect to the admin when the instance is first starting up.
This implies that perhaps the instance is never starting due to this problem.
Can anybody suggest what the issue might be? I had no success with creating a new quickstart instance and using the GUI to join the existing author in a cluster. The "Loading..." message showed for a long time before disappearing with nothing appearing to have worked and no error message.