and excuse the English :-)
I am installing IBM platform HPC v4.2 for learning purpose in high availability mode.
My lab is :
- physical machine Fedora 25 : nfs3 for shared storage
- VM1 : manager1 : RHEL 6.5
- VM2 : manager2 : RHEL 6.5
The installation goes like this :
- install phpc on first manager: manager1
- install phpc on second manager: manager2 (using the same options)
- configure an NFS3 share (home and data) for HA
- trigger the HA installation on manager1
All these steps were done successfully, but at then end of the HA installation, I couldn't log into phpc webmin interface, I had an error of some services/daemons which are not running.
After some digging, I have found the problem, I think I have at least :-)
The problem comes from the ActiveMQ process, the process starts but it doesn't bind to it's port (61616). - the port is free and open on firewall.
The activemq process uses /opt/pcm/activemq/data to store it's log file, pid file and kahadb directory which contains it's database (I don't know too much about activemq).
In HA mode, the data directory is copied into the nfs3 share, mounted on /shared_phpc/opt/pcm/activemq/data, and a symplink is created on both managers on /opt/pcm/activemq/data to point to that directory.
/opt/pcm/activemq/data -> /shared_phpc/opt/pcm/activemq/data
If I rename the symlink and use the real directory (which was backed-up by the HA installer), the activemq process starts successfully, but if I use the symlink it doesn't, it starts but doesn't bind to it's port.
I even deleted all the files on the shared data directory to see if it was some access rights problem, but it was not, the activemq process starts, re-creates its log file, pid file, the kahadb directory, the lock file in the kahadb directory, but it doesn't create the kahadb files, and doesn't start properly.
I have no error messages on the log file, I don't know where to look now.
Thanks in advance.