I have a 4 node cluster which has 7 virtual servers configured. I have noticed that when CommVault Galaxy has an issue with backing up the virtual servers, I can't unload the galaxy client, even though all jobs for that server have been killed to close down the pipes.

This results in loss of control over the cluster node (can't fail over resources to other node(s) & reboot problem server,) and in the worst case, the cluster view in Console 1 shows all 4 nodes greyed out (zero control).

I have resorted to loading Galaxy with a "?Galaxy" in the autoexec.ncf file, so that when I reboot the problem cluster nodes, I can regain control the cluster either by C1 or the CLI on the server.

If I do load Galaxy immediately after a reboot to resolve the hung galaxy unload, I am not always able to take resources off line/on line to revert them from the comatose state after the crash.

Two ways to shut down for this scenario are, if you have tried to unload Galaxy and the server console is hung, open another emergency console and unload NCS. This is not always successful, as the resources do not always migrate to the alternative node, and it the worse case, the server gets the poison pill.

The other is to try to unload NCS before rebooting the server, but this can also result as above.

Usually, you have to take both nodes down that manage the resources assigned to them, and in the worst case, I have not been able to control any resources to take them off line/on-line, I have to reboot all 4 nodes and bring them back up in sequence with out Galaxy loaded, get the resources online and then load Galaxy.

The unload screen from galaxy reports a warning on CVMADISP.NLM and CVLIB.NLM, but does not give any other info on the warning message.

The NLM that always hangs when unloading from this scenario is CVMADISP. Note however, you can unload Galaxy when no back jobs are running and no problems have occurred!

NetWare = 6.5 sp4a
Cluster Services = 1.7
CommVault Galaxy NLM = 5.90
CommVault Patches= nw_win_5.9.0B48_SP3 including abend patch nw_5.9.0B48_534

Galaxy is installed on all for physical nodes & all virtual servers for backup.