NetBackup Agent for VCS 5.0MP3 Issue

[chrisa72]

I've just done an install of NBU 6.5.4 as per the NetBackup High Availability Administrator's Guide on RHEL 5 and I've run into an issue regarding the NetBackup VCS Agent on the second node.
I forgot that RHEL has iptables on by default and this caused some issues when I ran the cluster_config script. I disabled iptables and my NetBackup service group is up on the node I ran the cluster_config script on, but it can't be switched as the NetBackup VCS Agent is not installed or running.
# haagent -display NetBackup
#Agent       Attribute      Value
NetBackup    AgentDirectory
NetBackup    AgentFile
NetBackup    Faults         0
NetBackup    Running        No
NetBackup    Started        No
# haagent -start NetBackup -sys sysname
VCS NOTICE V-16-1-10001 Please look for messages in the log file
*Logfile Output*
2009/10/27 10:45:09 VCS ERROR V-16-1-10042 Could not locate agent binary file for type NetBackup

I know the NetBackupAgent file should reside in the /usr/openv/netbackup/bin/cluster/vcs directory, it was already there on the first node so I copied it into the directory on the second node and ensured permissions are the same. All the other files in this directory also appear to be the same.
I've also checked and all entries for NetBackup in the VCS main.cf are the same on both nodes.

Has anyone else encountered a similar issue?

Thanks in advance for all/any help.

Cheers,
Chris

[Marianne · Accepted solution]

The cluster_config was supposed to create the symbolic link.
Try this:
cd /opt/VRTSvcs/bin
ln -s /usr/openv/netbackup/bin/cluster/vcs NetBackup

[chrisa72]

Ahaaaaa, Thanks very much Marianne. Sounds like something you've encountered before?
I actually decided to wipe it and start the install again shortly after posting. It appears to be going better this time although it insists on trying to communicate with the second node using RSH, even though RSH is not running and I deliberately set up authorized_keys for SSH.
Did I miss an -ssh off the end of the cluster_config or something similar?

[Marianne]

I recently had to add a 2nd node to a one-node cluster - had to do everything on node 2 manually. It took me about 3 days to figure it all out. My cluster is still running!
The fact that -ssh is not an option to the cluster_config script is a MAJOR shortcoming as far as I am concerned. I missed this pre-req in the HA guide as well when I configured my 1st cluster:
Extract
Make sure that each node in the cluster, on which you want to install NetBackup, is rsh equivalent. As the root user you need to be able to perform a remote logon to each node without entering a password. This remote logon is only necessary for installation, upgrades, and configuration of the NetBackup server and any NetBackup database agents and options. After installation and configuration is complete the nodes no longer have to be rsh equivalent.

Maybe we should add the request for -ssh under 'Ideas'?

[chrisa72]

I think that's a great plan Marianne. We're encouraged to set up authorized_keys, especially on Linux clients for VCS and then told we're supposed to run rsh for HA NetBackup. Sounds like an oversight on someone's part, that's for sure ;)