Veritas Enterprise Administrator not recognizing host properly

[Jamie A Dennis]

Hello,

Using VEA through SAM, and not seeing any of the disk groups, logical volumes, or filesystems I've created. Disk groups were created using vxdg, disks added using vxdg, logical volumes created using vxassist, filesystems created using mkfs (with vxfs as the type).

The VEA has configured itself to use the hostname of one of my secondary interfaces, and not the primary hostname/IP.

When entering VEA, the gui scrolls through all the agents (calling them out by secondary interface hostname) with a message that it cannot contact the agent, and it exits.

Does anyone know which files control how VEA sees the server and its interfaces, and how to correct this situation?

I've tried running installsf -configure and still see the same situation.

Thanks,

J

[Richard Hepworth]

Jamie,

What hostname does 'vxdctl list' show?

Also have you tried using the windows VEA client and connecting through to the IP address of your secondary interface?

[Richard Hepworth · Accepted solution]

Just found this, which should hopefully help:

http://seer.entsupport.symantec.com/docs/301651.htm

[Sharma Sanjeev]

Hi

Try vxdctl enable

Regards
Sanjeev

[Jamie A Dennis]

Richard,

/wdroot# vxdctl list
Volboot file
version: 3/1
seqno: 0.1
cluster protocol version: 70
hostid: (correct host name)
hostguid: {f098b42e-1dd1-11b2-b673-001a4b07d491}

I found a procedure last night similar to the one you sent in the link, which I tried to no avail. This procedure you sent has more detail, so I'll give that a shot as well. Thanks again for your help.

Sanjeev - The problem isn't that the disks aren't showing up properly in the OS, they just aren't showing up in VEA. All OS level commands function exactly as expected. Thanks for your input into the problem though, I do appreciate it.



Jamie

[Jamie A Dennis]

Richard,

Ok, I ran the procedure you sent me:
http://seer.entsupport.symantec.com/docs/301651.htm.

It's got most of what a person needs to do on HP-UX, despite being written for Solaris. :)

After running some of the steps, I realized that the agents had not stopped despite my using the /sbin/init.d/xvpal.{agent} stop command. I was forced to kill them manually (seemed orphaned anyway).

These are the commands for HP-UX
/sbin/init.d/vxpal.StorageAgent stop
/sbin/init.d/vxpal.VAILAgent stop
/sbin/init.d/vxpal.actionagent stop
/sbin/init.d/vxpal.gridnode stop

Then run vxsvcctrl stop.

Next, and this does seem to be very key -move the /etc/resolv.conf to /etc/resolv.conf.save and /etc/nsswitch.conf to /etc/nsswitch.conf.save, and make sure /etc/hosts contains ONLY the name of the server which you want the agents to be registered under. Not sure if the installer using some greps which pulled the wrong interface, since our primary interface is named the hostname, and our backup interface is essentially the same but with an indicator that it is on the backup network ( hostname-bu for example).

Finally, run the installsf -configure command to reconfig the agents to the proper hostname/fqhn.

At this point, we're looking good. I reran all the checks in this advisory, and all report the proper hostname and where needed, IP address. I can now see the objects in VEA with no problems, and am not getting the errors I was before.

Thanks again Richard! Great work!

Jamie

[Jamie A Dennis]

Closed as per my post above. Thanks to all who assisted.