Both NT Cluster nodes having the same FQDN.. how to backup?

[DennisN]

Hi there!

Both our NT Cluster nodes have the same FQDN: mail.lvnl.nl
I've changed the client resourcesin Solstice to set the aliases for both nodes to mail.lvnl.nl

Both nodes are in a backupgroup called NT_Clients.
Backup of the first group works, but as soon as the second one requests a backup, the second node fails to backup (RAP error)

Is there a solution for this problem without
1) renaming the nodes again (unacceptable)
2) Placing both of the NT nodes in different groups on different times (unacceptable)

We're using Networker 5.1. Our backupserver is a UNIX server.

The backup used to go smoothly until the NT-admins changed the FQDN from primary.lvnl.nl (node1) and secondary.lvnl.nl (node2) to mail.lvnl.nl.
Their idea was that only one of the nodes is server for our MS Exchange Server, so that it wouldn't matter that both nodes have the same FQDN. Too bad Solstice does a hostname command to the client which gives 'mail' on both nodes...

Has anyone seen this problem before?

Thanx in advance,
Dennis van der Meijden
Sysadmin
Air Traffic Control, the Netherlands

 

[john7]

You got big problems.

Networker 5.1 really does not support clusters, 5.5.1 supported MS CLuster and 6.x supports more cluster software like Sun Custer, Legato CLuster, HP service Guard.....

The new cluster support allows you to backup the physical nodes and virtual nodes spearately, and give them their own indexes. But you will need new CLuster CLient Connections.

By changing the FQDN to the same for each node in the cluster, you only have one index, this can be a real problem if you want to to a Disater Recovery.

Are you using MS Cluster? Have you tried to DR your Cluster? it is a real pain in the butt, it can be done, I spent 3 weeks (with Networker) figuring it out, before we clusters into production. Networker was the only product that could do it.

Have you tried using aliases?

Have you tried doing client side initiated backups?