crystalball07
Technical User
I am tasked with relocating a two-node Sun Cluster to a new data center where the nodes and logical hosts must be assigned new IP addresses.
Hardware:
E420R's running Solaris 7 (11/99 release)
2 x A1000 StorEdge Arrays dual-attached to cluster nodes
Software:
Sun Cluster version 2.2
RAID Manager v6.22
Veritas Volume Manager v3.04
The high level Plan:
1. Change IP's of the respective logical hosts.
2. Change IP of TC (Terminal Concentrator).
3. Shutdown cluster on node 2 and node 1 respectively.
4. Create appropriate backup files and "sys-unconfig" both nodes and cluster admin workstation.
5. Move servers, boot and reconfigure each node with it's new IP addresses. In my case, the "Primary public" interface is ge0 and "Secondary public" is qfe2.
6. Restart cluster on node 1, then node 2.
Questions:
1. Does it matter in which order the IP's are changed? ie. logical hosts first, then cluster nodes... or the other way around?
2. Can a successful "fail-over" occur if the logical host IP's have been changed but not the cluster nodes IP's?
3. Per scconf man pages, the IP address of the TC can be changed with the command: scconf <clustername> -t <TC_hostname> -i <new_IP_Address>
Does this actually change the IP of the TC or just update info in cluster configuration files?
Many Thanks!!
Hardware:
E420R's running Solaris 7 (11/99 release)
2 x A1000 StorEdge Arrays dual-attached to cluster nodes
Software:
Sun Cluster version 2.2
RAID Manager v6.22
Veritas Volume Manager v3.04
The high level Plan:
1. Change IP's of the respective logical hosts.
2. Change IP of TC (Terminal Concentrator).
3. Shutdown cluster on node 2 and node 1 respectively.
4. Create appropriate backup files and "sys-unconfig" both nodes and cluster admin workstation.
5. Move servers, boot and reconfigure each node with it's new IP addresses. In my case, the "Primary public" interface is ge0 and "Secondary public" is qfe2.
6. Restart cluster on node 1, then node 2.
Questions:
1. Does it matter in which order the IP's are changed? ie. logical hosts first, then cluster nodes... or the other way around?
2. Can a successful "fail-over" occur if the logical host IP's have been changed but not the cluster nodes IP's?
3. Per scconf man pages, the IP address of the TC can be changed with the command: scconf <clustername> -t <TC_hostname> -i <new_IP_Address>
Does this actually change the IP of the TC or just update info in cluster configuration files?
Many Thanks!!