I need to move my two node cluster (AIX 5.3, HACMP 5.4) to a different physical subnet. My private network will remain the same, but the routable IPs (service, persistants) will need to change. What is the best method to accomplish this (e.g., rm/create cluster, change IPs, etc.)?
Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations Mike Lewis on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!
Moving HA cluster to new subnet
- Thread starter sbf777
- Start date
I've never done that but i think it should be straight forward
you just need to change the IPs and update the /etc/hosts file
VERY IMPORTANT TO REMEMBER: make sure that all your IPs sharing the same subnet!!!
The cluster needs to be down for that!
R u using Replacement take-over or aliasing?
If replacement then the primary (service) IP addresses should in the same subnet with the boot address but different from the standby subnet
on aliasing then the three of them should be on different subnets!
Good Luck
Khalid
you just need to change the IPs and update the /etc/hosts file
VERY IMPORTANT TO REMEMBER: make sure that all your IPs sharing the same subnet!!!
The cluster needs to be down for that!
R u using Replacement take-over or aliasing?
If replacement then the primary (service) IP addresses should in the same subnet with the boot address but different from the standby subnet
on aliasing then the three of them should be on different subnets!
Good Luck
Khalid
I just performed something similar last weekend. I renamed all of my IP labels and upgraded to Gig Ethernet. It's not exactly the same, but uses the same steps.
I deleted the cluster on both nodes, updated the hosts table (and nodename) to reflect the new labels, and then just recreated the cluster.
If your cluster is well documented (networks, resource groups, dependencies, volume groups, etc.) it isn't that difficult. I had the luxury to test out all my changes on a test cluster first though.
The outage took about 20 minutes.
I deleted the cluster on both nodes, updated the hosts table (and nodename) to reflect the new labels, and then just recreated the cluster.
If your cluster is well documented (networks, resource groups, dependencies, volume groups, etc.) it isn't that difficult. I had the luxury to test out all my changes on a test cluster first though.
The outage took about 20 minutes.
once, due to some HA configuration problems my prod system I had to create test system to find the solution. to create exact configuration on the test system we directly cabled two test LPARS (2 physical ents on each LPAR), next we created appriopriate number of VLAN TAGS (smitty addvlan) on that "cross-cable" network and some "dummy" shared storage.
Next, we copied cluster snapshot from prod system (xxx.odm file from /usr/sbin/cluster/snapshots directory), carefully changed en# to appropriate ones (VLAN) and loaded such snapshot on our test system. I helpd us to deeply investigate the production system's problem and find the solution wich we implemented successfully on production later.
I think that the same would workk in your case but for security reasons it is better to do it from smitty hacmp menu (I had played on test system).
Next, we copied cluster snapshot from prod system (xxx.odm file from /usr/sbin/cluster/snapshots directory), carefully changed en# to appropriate ones (VLAN) and loaded such snapshot on our test system. I helpd us to deeply investigate the production system's problem and find the solution wich we implemented successfully on production later.
I think that the same would workk in your case but for security reasons it is better to do it from smitty hacmp menu (I had played on test system).
- Status
