Always run a 'clRGinfo -v' priory to starting cluster services on multi-site nodes.

When you move a PowerHA resource group to a new site, extra care should be taken as it could cause unwanted issues at a later date. This is because PowerHA sets the new target site as the 'Primary Site' albeit temporarily.

Imagine the following cluster configuration:

# clRGinfo -v

Cluster Name: Southbank_EE_Linked

Resource Group Name: Southbank_RG1
Startup Policy: Online On Home Node Only
Fallover Policy: Fallover To Next Priority Node In The List
Fallback Policy: Never Fallback
Site Policy: Prefer Primary Site
Node                             Primary State     Secondary State
----------------------------  --------------------- ----------------------------------------
pts114@SiteA              OFFLINE            OFFLINE
pts116@SiteA              ONLINE              OFFLINE
pts115@SiteB              OFFLINE            ONLINE SECONDARY
 

The primary site 'Site A' has 2 nodes (pts114 and pts116) and the secondary site 'Site B' has 1 node (pts115). If we move the 'Southbank_RG1' to Site B using: smitty hacmp -> System Management (C-SPoC) ->  Resource Group and Applications ->  Move Resource Groups to Another Site, it moves as expected however we now see a new line in the clRGinfo output:

# clRGinfo -v

Cluster Name: Southbank_EE_Linked

Resource Group Name: Southbank_RG1
Startup Policy: Online On Home Node Only
Fallover Policy: Fallover To Next Priority Node In The List
Fallback Policy: Never Fallback
Site Policy: Prefer Primary Site
Site settings:
The following site temporarily has the highest priority for this group:
SiteB, user-requested rg_move performed on Thu Aug  7 13:15:29 2014

Node                         Primary State    Secondary State
---------------------------- ---------------
pts114@SiteA                 OFFLINE          ONLINE SECONDARY
pts116@SiteA                 OFFLINE          OFFLINE
pts115@SiteB                 ONLINE           OFFLINE

We can see that PowerHA has set 'Site B' to the highest priority site. So lets imagine the node on site B crashes, pts115. The resource group will move to a node on site A as expected but the temporary site priority setting is not reset.

Once node pts115@SiteB is rebooted and PowerHA restarts it WILL take back the resource group as PowerHA thinks that site has priority. This is not something most people look for but worth nothing.

Always run a 'clRGinfo -v' priory to starting cluster services on multi-site nodes.

Tags:  outage secondary primary site cluster rgmove powerha failover clrginfo sites policy