Originally posted by: oldSE

I cannot find documentation on how to set up Full Concurrent Mode PowerHA so that all nodes are participating in the processing and "failover" is just the circumstance of one node falling out of the cluster.

I understand that this would require multiple service addresses or routing via a one-to-many configuration.  With a large database application running, regular failover leaves locks on the database and the failover results in mounted file systems only with an admin needing to clear lock files to restart the application.  What we need is true concurrent processing so that a Node leaves, at worst, an incomplete transaction to be completed or rolled back, but otherwise just quits processing until fixed and brought back in.  The users should see no "failover"; rather, they just continue albeit it on a potentially different Node.

Setting a concurrent VG and file systems is no issue.  I just can't find the documentation that describes the setup to accommodate the concurrent processing and redistribution of users with the advent of a failed node.