How to do a failback?

[John Williams]

Having some problems getting a failback (with SyncIQ) to work on a 5400S cluster with OneFS 7.0.0.
Trying to drive the process entirely from the command line to better understand what is going on.
Nothing fancy - After a failover, want to do a restore and demonstrate to my client (and to myself) how updates to the destination are copied back and how the replication can be later resumed.

The SyncIQ best practices document (2012) identify the following as the command to run to start the failover...
# isi sync target allow_write <replication-policy-name>.

And the following are for the failback...

# isi sync resync prep <replication-policy-name>

# isi sync policy run <replication-policy-name>_mirror

Now, upon running the second command, the initial source directory is still not writable.
It doesnt become writable until I delete the mirror policy as follows..

# isi sync policy delete <replication-policy-name>_mirror

But that step isnt in the best practices guide or the 7.0.0 Admin guide, so perhaps it wasnt the right thing to do??
But at least that makes the source writable again.

Then we are in a state where the destination cannot be written to on subsequent runs of the policy..
And under "isi sync target list" I see this message of "resync policy created".

CLUSTER-3# isi sync target list                  
Policy            | Source  | Target Path           | Run     | FOFB State          
------------------+---------+-----------------------+---------+----------------------
copytest1         | isi2    | /ifs/copytest1tgt     | Success | resync policy created
....

We are also having the question - Do we need to manually re-enable the initial replication policy?
Should we just disable the _mirror policy, or delete it?

I have seen (or heard of at least) others who use fail back just fine. My trials and failures suggest there are one or more critical steps that are part of the process that are not documented.

Any help is appreciated.. thanks!!

J