DFM keeps insisting that there is a deleted SnapVault relationship that it tries to recreate when I want to get a normal Conformance status. The point is, the relationship has been deleted on purpose, and DFM should just forget about it.
The error message (please see attachment) is:
Could not proceed automatically with this task. Found a deleted relationship for primary 6:/snapmirror_vol_cifs_homedir_2/- and dpReBaselineMode is set to Confirm.
Create backup relationship(s) between '6:/snapmirror_lun_exchangelog_sas_01_1/-' and '1:/snapmirror_vol_cifs_homedir'.
How can I tell DFM to stop wanting to recreate this last relationship?
You are probably well past this problem since it's been a couple months but I came across this thread trying to find a solution to a similar issue I was having with the dfpm dataset remove command. If you don't specify a node name with the -N option the command assumes the resource you are trying to remove is in the default or root node. Since most policies don't really apply much to the root node this is often not the case. In your case since this is a SnapVault relationship the node you want to target within the policy is probably the Backup portion. To confirm this you can run the dfpm policy node list policy-name (the name of the policy assigned to the dataset in question). This should give you something similar to this:
1 Primary data
The other option that may show in replacement or addition to Backup is DR Mirror, it all depends on the policy settings applied. In order to get the dataset remove command to work you need to add the -N Backup to your command so it should look like this:
dfpm dataset remove -N Backup 1018 2248
If you want you can add the -D command to do a dry run to see what would happen but I think this will get you where you need be. Of course assuming you are still having to deal with this and didn't just go and remove the resource via the management console. Just thought I would put in an answer to the problem in case someone else comes along and wants to script a mass dataset redesign like myself and was hoping to avoid the GUI.