Sorry, forgot to mention that the command 'symdg list' still totals up to a number that includes the original disks and the command 'symdg export' to a file still shows the original disks.

Use symld and remove from the DG the devices already removed from the link with the "deletepair"

Kook I've seen that you already put questions in the past .. but it seems that they are still unanswered .. Could you please review your previous threads and mark them as answered (if you received an answer) ?? Even better, before marking your threads as answered, assign also points to correct or helpful answers someone offered you ..

Okay, but then how do I get back to a valid RDF state ?

Since I can not see the devices in group using command 'symdg list' I was thinking of using 'symdg export' to a file. Edit the file leaving only the devices I need and then doing a symdg import. I have confirmed using symrdf -rdfg xx -sid xxx list that only the devices I want are replicating.

If I do this will this mean the group will become valid again.

Instead of suspending the whole DGs you can put them in adaptive copy .. you'll still be able to delete unneeded pairs .. and it will be faster and smoother to turn again to async.
If you want to try this way, please don't touch the skew and leave the default.

The "valid" state in "symdg list" refers to the fact that the DG is valid (contains the right volumes). Now your DG contains some RDF devices and some NON-RDF devices (the ones you deleted before from the RDFG but not from the DG) so the status of the DG lists as "invalid".

If you really want to know how does your link is working, issue a "symrdf -rdfg xx -sid yyy list"

Message was edited by:
Stefano Del Corno

Yes Group is a vlaid RDF1 group again and I will be enabling consistency.

The method I used was to suspend and delete the devices I do not need, then export the device group out to a file, edit the file and keep only the devices I need and import it again and it works.

Is this the best method ?

Are your groups valid again ??
Did you already reached the "consistent" state on your groups ??
Remember to enable consistency protection acros your RDFg

Yes the method you used is one of the possible ways of doing the same thing

My method:
0) clean DGs and remove to-be-deleted devices (use symld)
1) symrdf -rdfg $xx -sid $yyy list | grep -v " - " | grep "R.:$xx" | cut -c 1-10 > pairfile.old
2) create 2 new pairfiles from the first one ... the first one (pairfile.delete) will contain to-be-deleted devices while the second one (pairfile.new) will contain the remaining devices.
3) symrdf -rdfg $xx -sid $yyy -file pairfile.old disable
4) symrdf -rdfg $xx -sid $yyy -file pairfile.old set mode acp_disk
5) symrdf -rdfg $xx -sid $yyy -file pairfile.delete susp
6) symrdf -rdfg $xx -sid $yyy -file pairfile.delete deletepair
7) symrdf -rdfg $xx -sid $yyy -file pairfile.new set mode async
and later on
8) symrdf -rdfg $xxx -sid $yyy -file pairfile.new enable

Since we don't use DGs we usually jump to step 1

Thanks again for your help, I am sure I will be in the forums again soon.

Stefano:

How do I do this (remove device pairs) if I'm using Device Groups?

   Stuart

Thank you Me stefano.. those steps are great help for my troubleshooting today.

only your document says about this step

symrdf -rdfg $xx -sid $yyy -file pairfile.old disable