IBM Storage Defender

 View Only

 COLLOCATION BY GROUP

Ulf Hedlund's profile image
Ulf Hedlund posted Thu February 26, 2026 03:36 AM

We want to avoid using too many tapes. 

If we create lets say three collocation groups, then add nodes to those groups with wildcard, the next retset copy to tape will use 3 tapes unil they are filled.

If nodes added to those three collogroups group1, group2, group3, are added with wildcard ( group1: node_a* , group2: node_b, group3: node_c ) is this a onetime add ? Or will a new regged node that has a name that falls into same pattern as when the first add was done automatically mean it's added into an already existing group ? 

Or du you have to constantly check group members and see if any new nodes are added , and add those so they are not treated as not be found in any collogroup and then be copied by collo by node.

I want to use as few tapes as possible, and don't want to see the copy job all of a sudden write on 28 tapes because we added 25 nodes later onafter the first collo groups were made.

Hope you get my point...  

Nezih Boyacioglu's profile image
Nezih Boyacioglu IBM Champion posted Thu February 26, 2026 06:32 AM

You are exactly right about the wildcard issue: the wildcard is not dynamic. it only adds the nodes that exist at the moment you run the command. Any new nodes registered later will be ignored by the group and start grabbing their own tapes.
If your primary goal is to use as few tapes as possible, I would actually recommend create 3 policy domains and 3 storage pools with Collocation=None. When you reg node use the target policy domain and it will use the destination storage pool of that particular policy domain. 

Related Content