renaming offsite data

[gernthefish]

We use DRM on one of our TSM servers and have certain archives from several nodes sent to DR_COPY_POOL for offsite storage. To better organize things, we thought it would be best to create separate nodes just for offsite data - something like OFFSITE_FINANCE for instance:

- Does this sound like a good idea?

- How would the change be done? We use server-server communication, and I think we can do it the following way using serverB (ServerA is where the offsite data and DRM reside). Node1 and node2 are on serverA. Node2 needs to be kept as it has other data that is not in the DR_COPY_POOL. Does this process make sense?
1. serverB -- Register node node1
2. serverA -- Export node node1 filedata=archive toserver=serverB
3. serverB - Rename node node1 to node2
4. serverA - Export node node2 filedata=archive toserver=serverB merge=yes
5. serverA - Register node offsite_finance
6. serverB -- Rename node node2 to offsite_finance
7. serverB - Export node offsite_finance filedata=all toserver=serverA
8. serverB - Remove node node1
9. nodeB - Remove archives for node2 (using client)


Could this work or am I on drugs? Will TSM require reading in all the tapes that are offsite? That would make it a very messy prospect.

Thanks for your help!

 

[Eldoraan]

Perhaps not drugs, but just not really understanding how DRM module works. You shouldn't need a second server or renamed nodes or anything complex like that. Just clients backup to primary storage pools, backup stg commands copy that to copy storage pools and DRM module manages the copy storage pool media. The node name stays the same since the data in both pools is still that node's data. I'm not sure I understand what you're trying to accomplish here.

 

[gernthefish]

The idea is to keep all offsite data within "nodes" that have a standard naming convention (in this case OFFSITE_xxxx). Each of these nodes only contains data going offsite and potentially contain offsite data from multiple clients. Some of these clients (node2 in this case) need to keep their original node name for backups/archives that are NOT going offsite. To do this, I believe the offsite data will need to be merged into new node offsite_finance.

 

[BBB]

You're making it way more complicated than it needs to be.

1. Create management classes that send data to a separate primary storage pool to the normal one.
2. Get the clients to use those management classes only for data that needs to be sent offsite. (see include/exclude options, or for eg if archives, use -archmc flag)
3. Run "backup stgpool <newprimarypool> <offsitepool>" each day - and you're done. Much easier, and its the normal way to do things.