Unsolved
1 Rookie
•
15 Posts
0
30
December 4th, 2024 08:56
Copy saveset
Hello,
I would like to know how I can copy savesets from clone pool back to prod pool.
The scenario is the following:
* Prod tape has failed
* Copy of savesets left in clone pool
* Prod tape has been replaced
* I'd like to recreate the redundancy of having more than one copy of every saveset by copying the missing ones back from clone to prod
However:
1) nsrclone only copies to clone pools
2) nsrstage only MOVES data
Of course I could first move from clone to prod and then clone it again, but why would that need to be this complicated?
Am I missing something?
Thanks,
Marki
No Events found!
bbeckers1
2 Intern
•
191 Posts
0
December 5th, 2024 20:22
Your assumption is correct with the clear distinction of regular backup pools and clone pools.
But as it is transparent for the one needing to restore data on which pool data is located if they "just" need to restore a specific ssid (regardless of the cloneid associated with it unless specifically stated), you can also clone the data to yet another clone pool, so that you would have two copies again. It would not be needed specifically to exist on a regular pool of type backup.
jer0nim01
1 Rookie
•
15 Posts
0
December 5th, 2024 22:36
@bbeckers1 Another clone pool would require more tapes. It would not make sense. The goal would be to restore the state in which the system was before, and not add another layer of complexity.
In essence, the restriction that copying a saveset to a non-clone pool would need to be lifted. In essence you'd simply need a staging command that does not remove the initial ssid from the database. Let's just have a copy/move command that lets you do anything.
But maybe I'm missing some history. Certainly, 20 years ago, there was a reason why it was done this way. nsrstage was introduced with backup-to-disk systems IIRC, and the entire world has changed anyhow since then...
bbeckers1
2 Intern
•
191 Posts
0
December 6th, 2024 16:38
@jer0nim01 it changed hugely indeed. we don't even have tape anymore, as we only use data domain. Dealing with the product for so long, one tends to take things for granted (also the features that we always wondered about from day one, like the differentiation between backup and clone pools) and you work around any annoyances. So business as usual really...
why there still is a difference what one can do with the nsrstage command versus the nsrclone command is one for example? nsrclone gives one more control about source and target NW storage nodes to use, so if you wanted to control that with staging, it would be better to use nsrclone with the -m (stage) option instead of nsrstage.
However in reverse there is no option to do cloning to a regular pool as you envision to do in reverse...