Stage between storage nodes

[fraxx]

I might have asked this before, but I've been getting mixed answers from different people, so I thought I'd ask again. This time with a more detailed scenario:

Within the same zone I have several remote sites with storage nodes. On those storage nodes I have direct attached disk, to which I want to backup to.

I then want to stage the backups to _another_ storage node over a WAN connection to a central data centre. The WAN connection can be considered to have sufficient bandwith, and the storage node in the data centre will place the data on another disk so I don't have to worry about shoe-shining on the tape drive.

On the storage node in the central data centre I want to stage the backup one more time from disk to tape.

The purpose is to elminate tape handling on the remote sites.

Although I see this as technically possible in NetWorker, since it's possible to clone between storage nodes, how feasible is it?

Can I create such a multi-step staging policy within the GUI, and expect it to run as smooth as staging is when it's only done in one step in one single storage node from disk to tape?

I don't want to have to script the process.

With regards to restores to a client in case the remote storage node with it's disk is lost, can I restore directly from the central data centre tape to the client with no or minimal reconfiguration. I was thinking of marking media suspect.

 

[605]

Sorry, but your statement that you can not clone between storage nodes is simply wrong. A number of customers is using this feature to create copies of their saves in another building or even at a complete different location.

Although i have never done this in practice, why shouldn't it be possibe to create and link multiple staging policies? Such should not care where the data comes from (i.e. whether this is true backup data or clone data), these policies only monitor the device for fullness and stage appropriately.

 

[fraxx]

My claim was "...Although I see this as technically possible in NetWorker, since it's possible to clone between storage nodes, how feasible is it?"

I claim it is possible, not that it's not possible. And you agree it's possible.

My question is really how feasible it is from a management perspective, when trying to keep things simple for the operators and those in the help desk tasked with doing restores for single users.

 

[605]

You are right of course, i simply misread your statement.

It would not be a multi-staging policy (NW does not allow that) but multiple single-staging policies. If you only have one disk media for a pool, nothing should go wrong.

With respect to recoveries, you can recover from all media even without changing the save set to suspect. Simply use the combination of ssid/cloneid to uniquely identify the save set (and the media). In this case a so-called automatic recovery would be optimal:
recover -S ssid/cloneid -a pathname