Make Storage Node read from Data Domain

I am not a network specialist so I can't help with your specific setup. What I can say is that - with respect to NW - there is no special setup. We have used a similar configuration for years.

You mentioned "... dedicated backup addresses". I do not know exactly what you refer to.

For further investigations it would be a good idea to use the nsrclone command where you also can explicitely define the read and write storage nodes. Just make sure that there is a unique name resolution.

Besides that, may I also suggest that you get in touch with your network specialists to trace the data transfer streams.

in the NW client definition you can set the "server interface" if you have multiple networks over which a client, in this case the NW storage node could possibly reach the nw server. in our case a nw server alsways is able to reach the nw storage nodes over dedicated backup interface. so both nw server and nw storage node are defined within the nw server with their names belonging to the backup interfaces they have.

if in your case the traffic between nw server and nw storage node is not occurring over the intended interface, then you seem to have used the frontend name/ip of the ns sn instead of its backup interface/ip? you can address that by having the nw sn communicate to a specific nw server interface by setting the required fqdn/ip in the client definition of the nw sn in the setting "server network interface". I don't know if that might break anything else for you however?

If you would want to have the nw sn directly read from the DD, then you could define a 2nd ddboost device selecting the device that already exists for the nw server and share the same volume with the backup server. so then both nw server and nw sn would be able to directly use the same ddboost volume. This of course assumes that the nw storage node can access the DD, preferably via the same fqdn the backup server does, as it will use the "device access information" to actually access the DD. It is possible to have DD's with multiple interfaces in multiple subnets and still have various components access it, even from within the same networker data zone, but to be honest I am very happy that we have the intelligence in the network and that the customer/client networks are routed towards the backup network and that the backup server and dd's are in the same subnet as well as the nw storage nodes.

Of course I know the parameter server network interface. But did you ever read the explanation? . The only help clearly states: "Server network interface - The unique hostname associated with the network interface on the server to be used for saves and recovers." - not for cloning.

AFAIR we have not used the server network interface at all. Sorry.

So using the command line and especially defining the storage nodes is at least a method worth to try.

On NetWorker (NMC > Devices), configure the same DD device with the Storage Node.
You should then see there is two devices mounted with the same Volume Name, but with slightly different Device Name.
e.g. "DDname_abc" and "rd=StorageNodeName:DDname_abc"

Then modify the clone action and make sure both Source Storage Node and Destination Storage Node are set to the same StorageNodeName.

we even share the same ddboost volume on the same NW SN even. One device I set to read-only, intended for cloning (or restores as those prefer read-only devices), while the other device is used for backups.

For better scalability as NW allowed for maximum 60 sessions on a boost device, we created for backups even up to four devices on the same NW SN sharing the same volume, so to have 240 streams available to the same pool. This was especially done for SAP-on-Oracle backups as those are greedy as the want to allocate the backup channels right-away and were notorious for leading to media alerts if not all sessions where getting a device to backup to right-away. NMDA had an option to prevent this from occurring with Oracle, while NMSAP module did not offer that, hence being able to massively scale-up into the same pool in an environment with hundreds of SAP systems, was very helpful instead of having to create many pools which each their own devices. Kept the environment rather lean and mean, even though still to be regarded as fairly large.

I also use two DDBoost interfaces for cloning to tape. The device on the storage node is set to read-only.

I have a direct connection between Storage Node (192.168.2.1) and Data Domain (192.168.2.2). This is a private network. Both, Storage Node and DD, are also with another network interface connectec to our datacenter LAN.

How can I tell NetWorker to use the private network for cloning operations? I did set the Clone Storage Node field to 192.168.2.1 in the Storage Node resource. But that did not work.

Can anyone help?

Your idea seems to follow the right track. However, before you use resources (where you never know when they will be applied), may I suggest that you use the command line where you have much more details to control a process.

In your case, you should use the nsrclone command as follows:

  nsrclone -b pool -J recover_snode -d save_snode -S ssid

These are just the most important options. Please read the Command Line Reference for all details. If this command runs as expected, you are half way there ...