As @Damian Andre pointed out you could try to add all StoreOnce nodes based on their own management IP and COFC identifier, then add the data paths for each StoreOnce to one Disklibrary configuration. If successful you will be able to continue backups when a path goes offline.
But I never tested a combination of different StoreOnce nodes within one disklibrary configuration.
If not successful you might want to take a closer look at the following options.
Please keep in mind, I never tested them as I do not own StoreOnce nodes, but hope it will help identify a solution for you:
- VT replication, this way data can be replicated and nodes should then in theory be interchangeable in case of node failure.
- Another idea could be, if you have VMWare for example, to use StoreOnce VSA.
This will allow you to position a “software defined” layer which can attach to StoreOnce nodes via a VM-direct path I/O config. And then publish storage towards media agents. - Depending on your model and the capabilities present you could create a multi-node StoreOnce cluster. This might involve additional hardware such as an Interlink Kit.
Not sure though how the last two options would work with an existing Commvault data path.
You might need to start a new disklibrary and auxcopy the existing data to a new copy based on the newly created disklibrary. And when finished promote the new copy as primary so you can remove the old copy, clean up the node and add it to the vsa / cluster.