Data Backup and Recovery

SnapCenter Spanning Disks Option


We have just recently deployed SnapCenter for the first time, and transitioned our VSC backup jobs to the SnapCenter plug-in. I notice when setting up new Resource Groups, there is a Spanning Disks screen that presents the following choices (details from documentation):



Always exclude all spanning datastores [This is the default for datastores.]
Always include all spanning datastores [This is the default for VMs.]
Manually select the spanning datastores to be included



Does anyone have insight into these options, and what best practices would be? We only back up at the datastore level so the first option is default, however our transitioned VSC jobs all defaulted to the second option. Any insight would be appreciated!



Hi mate, consider this scenario:


VM1 has disk1 on DATASTORE1 and disk2 on DATASTORE2. If you select the span option, a NetApp snapshot will be taken on both volumes backing the Datastores. If you select not to span, it will only backup on the datastore/volume where most disks reside/the OS usually. If you have both DATASTORE1 and 2 in seperate jobs, there is no need to span anything because you already back it up seperately, but when it comes to restoring you would have to manually attach the disks from the flexclone restore job VSC/SnapCenter performs, so in terms of restorability you need to know more about you environment if you do not span all datastores. 


If you do span, you will be able to right click the VM > SnapCenter > Attach Virtual Disk option for e.g. and attach any disk belonging to that VM quicker. But if you set all jobs to span, and your VM/datastore layout is not good, you will end up taking multiple backups of the same data, but then again it's only a snapshot so only records what changes, so backups are small if your rate of change is small either way.


Hope this helps somewhat, how did you find deploying SnapCenter and performing backups compared to VSC out of interest? 


Thanks @HyperDiskhead! I found the actual deployment to have a few minor issues, as with any new software, but once it's been up and running it's been pretty smooth. Of course we haven't had to do a restore yet so we'll see how that goes!