As luck would have it, the NAS that died was the main file server, the most complicated one to test and restore. Backups were performed to other devices from it, and it was the main target for workstation backup. There was a secondary target to keep errors from propagating, but I didn’t want to take the time to restore from the secondary target, now that the NAS seemed happy.
On the other hand, I wasn’t going to trust the NAS box to tell me that all the data was good. Now is the time that carelessness can result in wholesale damage. I’d changed the name of the NAS server to make sure that no scripts ran against it while I was testing it, and because I don’t like to reuse names in a Windows domain. I ran all of the server-side scripts in compare mode to see if the results were reasonable. They appeared to be so. I wasn’t as careful with the workstation scripts, since errors in the target tend to be self-correcting, but I did test a few. I changes all the workstation scripts to back up to Anna, the new server. Then I changed all the server scripts to point to Anna, and put them in automatic mode.
This is about as good as a major failure in a RAID box gets. It probably would not have been anywhere near so pleasant if my hardware had become obsolete between purchase and failure, but with Synology’s one-size-fits-all software, even that might have been OK.
Leave a Reply