Understanding restoring files from Snapshots?

[jono]

I've been backing up with Snapshot Replication for a few months since I bought an new DiskStation (as well as backing up via Hyper Backup), but haven't really used it to restore files. I kind of turned it on because I had the space & left it running without 'interfering with it'.

It seems that SR is good for file versioning and quickly restoring documents, files etc, whereas HB is more suited to long term backup/archiving. Is my understanding correct?

Now that I'm needing to restore the odd few files or folders here and there I've tried doing it from Snapshot Replication, but it feels 'kinda clunky' compared to restoring files with Hyper Backup. What I mean by that is if I want to restore files with HB I open Backup Explorer for the a specific backup, pick a date or specific backup from the horizontal timeline at the bottom and either restore or copy to another location.

If I want to restore a file or folder via Snapshot Replication I go to 'Recovery' in the SR app, I see I can click on a snapshot from the list and 'Restore in-place' or 'Clone with a new name', but I never want to do this as I only want to restore a couple of files, not overwrite the whole (top level) shared folder with a previous snapshot.

I can select a snapshot from the list and click on 'Browse' instead of a restoring a full snapshot and use the File Station UI and navigate through the folders to get the files, but this feels a little clunkier than the Backup Explorer in HB.

Is that the correct way to restore a few files from SR or am I missing something?

 

[Rusty]

Is that the correct way to restore a few files from SR or am I missing something?

HB and SR are def 2 separate tools for 2 separate usage scenarios. You have understood them both so there is nothing wrong there. The reason SR feels "clunkier" is because you don't really need it in your case/usage.

SR primary usage is restoring whatever you have snaped as a whole back to its original (or alternate) destination. In most cases that will mean multiple files at once. For example if you have a flat file DB structure like MongoDB and there is a problem after you do an update of mongoDB and/or its DB's you want to restore the the previous state. This will include restoring all the files to the exact point in time, so snapshot is the best solution. Reason is that its fast, local (or remote) and its a simple one click solution.

If on the other hand you want to pull a specific file from a folder with multiple files you can do that as well with SR. Just restore the whole snap to a separate location and then handpick it. Sure this might sound clunky but its also a way to do it. If on the other hand you don't wanna do it that way then you can use HB and as you said, use the time line to search for a file that you want and pull it out of the backup.

SR has a bit of a faster response overall, it can hit multiple times and will just snap delta changes so in general for things that I personally need "hands on" I use SR for almost instant recovery. For things that I wanna restore at a slower pace (less frequent) I use HB.

Again, you have understood fine how both work its just a matter do you really need both depeding on your needs, services and up/downtime you can afford.

SR is oriented a bit more towards enterprise usage/scenarios and its not something that you have to use but ofc can use in a combination with HB.

 

[jono]

Thanks a lot for the confirmation/explanation

If on the other hand you want to pull a specific file from a folder with multiple files you can do that as well with SR. Just restore the whole snap to a separate location and then handpick it. Sure this might sound clunky but its also a way to do it.

Ah, I hadn't thought of that.

Now that I know I'm not really doing anything wrong in the way I use it I might have another 'proper' go. I really like the fact that SR is so fast compared to HB.

When looking at Synology's website and YouTube videos they recommend using Snapshot Replication as a way of backing up your data to another NAS, and haven't seen them mention HB so much for this. Do you know of any pros or cons for/against Snapshot Replication or Hyper Backup for backing up to another (offsite) NAS?

 

[Rusty]

Thanks a lot for the confirmation/explanation


Ah, I hadn't thought of that.

Now that I know I'm not really doing anything wrong in the way I use it I might have another 'proper' go. I really like the fact that SR is so fast compared to HB.

When looking at Synology's website and YouTube videos they recommend using Snapshot Replication as a way of backing up your data to another NAS, and haven't seen them mention HB so much for this. Do you know of any pros or cons for/against Snapshot Replication or Hyper Backup for backing up to another (offsite) NAS?

Well offsite backup using HB can be used to backup to other non Synology (HB) capable servers using rsync or some cloud providers you can see in the list of HB destinations.

SR on the other hand can be used only with Synology SR capable NAS on the other end (not the same model but a model that can use and run SR) running on BTRFS file system (HB doesn't have this prerequisite).

Again, reason is simple, SR is more towards business continuity so if for example on a primary node you have a service running and it goes down, you can activate that service on your second NAS and just use SR to apply the fresh data from your main NAS (for example a DB state that was running on your main nas). HB would be in this case much slower solution. You would have to restore the backup, move it into place and then start the app. Also this is in case that you have a fresh up to date state that you need, not xy hours old one. With snaps you can hit them 5min apart for example.

Saying this, again, I use both apps for different type of scenarios. Fast business continuity = SR, slower backup restores = HB.

 

[jono]

Thanks again.

I think I'll stick with Hyper Backup to the offsite NAS, maybe replicate snapshots of my main work folder to it also.

 

[Telos]

I like snapshotting because the time between snapshots is minimal where you need that.