I've recently setup Unraid on a TerraMaster NAS, with the view to it becoming the primary and the Synology NAS it replaced becoming the backup target.

I've done quite a bit of research and am conflicted in the best way to achieve a backup solution, as each way has pros and cons. I'm only looking to mirror really, version control isn't required. It's more so to have a remote/physically separated copy of the data in case there is huge failure or theft of the Unraid unit. The backup should only copy new/changed data, and delete anything that no longer exists in the source, on the target.

For example, backing up my OneDrive and Dropbox repositories Synology excels at that, with Cloud Sync in-built that 'just works'. Unraid allows you to install rclone, but most of the setup guides I've seen for that reference a plugin by waseh, which seems to be deprecated and there is now an 'official' docker container that shows up in CA. When installing the docker container it asks for a few settings such as "Rclone config" and "Data path", where it's not clear if I'm setting the path in the filesystem on the Docker container, or whether that path references the underlying Unraid filesystem. Whereas with the older waseh plugin version, it says just install that (no input required during the install unlike the newer version) and then type rclone config on the command line.

If I manage all the backups from Unraid to Synology, I've been looking at Duplicacy, Kopia and Restic. Rsync also factors in obviously (with some GUI front ends to it e.g., luckyBackup). I have AppData backup installed already, so that's taking care of the flash drive backup and Docker containers/metadata(?).

If I manage all the backups from Synology to Unraid, you can use either Hyper Backup or Active Backup for Business. With Hyper Backup for example, I think I would need to mount my Unraid share on the Synology, then essentially backup from 1 share to the other as otherwise the way the software works is more selecting another NAS as a target, rather than itself.

When I did the initial sync from my Synology to Unraid, I used the following command, but note that the date/time stamp on the files was set as the time of the sync, rather than retaining their original date/time stamp.

rsync -r --info=progress2 --info=name0 --delete-delay --size-only "/mnt/remotes/SOURCE" "/mnt/user/data/DESTINATION"

So, I see the options as:

1. Hyper Backup/ABB on Synology - mount the Unraid shares, and then backup/mirror to the same file structure on the Synology, Cloud Sync to pull down OneDrive/Dropbox.

2. Unraid crontab rsync with the command above, but in the reverse (although timestamp issue would still exist), Cloud Sync on the Synology (e.g., no mirror/sync will exist on Unraid)

3. Unraid - install rclone (to cover OneDrive/Dropbox), then either rsync as per point 2 or one of the various backup apps available (Duplicacy, Kopia, restic etc.) to the Synology.

Any thoughts? Anything else I've not considered?

I do also plug in a USB hard drive every few weeks just to give another critical items backup option, which covers things like Photos etc.

On 11/9/2025 at 6:39 AM, flashback said:

For example, backing up my OneDrive and Dropbox repositories Synology excels at that, with Cloud Sync in-built that 'just works'. Unraid allows you to install rclone, but most of the setup guides I've seen for that reference a plugin by waseh, which seems to be deprecated and there is now an 'official' docker container that shows up in CA. When installing the docker container it asks for a few settings such as "Rclone config" and "Data path", where it's not clear if I'm setting the path in the filesystem on the Docker container, or whether that path references the underlying Unraid filesystem. Whereas with the older waseh plugin version, it says just install that (no input required during the install unlike the newer version) and then type rclone config on the command line.

I'm clear on this item now, as it says Container Path: /config/rclone so in that case I'd just create /mnt/user/rclone/config and set that, ditto for Container Path: /data I'd set that to /mnt/user/rclone/data.

Does anyone have thoughts on the strategies above taking into account the pros/cons of each?

Bump

I am confused by your posts and the subject of this post. You seem to be asking two different questions.

First, if you have installed a Docker container and you have questions about that container, you should be asking itin the support thread for that Container. You can find that by clicking on the 'Support' link for that Container. See here for an example:

IF nothing else, you should name the specific Container that you are having issues with.

I don't use any container to do exactly what you seem to be wanting to do. I use rsync however. I use the Unassigned Devices plugins. (All three of them so that I had tmux so that I could shut down Unraid's GUI terminal and allow rsync to run in the background. You can find info on tmux by googling.). It is necessary to have rysnc in stalled on both devices. This is not normally not a problem as rysnyc has been ported to virtually every OS in existence. The rsync command is formatted like this:

rsync -Options source destination

It works in either a push or pull mode. If the source is remote, you are pulling. If the source is local, you are pushing. You can find much info and tutorials on rsync on the web. I know it might seem bit overwhelming initially but it is fairly easy to use once you do a bit of research. (If you can think of a set of conditions that you want to apply rsync can probably implement them...) Simple task are very easy to implement as there are some switches (like "-a") that combine many common backup features into a single switch!)

I use the Unassigned Devices to mount the remote server. (This plugin does all of the really confusing stuff behind the scene which make the use of rsync so much easier!)

My command line looks like this. I have it stored in a text file on my Windows desktop. When I want to do a backup, I copy-and-paste it from the file to the GUI terminal and run it. (If I run it in tmuc session, I can then close the GUI terminal and even shut the Windows machine down and rsync will continue to run.)

rsync -avhPX /mnt/remotes/ELSIE1_Media/All\ Movies/ /mnt/user/BackupMedia/All\ Movies/

In this case, it is pull operation since the source is on the remote machine. (It is really hard to say what is a 'pull' vs a 'push' operation with rsync since rsync is running on both machines. I believe it is more in the mind of the user rather than true definition.)

PS: In my command, only new and changed files are copied so it does an incremental backup each time it is run. It does display the progress on the terminal screen so you can watch what is happening. There is a 'switch' which will allow you to delete files in the distinction that are not in source. (By default, it does not auto-delete files so a bit of stupidity by the user on the source does repeat itself in the backup!)

3 minutes ago, Frank1940 said:

I am confused by your posts and the subject of this post. You seem to be asking two different questions.

First, if you have installed a Docker container and you have questions about that container, you should be asking itin the support thread for that Container. You can find that by clicking on the 'Support' link for that Container. See here for an example:

IF nothing else, you should name the specific Container that you are having issues with.

I don't use any container to do exactly what you seem to be wanting to do. I use rsync however. I use the Unassigned Devices plugins. (All three of them so that I had tmux so that I could shut down Unraid's GUI terminal and allow rsync to run in the background. You can find info on tmux by googling.). It is necessary to have rysnc in stalled on both devices. This is not normally not a problem as rysnyc has been ported to virtually every OS in existence. The rsync command is formatted like this:

rsync -Options source destination

It works in either a push or pull mode. If the source is remote, you are pulling. If the source is local, you are pushing. You can find much info and tutorials on rsync on the web. I know it might seem bit overwhelming initially but it is fairly easy to use once you do a bit of research. (If you can think of a set of conditions that you want to apply rsync can probably implement them...) Simple task are very easy to implement as there are some switches (like "-a") that combine many common backup features into a single switch!)

I use the Unassigned Devices to mount the remote server. (This plugin does all of the really confusing stuff behind the scene which make the use of rsync so much easier!)

My command line looks like this. I have it stored in a text file on my Windows desktop. When I want to do a backup, I copy-and-paste it from the file to the GUI terminal and run it. (If I run it in tmuc session, I can then close the GUI terminal and even shut the Windows machine down and rsync will continue to run.)

rsync -avhPX /mnt/remotes/ELSIE1_Media/All\ Movies/ /mnt/user/BackupMedia/All\ Movies/

In this case, it is pull operation since the source is on the remote machine. (It is really hard to say what is a 'pull' vs a 'push' operation with rsync since rsync is running on both machines. I believe it is more in the mind of the user rather than true definition.)

PS: In my command, only new and changed files are copied so it does an incremental backup each time it is run. It does display the progress on the terminal screen so you can watch what is happening. There is a 'switch' which will allow you to delete files in the distinction that are not in source. (By default, it does not auto-delete files so a bit of stupidity by the user on the source does repeat itself in the backup!)

Thanks for your comments, my 2nd post clarified that I have resolved the issue/query I had with the 'rclone' official Docker container that's available via CA. Having had a look at it, it feels like using Cloud Sync on my Synology target is going to be nicer to use.

Did you have any thoughts on why my rsync command isn't keeping the original timestamp on the files/folders?

When I did the initial sync from my Synology to Unraid, I used the following command, but note that the date/time stamp on the files was set as the time of the sync, rather than retaining their original date/time stamp.

rsync -r --info=progress2 --info=name0 --delete-delay --size-only "/mnt/remotes/SOURCE" "/mnt/user/data/DESTINATION"

Would using -a (archive) or -X (preserve extended attributes) fix that?

IF you do a bit of Googling on rsync, you will find that the "-a" option is the one most widely used because it does what most people want--- Copy all new or updated files. It preserves most of the source's original properties of each resource as you would expect in an archive. The "-v" tells it to verbose as it does each step. (You probably would not want this if you are using it in a shell script!) The "-h'" option makes the numbers easily human readable. The "-P" option keeps any partially transferred file(s) on the destination so they can be completed later. (This is in case you have a 'flakey' and, perhaps, slow connection situation. If you transfer 90% of the file before interruption, you only have to transfer the remaining 10% on the next pass. Usually not an issue for most Unraid users...) The '-X' option is for preserving the extended attributes. This is the one option that I don't think is really needed unless your Unraid server is running in an Active Directory environment. (Active Directory is part of Microsoft's server/networking package.) My first use of rsync was when I switched from reiserfs file system to XFS file system and I was using following a set of 'How-to' instructions. I just reused the same set of switches for my backups as i did for the conversion.

IF you haven't found it, here is a link to the manual page for rsync:

https://linux.die.net/man/1/rsync

IF you want to try out rsync, setup a small transfer of small part of the data set and see if it delivers what you want. Then you can play around with the various parameters and see that it is doing what you want.

In my initial backup of my one server to the other I moved almost 20TB of data using rsync. I did break it up into more manageable chunks of about 1TB each but looking back, doing it in one move would have been simpler and perhaps faster! (The only caution here is how much bandwidth rsync might being consuming if you try to use either server for something else.)

In a more general note, I do have my backup server set up to use 'reconstruct write' mode and I do not have a cache drive for the backup shares. The cache drive (500GB) could easily fill up when the backup is being done. These settings often allow the backup speed to be about 100MB/s depending on file size.

12 minutes ago, Frank1940 said:

IF you do a bit of Googling on rsync, you will find that the "-a" option is the one most widely used because it does what most people want--- Copy all new or updated files. It preserves most of the source's original properties of each resource as you would expect in an archive. The "-v" tells it to verbose as it does each step. (You probably would not want this if you are using it in a shell script!) The "-h'" option makes the numbers easily human readable. The "-P" option keeps any partially transferred file(s) on the destination so they can be completed later. (This is in case you have a 'flakey' and, perhaps, slow connection situation. If you transfer 90% of the file before interruption, you only have to transfer the remaining 10% on the next pass. Usually not an issue for most Unraid users...) The '-X' option is for preserving the extended attributes. This is the one option that I don't think is really needed unless your Unraid server is running in an Active Directory environment. (Active Directory is part of Microsoft's server/networking package.) My first use of rsync was when I switched from reiserfs file system to XFS file system and I was using following a set of 'How-to' instructions. I just reused the same set of switches for my backups as i did for the conversion.

IF you haven't found it, here is a link to the manual page for rsync:

https://linux.die.net/man/1/rsync

IF you want to try out rsync, setup a small transfer of small part of the data set and see if it delivers what you want. Then you can play around with the various parameters and see that it is doing what you want.

In my initial backup of my one server to the other I moved almost 20TB of data using rsync. I did break it up into more manageable chunks of about 1TB each but looking back, doing it in one move would have been simpler and perhaps faster! (The only caution here is how much bandwidth rsync might being consuming if you try to use either server for something else.)

In a more general note, I do have my backup server set up to use 'reconstruct write' mode and I do not have a cache drive for the backup shares. The cache drive (500GB) could easily fill up when the backup is being done. These settings often allow the backup speed to be about 100MB/s depending on file size.

Yes, when I did the initial sync from my older Synology to my new Unraid box, I used rsync. 2.5GbE on either side meant it handled the 10TB of data in reasonably good fashion. I thought I had made it clear in my original post, but I'm not coming at this completely fresh and have tried a lot of things in the background. Given rclone doesn't function as nicely, I think I'm still going to backup OneDrive and Dropbox using Cloud Sync on the Synology. Again, given the lack of simple reporting and scheduling on the Unraid side, it seems I'm better to also use Active Backup for Business on the Synology side and pull the data down that way to backup.

The new Unraid box is primary and it will be backing up to the older Synology NAS it replaced.

Ok, I ended up using rsync -avhP and that reseeded things with the correct date/timestamp, although it seemed to actually copy the data again when it did it rather than just update the timestamp on the file.

I think if I add --delete-delay that will then remove files on the backup destination (Synology) from the source (Unraid) anything that has been deleted on the Unraid data share.

I should probably mention as well, I've been running mine using screen - which I added via a non-CA route. Any debate on why you prefer tmux over screen for doing simple 1 line commands like this?

Edited November 14, 2025Nov 14 by flashback

26 minutes ago, flashback said:

Any debate on why you prefer tmux over screen for doing simple 1 line commands like this?

Not really. Except that tmux is a part of a LimeTech Plugin that is supported on an active basis so it is quick and easy for most folks to install. I used screen many years ago when it was a part of tool set plugin (not longer supported) and it worked just fine. I prefer to suggest the Unassigned Devices plugin as I don't want to have to figure out how to find a non CA plugin, and then write up instructions to get and install it. (Plus, a fair number of folks already have Unassigned Devices installed..)

I can remember back when there was no Docker in Unraid and everything had to be added via the plugin route. There were always problem with common dependencies (probably wrong term) being updated to a new version which would either break Unraid or a plugin. Those memories make me very hesitant to recommend any non-support plugin.

1 hour ago, Frank1940 said:

Not really. Except that tmux is a part of a LimeTech Plugin that is supported on an active basis so it is quick and easy for most folks to install. I used screen many years ago when it was a part of tool set plugin (not longer supported) and it worked just fine. I prefer to suggest the Unassigned Devices plugin as I don't want to have to figure out how to find a non CA plugin, and then write up instructions to get and install it. (Plus, a fair number of folks already have Unassigned Devices installed..)

I can remember back when there was no Docker in Unraid and everything had to be added via the plugin route. There were always problem with common dependencies (probably wrong term) being updated to a new version which would either break Unraid or a plugin. Those memories make me very hesitant to recommend any non-support plugin.

Absolutely!

I'd already using the Unassigned Devices plugin, which works well for mounting the shares from the Synology NAS in r/o mode.

It sounds like that's the next thing to do on my list, investigate how to spin up a Docker container for one that does not show up in CA e.g., https://bahmni.atlassian.net/wiki/spaces/BAH/pages/3117744129/Getting+Started+Quickly+with+Bahmni+on+Docker