Duplicacy can deduplicate those uncompressed tar files. There are several users who discussed about this.

So after struggling for a few days to figure out why the plugin kept erroring out on a couple of my containers (specifically binhex-krusader and qBittorrent) I finally discovered the root cause.

They both have mappings to /mnt/user, for obvious reasons.

Having /mnt/user as an exclusion will cause the plugin to ignore everything in /mnt/user, which includes the container's appdata folder, even though the appdata folder is explicitly chosen to be backed up.

For anyone else who comes across this issue, the proper way to handle a container with a /mnt/user mapping is to not set an exclusion for those but instead set "Save External Volumes" to No.

Also on version 7.2.0, I'm assuming due to the new responsive WebUI backend, the option boxes for the plugin are a bit discombobulated, so make sure you're setting the correct option (the middle one) to No.

4 hours ago, KluthR said:

Duplicacy can deduplicate those uncompressed tar files. There are several users who discussed about this.

Sure it can, but there would be no point. You would just waste IO and resources.

Correct me if I am wrong but as far as I know deduplicating tools like Duplicacy/Restic/Borg work on a block/chunk level meaning they will identify identical chunks of data and store only the differences between backups. This only works efficiently if the underlying binary data remains aligned and unchanged. With .tar files each new backup folder appears entirely different to the deduplication tool, meaning you’ll end up re-backing up almost all of the same data and deduplication is practically pointless. But please educate me if I am missing something. I don't think I can find that particular conversation in 82 pages of posts easily 😂

Update: You might be right- This post confirms what you are saying, though I haven't tested it. Though this post suggests it might be sub-optimal to do so.

Edited November 12, 2025Nov 12 by Bob-omb

2 hours ago, CMDMC12 said:

There is absolutely no reason to backup mnt/user That is what causing problems. I assume you have added that path to the docker container, but you should remove it. There is no need to map it. The two other paths are fine though.

16 hours ago, CMDMC12 said:

They both have mappings to /mnt/user, for obvious reasons.

Normally no container should have a mapping to the whole of /mnt/user, only the required shares under it.

Hi guys, I keep trying to create a backup group for two Docker containers, and it says, "type something in to create one" - however, I have no idea where exactly I should type this. I guess I'm blind or something. Can anyone help me out?

In the group field, type the group name you want instead of selecting one.

There is no group field for me, or maybe I just can't see it:

It's where it shows ""None - Double click for a list".

Layout is messed up on 7.2, normally the fields should be at the right of the labels.

Hopefully @KluthR will find time to do the 7.2 fixes at some point

Edited November 17, 2025Nov 17 by Kilrah

On 11/11/2025 at 11:17 AM, JTVUS said:

Good news: backup was successfully completed this morning!

I am interested in still having app updates automatically applied. I may try to see if the problem with the updates is connected with the stopped / non-auto-start apps. Especially (VPN linked) Bazaar in my case. I have removed Bazaar, set Handbrake to auto-start. And turned on updates in this plugin. I will see if it works next week.

Just to close the loop on my issue in case anyone else is having it: It does appear to be the update setting.

I apparently forgot to save the setting change last week, so the backup ran successfully this morning without automatically applying updates. When I changed the setting to automatically update and ran a manual backup, it had the same errors starting the Docker apps again. So I will just keep the auto-update setting off for now. Thanks all.

Just a quick THANK YOU for this plugin. Had the file system on my NVMe drive become corrupted, and after spending some time attempting to repair it, I needed to take the nuclear approach of reformatting the drive and recreating all of my docker containers.

I had almost 40 containers at the time of the crash. I have been meaning to remove a number of them, as they are no longer used. I ended up reinstalling 25. Every one started right up, and I didn't need to do any reconfiguration (short of stopping/restarting the array and Docker service).

It would have been a complete nightmare if it wasn't for this plugin. Kudos for providing and maintaining this plugin.

Hi there, I'm a bit unsure about the proposed directories to exclude. You mention:

• Library/Application Support/Plex Media Server/Cache

• Library/Application Support/Plex Media Server/Media

• Library/Application Support/Plex Media Server/Metadata

Cache I feel safe exluding as it would be a terrible decision of the software developers to put irrecoverable files in a dir called Cache, however, I cannot find any official statement from Plex that it is safe to exclude the Media or Metadata dirs. I'm most interested in Media since it is several hundred gigabytes on my system. What do I actually lose by excluding Media? I think it would be an improvement if you could add a bit of information about that in the main post where exclusions are mentioned! : )

I thought I had this working properly since it was creating backup files. Finally needed to actually use it and turns out it's only backing up 5 of my 40 containers and of course not the one I need to restore. How do I go about trying to figure out why it's not backing up everything else. The log window in the app tab says that the normal and debug logs are empty.

Check the settings for each container.

8 hours ago, Kilrah said:

Check the settings for each container.

Yeah, that's zero help. Check what? They're set to skip - No

14 hours ago, shiftylilbastrd said:

The log window in the app tab says that the normal and debug logs are empty

Do a manual run and check both logs for hints why some are being skipped

6 hours ago, shiftylilbastrd said:

Yeah, that's zero help. Check what? They're set to skip - No

Click on the container name, that'll expand the settings, check the paths are seen as internal/external correctly, etc...

I continue to have errors like the following:

[PostgreSQL_Immich] tar verification failed! Tar said: tar: Removing leading `/' from member names; mnt/user/appdata/PostgreSQL_Immich/base/16384/243355: Contents differ; mnt/user/appdata/PostgreSQL_Immich/base/16384/17700: Contents differ

But have not been able to solve them. I have tried turning off compression on when the back up is run but still compresses the files and receives this error.

Anyone have any ideas on how to stop this error or get the plugin to stop compressing the data?

This usually means something accessed the file during the backup.

2 hours ago, Kilrah said:

This usually means something accessed the file during the backup.

It's there anything you recommend to stop this? Containers are set to stop before backing up.

Edited December 5, 2025Dec 5 by BeakerFl

Hey @KluthR have you noticed the broken layout since 7.2?

Sure but I wont fix it. :D

I would accept PRs on github

I've got this functioning properly on the majority of my containers. I am having an issue with Plex and Jellyfin though. I have both of those mapped to their own share outside of the appdata share but none of the internal volumes in those containers are getting detected. I have the following in appdata sources block.
/mnt/user/appdata

/mnt/user/jellyfin

/mnt/user/plex

But

Plex setting

Jellyfin settings

I do see the following about source path in the logs for both containers but not sure how to address it.

Edited December 13, 2025Dec 13 by shiftylilbastrd
Add info from log

1 hour ago, shiftylilbastrd said:

/mnt/user/jellyfin

/mnt/user/plex

You should not have those. Like the default /mnt/user/appdata that's supposed to be the parent folder of where appdata is located.

Why not have appdata in the default share? Media you sure want separate, but...

58 minutes ago, Kilrah said:

Why not have appdata in the default share?

They're in their own share and then have exclusive access to their own cache disks. Made my libraries way more responsive. Besides, isn't that the entire purpose to be able to add multiple appdata locations?

Edited December 13, 2025Dec 13 by shiftylilbastrd