Using the GUI from Unraid getting an error on the directory i want to backup

[acrusso]

I am backing up my “appdata” zfs backups on a diff drive to google drive

within the docker console i can browse and list the directory but the gui log says its skipping, not sure where to go - i even chmod 777 hte directory to see and its the same result.

2023-09-27 14:03:43.058 INFO REPOSITORY_SET Repository set to /1_appdata
2023-09-27 14:03:43.058 INFO STORAGE_SET Storage set to gcd://ServerBackups/AppDataBackup_Duplicacy
2023-09-27 14:03:47.054 INFO BACKUP_START Last backup at revision 4 found
2023-09-27 14:03:47.129 INFO BACKUP_INDEXING Indexing /1_appdata
2023-09-27 14:03:47.129 INFO SNAPSHOT_FILTER Parsing filter file /cache/localhost/0/.duplicacy/filters
2023-09-27 14:03:47.129 INFO SNAPSHOT_FILTER Loaded 6 include/exclude pattern(s)
2023-09-27 14:03:47.130 WARN LIST_FAILURE Failed to list subdirectory cache_appdata/: lstat /1_appdata/cache_appdata/duplicacy: input/output error
2023-09-27 14:03:48.747 WARN SKIP_DIRECTORY Subdirectory cache_appdata/ cannot be listed

[saspus]

Can you list the subdirectory it actually complains about /1_appdata/cache_appdata/?

I/O failure is not permissions, it’s media failure.

[acrusso]

yes (this is from the docker console)

[saspus]

Interesting.

I’m not familiar with unraid, but general Linux guidance:

• check who is the owner of these files (ls -l) as seen from within the container vs as seen from the host. Ensure the user ID are the same, if using Linux permissions. If SELinux is enabled — they will be in different namespaces.
• if unraid is using ACLs then Linux permissions won’t matter, and you would need to ensure the ACLs are propagated though-out the directory structure.

The IO errors seems to indicated it’s something outside of permissions that the os in the container assumes is media failure.

[acrusso]

hmm strange i restarted the container and it seems to be backing up now, my apologies

[saspus]

This is the worst… now we don’t know what it was and how to fix it when it happens again.

Is unraid using docker or podman? Or maybe you want to consider running duplicacy natively, to avoid unnecessary dependencies and failure points like these in additional layers. Duplicacy is already monolithic go program, it does not benefit from containerization.

[acrusso]

i put it in unraid in the docker because that box is 24/7 and its the only one that stays on in my environment, i may be able to put it on a small VM

[saspus]

I’m confused. If docker is up 24/7 then the host that runs docker is too. I assume it’s unraid. Unraid is Linux, and hence you can run duplicacy as a service on unraid with its systemd daemon.

in comments to this article there is a sample systemd config for duplicacy: Duplicacy Web on Synology Diskstation without Docker | Trinkets, Odds, and Ends