The start order is used for stop, backup, start. If you put both in one group, both get stopped, then backed up and then started like the group order is. So: yes, that would be your solution
New observings: Its not happening with bridge network interface. See https://forums.docker.com/t/occasional-ping-issues-over-ipv6/140296/2?u=rkluth
@ich777 You also have IPv6 experiences with Unraids docker. Did you ever came accros such issue?
Check settings. Whats inside allowed appdata source paths? Is the volume in question maybe "external" and therefore excluded?
Some of your containers excluding /mnt which leads to empty zips..?
Point to the backup and restore all XMLs and appdatas. Use Previous apps to install all containers. Currently you need some manual work.
Another check here: I bet this container is inside of a group?
Interesting. Is there something special about /mnt/user/unraid_backup ?
Whats happening if you run "mkdir /mnt/user/unraid_backup/appdata_backup/test"?