Upgrading from 1.4-rolling-202210020218 to 1.4-rolling-202302050317 removes my container configuration

Tested running an update via:

add system image https://s3-us.vyos.io/rolling/current/vyos-1.4-rolling-202302050317-amd64.iso
set system image default-boot 1.4-rolling-202302050317
reboot

When the system came up my container section was entirely missing from the config. I rolled back to the previous release I was using, 1.4-rolling-202210020218 and now see an error when running podman commands:

ERRO[0000] User-selected graph driver "vfs" overwritten by graph driver "overlay" from database - delete libpod local files to resolve.  May prevent use of images created by other tools

And here’s my show container output from configuration mode:

 name adguardhome {
     cap-add net-bind-service
     description "Adguardhome DNS"
     image adguard/adguardhome:v0.107.23
     memory 4096
     network adguard {
         address 172.29.0.18
     }
     restart always
     volume conf {
         destination /opt/adguardhome/conf
         source /usr/lib/live/mount/persistence/volumes/adguardhome/conf
     }
     volume hostfile {
         destination /etc/hosts
         source /etc/hosts
     }
     volume work {
         destination /opt/adguardhome/work
         source /usr/lib/live/mount/persistence/volumes/adguardhome/work
     }
 }
 name unifi {
     environment DB_MONGO_LOCAL {
         value false
     }
     environment DB_MONGO_URI {
         value mongodb://172.29.0.34:27017/unifi
     }
     environment STATDB_MONGO_URI {
         value mongodb://172.29.0.34:27017/unifi_stat
     }
     environment TZ {
         value EST
     }
     environment UNIFI_DB_NAME {
         value unifi
     }
     image goofball222/unifi@sha256:df7f59207ca00a855dd3d9ca3a3913900e4cf21b567fdfb108a90861674b8034
     memory 2048
     network unifi {
         address 172.29.0.37
     }
     restart always
     volume data {
         destination /usr/lib/unifi/data
         source /usr/lib/live/mount/persistence/volumes/unifi/data
     }
 }
 name unifidb {
     image mongo@sha256:19c11a8f1064fd2bb713ef1270f79a742a184cd57d9bb922efdd2a8eca514af8
     network unifi {
         address 172.29.0.34
     }
     restart always
     volume db {
         destination /data/db
         source /usr/lib/live/mount/persistence/volumes/unifi/db
     }
 }
 network adguard {
     prefix 172.29.0.16/28
 }
 network unifi {
     prefix 172.29.0.32/28
 }

I do not mind redoing my container configuration, but I am not sure what the correct place to put persistent volume storage is for use with containers. I thought I was using a persistent mounted disk but the directories I specified for my container volumes are also missing along with the container config.

No rush, appreciate any input!

Driver was changed in T4870

It seems no another way to change driver without deleting images/etc

How much disk space does your VyOS have? I implemented a way in archiving and restoring container imGes when filesystem driver is changed. This will only work on sufficient diskspace. Let me take a look at this

I see, this makes a lot of sense, I should have checked the changelogs more closely. Though it also appears that there were changes added to mitigate the effects of the change, those appear to have not worked for me. Oh well, it is not a problem for me to re-do the configuration.

My only issue now is that when I roll forward, the contents of /usr/lib/live/mount/persistence/volumes (which is a directory I created) appear to be wiped out. I will check out the migration scripts and see if I can grok it.

I have a lot of extra space:

/dev/nvme0n1p3  469G   22G  423G   5% /usr/lib/live/mount/persistence