[pkg-go] Bug#1121681: Future podman metadata backend deprecation/migration (nag message)

[Reinhard Tartler]

Jan Binder <jb-debbugs at nlogn.org> writes:
> documenting this upcoming issue is a good first step.
>
> I have a hard time judging how generally applicable the tool for semaphoreui 
> boltdb to sqlite migration is, I thought it was mentioned more as an example 
> of how such a migration might be done in general.

Agreed, it is a starting point at best, not a tool that is ready to use
at this point.

> In my specific case, everything except the volumes and their contents
> is easily recreated from podman-compose.

I wonder what's the story around data backups for those volumes?

> How do you arrive at the conclusion that volumes and images survive
> podman system reset?

I mispoke, I should have written "don't survive", my thinking was that
images can always be fetched from the remote registry and critical
volumes with non-volatile data are generally on NFS, host mounts or
similar storage locations that are easy to backup and restore. I realize
and acknowledge that users may have less sophisticated setups and wish I
had better answers for them.

> The documentation [0] and the warning printed by the example command therein 
> clearly state that everything will be removed.

Unfortunately, that warning is accurate.

Best,
-rt