Restic backups - can fail on lost+found directory

[tesshuflower]

Describe the bug

Restic backups may fail due to lost+found directory created for ext4 filesystems. Should exclude this dir from restic backups.

Steps to reproduce

Expected behavior

Actual results

Additional context

[tesshuflower]

Reported by @erenfro .

Interestingly, we do have ext4 in use in e2e tests that run on OpenShift in AWS, however they haven't caught this. My best guess is it's because in OpenShift fsGroup is set automatically and lost+found ends up being readable by the group.

@erenfro Do you currently set fsGroup in your spec.restic.moverSecurityContext?

[erenfro]

I don't generally set that, but despite that, lost+found technically should always ever be root:root owned, as this is an innate part of the filesystem design itself. It simply should not be included in the backup regardless of what permissions it has.

[erenfro]

And actually, by default, yes, it's actually set in my moverSecurityContext template I use.

[tesshuflower]

And actually, by default, yes, it's actually set in my moverSecurityContext template I use.

thx for the confirmation

[erenfro]

For reference, I use this setup for my VolSync, in a templated fashion:
Template: https://github.com/erenfro/homelab-flux/tree/main/kubernetes/templates/volsync
Application: https://github.com/erenfro/homelab-flux/tree/main/kubernetes/apps/selfhosted/mealie