Incus move zfs backend snapshot gone

[ckruijntjens (Chris Kruijntjens)]

Hi All,

When i move a container from one incus server to another all my zfs snapshots of that machine are gone?

I use syncoid/sanoid.

Why is incus not moving the snapshots together with the container?

[yala (jon ⚝)]

This sounds interesting. Please help me understand:

Are you using incus move to transfer instances between hosts, or do you use syncoid/sanoid to transfer the snapshots? In case of the latter, it would depend on your synchronisation strategy what is transferred.

How are you transferring the Incus metadata of an instance, that is what does your restore procedure look like? Did you play with incus move --instance-only in combination with your syncoid/sanoid setup?

I have read that Incus itself on its own will always replicate the last snapshot first, also when using --refresh*.

@stgraber

The reason is that during refresh we need to:

• Revert the target to the most recent snapshot
• Transfer any new snapshots

This is potentially to safe transfer costs, only looking to move forward.

@stgraber

The optimized transfer on top of ZFS basically refers to the use of send/receive with delta for snapshots as opposed to the basic rsync default behavior.

The migration API doesn’t expose the availability of other instances on the target, even if that could be used to further reduce the transfer and eventual dataset/volume size.

With “other instances” here, could “earlier snapshots” have been meant as well?