[DRBD-user] Deleting deployed snapshot question

[Gianni Milo]

Don't think it's a bug. In my setup I'm using ZFS as storage backend. When
repeating steps you described, ZFS complains and it does not delete the
snapshot, as the newly created clone depends on it. This by it's turn
propagates above to the LINSTOR layer, which eventually aborts the action.
In the case you are using LVMThin, I think lvm *does* allow deleting the
snapshot, since it works differently than ZFS in this matter (data from the
deleted snapshot is automatically merged into the newly created volume, so
there's no dependency between them). You can verify if this is the case by
doing same tests directly on the lvm layer. In any case LINSTOR relies on
storage backend utils output and it does not by default prevent snapshot
removal.


On Sun, 26 May 2019 at 00:49, kvaps <kvapss at gmail.com> wrote:

> Hi
>
> I have a question and I can't fully determine is that bug or not.
>
> Example I have image A:
>
> linstor rd c A
> linstor vd c aaa 1G
> linstor r c aaa --auto-place 2
>
> now I can create new snapshot:
>
> linstor s c aaa snap1
>
> and deploy new resource from this snapshot:
>
> linstor rd c bbb
> linstor s vd rst --from-resource aaa --from-snapshot snap1 --to-resource
> bbb
> linstor s r rst --from-resource aaa --from-snapshot snap1 --to-resource bbb
>
> after that I can remove the snapshot:
>
> linstor s d aaa snap1
>
> Is that ok? Shouldn't it throw an error due having childrens?
>
>
> - kvaps
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20190526/90d86f43/attachment.htm>