[pve-devel] ceph osd over zfs volume

[Dietmar Maurer]

> > I don't want to relax that requirement now. For easy maintenance, it
> > is very important that we can replace whole disks.
> 
> It is difficult to believe that using slices (or partitions) makes administration
> difficult.

It is also difficult to believe that zfs is useful here. 

> Consider that using zvols and their snapshotting features gives more options for
> recovering and more flexibility for allocating resources where needed,
> particularly in small clusters. It also gives more options for caching on ssds and
> for dealing with the sync writes.
> These might or not might be useful depending on the application.
> 
> I do not want or need a gui, but I also don't like the "one solution fits all"
> approach taken here. 

I suggest you use a separate cluster for ceph then. Or simply use the tools provided by ceph
if you do not need a GUI anyway.

> To me this is a limiting factor without a tangible advantage.

We do not set any limits on ceph - you can use the original packages like you want.