[rescue] Fun with ZFS (was: Friend needs storage solution)

[Mike Meredith]

Digging up an ancient post :-

On Tue, 30 May 2006 12:23:05 -0500 (CDT), Jonathan C. Patschke wrote:
> uses its fstab equivalent (/etc/filesystems) to note mountpoints.  ZFS
> filesystems just get mounted where they think they should go: they
> don't show up in /etc/vfstab, and they don't show up usefully in the
> mount list.  The data is just out there...somewhere.

I haven't verified it myself, but 'zfs set mountpoint=legacy' might get
you back towards something more conventional. I must admit I like zfs
almost as much as AIX's storage management ... in some areas more
(snapshots, clones).

> filesystems thereon are separate entities.  You can't just carve a UFS
> or FAT filesystem out of a RAIDZ or use mirrored zpool for swap--the
> only things that will go on them are ZFS filesystems, and I never

Er ... not quite :-

% swap -l
swapfile             dev  swaplo blocks   free
/dev/dsk/c3t1d0s1   118,9      16 1058288 1015776
/dev/zvol/dsk/zp0/swap/0 252,1      16 2097136 2097136

Just create a fixed size 'filesystem' (zfs create -V size ...) and use
what I think is referred to as the 'emulated block device'. From the
sounds of it, you could use such a device for anything. Certainly UFS
... I just checked.

> So, this means on a large Solaris system with lots of
> compartmentalized storage, you can potentially have:
>    * Regular disklabel devices with UFS and other filesystems.
>    * SVM metadevices with UFS and other filesystems.
>    * ZFS devices with ZFS filesystems.
>    * zpool metadevices (stripes, mirrors, and permutations thereof)
>    with
>      ZFS filesystems.
>    * RAIDZ metadevies with ZFS filesystems.
> 
> I don't know about you, but I would not want to maintain such a
> system.

You've left out Veritas, QFS, EMC PowerPath devices, SunSAN devices, and
probably other stuff :)

I think you've over complicated the ZFS complexity; perhaps by being too
tied to "devices". Pools contain any number of devices each of which can
be a simple disk device (or slice or file), or a complex device being a
mirror of simple devices or a raidz collection of simple devices. ZFS
filesystems/snapshots/clones aren't tied to a device at all ... just a
pool (except with 'emulated block devices').

For me the big problems with ZFS currently are :-

* Isn't supported by NetWorker (I know, but I'm stuck with it).

* You can't reduce the size of a pool.

* Can't make / a ZFS filesystem

Both of these will be solved in time I guess.