zfs discuss - Jul 2007 - zfs iscsi storage for virtual machines

I''m going to be setting up about 6 virtual machines (Windows &
Linux) in
either VMWare Server or Xen on a CentOS 5 box. I''d like to connect to a
ZFS
iSCSI target to store the vm images and be able to use zfs snapshots for
backup. I have no experience with ZFS, so I have a couple of questions
before I move forward.

1. Is this a feasible setup? If not, is there any way to make something like
this work reliably?

2. Since I''d most likely want to restore single machines at a time, is
it
best to have a zpool for each machine?

Any insight is appreciated.

-- 
Pete
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20070716/57ef7315/attachment.html>

I had originally considered something similar, but... for ZFS snapshot
abilities, I am leaning more towards zfs-hosted NFS... Most of the other VMs
(FreeBSD, for example) can install onto NFS, it wouldn''t actually be
going
over the network, and it would allow file-level restore instead of
drive-level restore.

Just my untested 2 cents

Malachi

On 7/16/07, Peter Baumgartner <sgt.hulka at gmail.com>
wrote:
>
> I''m going to be setting up about 6 virtual machines (Windows &
Linux) in
> either VMWare Server or Xen on a CentOS 5 box. I''d like to connect
to a ZFS
> iSCSI target to store the vm images and be able to use zfs snapshots for
> backup. I have no experience with ZFS, so I have a couple of questions
> before I move forward.
>
> 1. Is this a feasible setup? If not, is there any way to make something
> like this work reliably?
>
> 2. Since I''d most likely want to restore single machines at a
time, is it
> best to have a zpool for each machine?
>
> Any insight is appreciated.
>
> --
> Pete
>
> _______________________________________________
> zfs-discuss mailing list
> zfs-discuss at opensolaris.org
> http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20070716/d3220235/attachment.html>

Peter Baumgartner wrote:
> I''m going to be setting up about 6 virtual machines (Windows &
Linux) in
> either VMWare Server or Xen on a CentOS 5 box. I''d like to connect
to a
> ZFS iSCSI target to store the vm images and be able to use zfs snapshots 
> for backup. I have no experience with ZFS, so I have a couple of 
> questions before I move forward.
> 
> 1. Is this a feasible setup? If not, is there any way to make something 
> like this work reliably?
Use some sort of data redundancy on the data.  The simplest is a mirrored
zpoool.
> 2. Since I''d most likely want to restore single machines at a
time, is
> it best to have a zpool for each machine?
I''d recommend one zpool, multiple file systems.  That way you can
manage
each file system (iSCSI target) separately, but still have the flexibility
of a large zpool.
  -- richard

zfs-discuss-bounces at opensolaris.org wrote on 17/07/2007 05:12:49 AM:
> I''m going to be setting up about 6 virtual machines (Windows &
> Linux) in either VMWare Server or Xen on a CentOS 5 box. I''d like
to
> connect to a ZFS iSCSI target to store the vm images and be able to 
> use zfs snapshots for backup. I have no experience with ZFS, so I 
> have a couple of questions before I move forward. 
> 
> 1. Is this a feasible setup? If not, is there any way to make 
> something like this work reliably?
> 
> 2. Since I''d most likely want to restore single machines at a
time,
> is it best to have a zpool for each machine? 
> 
> Any insight is appreciated.
I have hundreds of Xen-based virtual machines running off a ZFS/iSCSI 
service; yes, it''s viable. I can''t speak for CentOS
specifically; our
infrastructure is using Debian Etch with our own build of Xen.

Your success criteria are:
1. Staggering virtual machine cron entries to avoid high I/O contention,
2. Reliable gigabit/10gigabit switching infrastructure. Skimp on this and 
your project is sunk. Use a high-quality managed switch, good NICs, and 
well manufactured cables.
3. If using a smart, battery-backed, order-preserving storage array, 
append "set zfs:zfs_nocacheflush = 1" (sans quotes) to /etc/system.

Splitting your storage into multiple zpools will just cause wastage and 
administrative complexity.  I strongly advise using a single zpool.

JG



This email, including any attachments, is intended only for the use of the
individual or entity named above and may contain information that is
confidential and privileged. Any information contained in this email is not to
be used or disclosed for any purpose other than the purpose for which you
received it. If you are not the intended recipient you are notified that
disclosing, copying, distributing or taking any action in reliance on the
contents of this information is strictly prohibited. If you have received this
email by mistake, please delete this email permanently from your system.
WARNING: Although Editure has taken reasonable precautions to ensure no viruses
are present in this email, Editure can not accept responsibility for any losses
or damages whatsoever, arising from the use of this email and/or its
attachments.
www.editure.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20070717/a9421ee8/attachment.html>

>
>
> I have hundreds of Xen-based virtual machines running off a ZFS/iSCSI
> service; yes, it''s viable. I can''t speak for CentOS
specifically; our
> infrastructure is using Debian Etch with our own build of Xen.

How does ZFS handle snapshots of large files like VM images? Is replication
done on the bit/block level or by file? In otherwords, does a snapshot of a
changed VM image take up the same amount of space as the image or only the
amount of space of the bits that have changed within the image?

I''m also strongly considering going with NFS or AFS instead of iSCSI so
I
don''t have to deal with management of an extra filesystem layer. The
VMWare
community is split on which is faster. Are there any significant benefits to
either one on the ZFS side?
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
<http://mail.opensolaris.org/pipermail/zfs-discuss/attachments/20070808/1dc24590/attachment.html>

> How does ZFS handle snapshots of large files like VM images? Is 
> replication done on the bit/block level or by file? In otherwords, does 
> a snapshot of a changed VM image take up the same amount of space as the 
> image or only the amount of space of the bits that have changed within 
> the image?  
ZFS uses Copy On Write to implement snap shots.
No replication is done. When changes are made only the
blocks changed are different (the originals are kept by the
snapshot).

Neil.