X-Sylpheed-Account-Id:1
S:andy.sharp@onstor.com
SCF:#mh/Mailbox/sent
X-Sylpheed-Sign:0
X-Sylpheed-Encrypt:0
X-Sylpheed-Privacy-System:
RMID:#imap/andys@onstor.net@exch1.onstor.net/INBOX	0	2779531E7C760D4491C96305019FEEB52AC9100B8B@exch1.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Fri, 24 Apr 2009 10:35:36 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: Ben Piela <ben.piela@onstor.com>
Cc: dl-Leopard <dl-Leopard@onstor.com>, dl-se <dl-se@onstor.com>, Bob Condon
 <Bob.Condon@onstor.com>
Subject: Re: Outstanding questions
Message-ID: <20090424103536.01ab574b@ripper.onstor.net>
References: <ECB5BD0E84FDAC47A9062CB57901DD64E15D7E@mail.int.camcom.com>
	<2779531E7C760D4491C96305019FEEB52AC9100B8B@exch1.onstor.net>
Organization: Onstor
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

On Fri, 24 Apr 2009 09:59:41 -0700 Ben Piela <ben.piela@onstor.com>
wrote:

> Hello all,
> 
> I have a reseller that is asking a bunch of questions regarding the Leopard.  I put my answers in red but I really do not have answers for all of these.  I need to get a response back asap as always so any input would be greatly appreciated.
> 
> Thanks,
> 
> Ben Piela
> ben.piela@onstor.com
> Mobile: (201) 819-9537
> Office: (201) 251-1167
> 
> From: Sam Farmer [mailto:sfarmer@cambridgecomputer.com]
> Sent: Friday, April 24, 2009 9:16 AM
> To: Ben Piela
> Subject: Outstanding questions
>
>
> 1.       Is NFS v4 backward compatible  to NFS v3 clients?
> No.  The NFSv3 and NFSv4 protocols are not compatible. A NFSv4 client
> cannot access a NFSv3 server, and vice versa.

You are correct, the answer is 'no', but the question suggests a lack
of understanding on the part of the questioner.  The filer can provide
both v3 and v4 at the same time.  A client can also be capable of both,
or all 5 variants of NFS, but a client only connects with one of them,
and what protocol will have profound effects on usage.  v2, for
instance, is 32-bit only, so has severe size limitations in todays
storage world, v3 is 64-bit and several speed improvements over v2, but
v4 is quite different in that it adds ACLs as well as a few other
facilities.  Which implies that it is more involved to set up and
maintain.  So there is a choice to be made there for clients.

In general, if you're sharing a dataset with v4, you're doing so for a
reason, and you wouldn't want to also share it to other clients with
v3.

> 2.       Automated storage tiering: block-level or file-level migration,
> or would this depend on whether the volume is a NAS share or an iSCSUI
> LUN?
>
> This automated storage tiering is confusing to me.  From the training,
> it appeared that auto-tiering is just for migrating subfolders from
> one location to another without keeping the original file metadata.
> This seems different than what is being marketed.  My salesperson is
> under the impression that when the Leopard is configured with a SSD,
> data can be automatically moved between SSD, SAS and SATA depending on
> preconfigured policies, such as an ILM system.  Which is correct?
>
>
>
> 3.       Can a given share support NFS v3 and NFS v4 concurrently? Can a
> given share support NFS v3, NFS v4 and CIFS concurrently?

> No for the first one.  For the second one, it would be CIFS and one of
> the NFS versions.

I believe it's yes for all of them (not to the same client).  But the
akwardness of it boggles the mind.  In other words, it's probably not
something that has any real-world applications.  For any practical,
sane, real-world case, your answer is quite correct.

> 4.       Will Onstor support mixing SAS and SATA drives within a
> single 15-bay enclosure? If so, can I have a 7xSATA/8xSAS config or an
> 8xSATA/7xSAS config?
>
> I am guessing no, but I would like to hear an official answer on this.
>
>
> 5.       Will it be possible to configure dual controllers for
> active-active or active-standby mode when dual controller functionality is
> released in a few months? If active-active redundancy will be supported,
> will both controllers be able to concurrently serve up a single filesystem
> OR will true active-active functionality be contingent upon deploying
> two or more filesystems OR will active-active functionality be feasible
> with one filesystem with at least 2 shares?
>
> Does anyone have an idea as to how the HA will work?  Will it be kinda
> like a NetApp 2-node cluster?  Perhaps if we add something like AVS we
> could do the concurrent serving of a single filesystem out of both nodes.
> I would assume an active-active model would be 2 separate filesystems,
> each being served out of a different node.

Correct, separate storage pools.  Remember that "filesystem" is a
bit of a blurry term when talking about ZFS and Leopard.

> 6.       Does the Pantera LS-2150 include a native CIFS implementation
> or is CIFS merely "add-on" functionality? For example, with Windows
> Storage Server, CIFS is native though NFS is added by using the Services
> for UNIX.
>
> The CIFS service is a native kernel implementation.

Well, what can I say, there are a lot of Windows concepts that don't
really apply to the PanteraLS, so yes, it's an in-the-box
implementation. The customer won't have to go to a screen and click on
"Add Services for Windows" ~:^)
