[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Xen-devel] [RFC] [0/4] PV driver for FC transport layer



Hi Ian-san,

Thank you for your reply.

On Tue, 3 Jul 2007 01:07:04 +0100
"Ian Pratt" <Ian.Pratt@xxxxxxxxxxxx> wrote:

> > We developed a prototype of para-virtual FC(Fibre Channel) SCSI
> driver.
> > It's a extension of the "pv-scsi" driver, that Horikoshi-san posted on
> > 16 May 2007 ([Xen-devel] [RFC] pv-scsi driver (scsiback/scsifront)),
> > in order to support FC transport layer.
> 
> I'm struggling slightly to understand the usage scenario planned for
> this stuff. Mapping a SCSI LUN through to a guest makes perfect sense
> (e.g. for performing SCSI reservations, special SCSI commands like FUA,
> controlling a tape robot etc), but mapping a whole HBA through to a
> guest seems less useful -- usually it's the case you want to hide all
> that nastiness from the guest, taking care of multipath etc in dom0
> rather than exposing it to the guest. 
> 
> Have you a particular scenario in mind? 

We are planning to run a storage management software, which controls
bindings storages on FC network to hosts, on group of guest domains.
The software expect that each guest domain has each HBA, and control the
HBA directly. (Ex. resetting SCSI bus and getting WWN, ...)

On the other hand, we consider that NPIV (N_Port ID Virtualization)
technology makes it possible to assign whole "logical/virtual" HBA to
each gueast domain.

That's our scenario in mind.

Best regards

-----
Jun Kamada
Linux Technology Development Div.
Server Systems Unit
Fujitsu Ltd.
kama@xxxxxxxxxxxxxx



_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel


 


Rackspace

Lists.xenproject.org is hosted with RackSpace, monitoring our
servers 24x7x365 and backed by RackSpace's Fanatical Support®.