Re: [SPDK] iSCSI initiator

Re: [SPDK] iSCSI initiator

[Walker, Benjamin]

[-- Attachment #1: Type: text/plain, Size: 960 bytes --]

On Wed, 2017-11-01 at 07:02 +0300, Иванов Алексей wrote:
> Hello!
> As i understand SPDK can realize iSCSI target application. But is it possible
> to use it as iSCSI initiator?

There is no iSCSI initiator in SPDK today, so if you're just looking for a
userspace iSCSI initiator that already exists see
https://github.com/sahlberg/libiscsi.
I'm not familiar with this project beyond a cursory glance, but it seems to fit
the bill.

That said, there is both an NVMe-oF initiator and a virtio-scsi initiator in
SPDK already, so an iSCSI initiator would definitely fit into the toolkit. In
order to consider adding one, I think we have to figure out how an SPDK
implementation would be fundamentally better than the already existing
implementations (i.e. libiscsi). If anyone has any insight into that
implementation or any other, and has thoughts on how SPDK could do a faster one,
it would be much appreciated.

Thanks,
Ben


[-- Attachment #2: smime.p7s --]
[-- Type: application/x-pkcs7-signature, Size: 3274 bytes --]

Re: [SPDK] iSCSI initiator

[Harris, James R]

[-- Attachment #1: Type: text/plain, Size: 1507 bytes --]


> On Nov 1, 2017, at 9:47 AM, Walker, Benjamin <benjamin.walker(a)intel.com> wrote:
> 
> On Wed, 2017-11-01 at 07:02 +0300, Иванов Алексей wrote:
>> Hello!
>> As i understand SPDK can realize iSCSI target application. But is it possible
>> to use it as iSCSI initiator?
> 
> There is no iSCSI initiator in SPDK today, so if you're just looking for a
> userspace iSCSI initiator that already exists see
> https://github.com/sahlberg/libiscsi.
> I'm not familiar with this project beyond a cursory glance, but it seems to fit
> the bill.
> 
> That said, there is both an NVMe-oF initiator and a virtio-scsi initiator in
> SPDK already, so an iSCSI initiator would definitely fit into the toolkit. In
> order to consider adding one, I think we have to figure out how an SPDK
> implementation would be fundamentally better than the already existing
> implementations (i.e. libiscsi). If anyone has any insight into that
> implementation or any other, and has thoughts on how SPDK could do a faster one,
> it would be much appreciated.

I think libiscsi should work well.  it has async interfaces so it fits the SPDK asynchronous polling model.

Building an iSCSI bdev module on top of libiscsi should be fairly straightforward.  This would be really useful for presenting iSCSI LUNs to VMs via vhost.

> 
> Thanks,
> Ben
> 
> _______________________________________________
> SPDK mailing list
> SPDK(a)lists.01.org
> https://lists.01.org/mailman/listinfo/spdk

Re: [SPDK] iSCSI initiator

[Yang, Ziye]

[-- Attachment #1: Type: text/plain, Size: 763 bytes --]

Hi ,

I think you can do that. For example, currently, SPDK implements virtio scsi driver, though it is virtio scsi driver, but the idea can be similar.

Virtio SCSI driver

A userspace driver for Virtio SCSI devices has been added. The driver is capable of creating block devices on top of LUNs exposed by another SPDK vhost-scsi application.

From: SPDK [mailto:spdk-bounces(a)lists.01.org] On Behalf Of Иванов Алексей
Sent: Wednesday, November 1, 2017 12:02 PM
To: spdk(a)lists.01.org
Subject: [SPDK] iSCSI initiator

Hello!
As i understand SPDK can realize iSCSI target application. But is it possible to use it as iSCSI initiator?

--
Отправлено из мобильного приложения Яндекс.Почты

[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 4384 bytes --]

[SPDK] iSCSI initiator

[@ 2017-11-01 4:02 UTC (permalink / raw)]

[-- Attachment #1: Type: text/plain, Size: 0 bytes --]



[-- Attachment #2: attachment.html --]
[-- Type: text/html, Size: 237 bytes --]