Configured nqn

Configured nqn

[Gili Buzaglo]

Hi
I was wondering if there's a special reason why not to support preconfigured initiator nqn from a file.
(similar to iscsi in /etc/iscsi/initiatorname.iscsi).

For example libvirt iscsi volume driver in nova in openstack, uses the iscsi file to retrieve the initiator name and send it
To cinder to allow access.
We thought it would be required to support nvmf in the same manner.

Thanks
-gili

CONFIDENTIAL COMMUNICATION:  This communication including any attached file(s) is CONFIDENTIAL, intended only for the named recipient (s) above and contains trade secrets or other information that is exempt from disclosure under applicable law.  Any use, dissemination, distribution or copying of this communication and any file attachments by anyone other than the named recipient (s) is strictly prohibited. If you have received this communication in error, please immediately notify us by reply electronic mail and permanently delete the original transmission and any attachments from your computer system and any copies of this message and file attachments. Thank you.

Configured nqn

[Sagi Grimberg]

> Hi

Hey Gili,

> I was wondering if there's a special reason why not to support preconfigured initiator nqn from a file.
> (similar to iscsi in /etc/iscsi/initiatorname.iscsi).

It was raised on the list just last week. I think Jay
has (of was :)) volunteered.

> For example libvirt iscsi volume driver in nova in openstack, uses the iscsi file to retrieve the initiator name and send it
> To cinder to allow access.
> We thought it would be required to support nvmf in the same manner.

Its definitely useful and easy enough to do. You are welcome to send
a patch to nvme-cli ;)

Configured nqn

[J Freyensee]

On Mon, 2016-08-08@11:47 +0300, Sagi Grimberg wrote:
> > 
> > Hi
> 
> Hey Gili,
> 
> > 
> > I was wondering if there's a special reason why not to support
> > preconfigured initiator nqn from a file.
> > (similar to iscsi in /etc/iscsi/initiatorname.iscsi).
> 
> It was raised on the list just last week. I think Jay
> has (of was :)) volunteered.

I was going to do it, but I was waiting for a reply on confirming what
behavior people wanted to see:


"So is the request to have nvme-cli use whatever is in
/etc/nvme/hostnqn (if it exists) over the default in the nvme-fabrics
driver???So preference of hostnqn input would be:

1. --hostnqn flag
2. /etc/nvme/hostnqn
3. default hostnqn generated by nvme-fabrics driver (ex: nqn.2014
-08.org.nvmexpress:NVMf:uuid:a2d7752c-a31b-477a-a003-31a5e1c424a9)

(note 2. may contain a hostnqn different from the default)"


Gili if you are itchin' to do it, go ahead and give it a shot, just let
me know (or I can still do it w/confirmation to above). ?The discovery
fix I added to nvme-cli:

http://lists.infradead.org/pipermail/linux-nvme/2016-July/005562.html

has been merged into the nvme-cli tool by Keith.

Configured nqn

[Gili Buzaglo]

Hi
Sorry for responding late.
My original thought was that the nvme-fabrics driver will initiate the nqn from the file,
But your suggestion is also suitable and less complex.
Who is responsible for creating the file  /etc/nvme/hostnqn ?
Maybe nvme-cli first installation ?

Thanks
-gili

-----Original Message-----
From: J Freyensee [mailto:james_p_freyensee@linux.intel.com] 
Sent: Monday, August 08, 2016 8:55 PM
To: Sagi Grimberg; Gili Buzaglo; linux-nvme at lists.infradead.org
Subject: Re: Configured nqn

On Mon, 2016-08-08@11:47 +0300, Sagi Grimberg wrote:
> > 
> > Hi
> 
> Hey Gili,
> 
> > 
> > I was wondering if there's a special reason why not to support 
> > preconfigured initiator nqn from a file.
> > (similar to iscsi in /etc/iscsi/initiatorname.iscsi).
> 
> It was raised on the list just last week. I think Jay has (of was :)) 
> volunteered.

I was going to do it, but I was waiting for a reply on confirming what behavior people wanted to see:


"So is the request to have nvme-cli use whatever is in /etc/nvme/hostnqn (if it exists) over the default in the nvme-fabrics driver???So preference of hostnqn input would be:

1. --hostnqn flag
2. /etc/nvme/hostnqn
3. default hostnqn generated by nvme-fabrics driver (ex: nqn.2014
-08.org.nvmexpress:NVMf:uuid:a2d7752c-a31b-477a-a003-31a5e1c424a9)

(note 2. may contain a hostnqn different from the default)"


Gili if you are itchin' to do it, go ahead and give it a shot, just let me know (or I can still do it w/confirmation to above). ?The discovery fix I added to nvme-cli:

http://lists.infradead.org/pipermail/linux-nvme/2016-July/005562.html

has been merged into the nvme-cli tool by Keith.

Configured nqn

[J Freyensee]

On Tue, 2016-08-16@14:08 +0000, Gili Buzaglo wrote:
> Hi
> Sorry for responding late.
> My original thought was that the nvme-fabrics driver will initiate
> the nqn from the file,
> But your suggestion is also suitable and less complex.
> Who is responsible for creating the file??/etc/nvme/hostnqn ?
> Maybe nvme-cli first installation ?

Yes, I would say that if nvme-cli is told to use /etc/nvme/hostnqn and
the filename is not there for nvme-cli to create the file containing a
generated hostnqn per NQN naming standard in the NVMe spec.

Jay

> 
> Thanks
> -gili
> 
> -----Original Message-----
> From: J Freyensee [mailto:james_p_freyensee at linux.intel.com]?
> Sent: Monday, August 08, 2016 8:55 PM
> To: Sagi Grimberg; Gili Buzaglo; linux-nvme at lists.infradead.org
> Subject: Re: Configured nqn
> 
> On Mon, 2016-08-08@11:47 +0300, Sagi Grimberg wrote:
> > 
> > > 
> > > 
> > > Hi
> > 
> > Hey Gili,
> > 
> > > 
> > > 
> > > I was wondering if there's a special reason why not to support?
> > > preconfigured initiator nqn from a file.
> > > (similar to iscsi in /etc/iscsi/initiatorname.iscsi).
> > 
> > It was raised on the list just last week. I think Jay has (of was
> > :))?
> > volunteered.
> 
> I was going to do it, but I was waiting for a reply on confirming
> what behavior people wanted to see:
> 
> 
> "So is the request to have nvme-cli use whatever is in
> /etc/nvme/hostnqn (if it exists) over the default in the nvme-fabrics 
> driver???So preference of hostnqn input would be:
> 
> 1. --hostnqn flag
> 2. /etc/nvme/hostnqn
> 3. default hostnqn generated by nvme-fabrics driver (ex: nqn.2014
> -08.org.nvmexpress:NVMf:uuid:a2d7752c-a31b-477a-a003-31a5e1c424a9)
> 
> (note 2. may contain a hostnqn different from the default)"
> 
> 
> Gili if you are itchin' to do it, go ahead and give it a shot, just
> let me know (or I can still do it w/confirmation to above). ?The
> discovery fix I added to nvme-cli:
> 
> http://lists.infradead.org/pipermail/linux-nvme/2016-July/005562.html
> 
> has been merged into the nvme-cli tool by Keith.
> 
> 
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme

Configured nqn

[Gili Buzaglo]

Hi
I think the host nqn file should exist even before the first connection attempt, so
That a management software will be able to first configure the storage to allow connection from this nqn(with ACL).
The rpm spec can run a script during first install to take care of that.

Thanks
-gili

-----Original Message-----
From: J Freyensee [mailto:james_p_freyensee@linux.intel.com] 
Sent: Thursday, August 18, 2016 1:04 AM
To: Gili Buzaglo; Sagi Grimberg; linux-nvme at lists.infradead.org
Subject: Re: Configured nqn

On Tue, 2016-08-16@14:08 +0000, Gili Buzaglo wrote:
> Hi
> Sorry for responding late.
> My original thought was that the nvme-fabrics driver will initiate the 
> nqn from the file, But your suggestion is also suitable and less 
> complex.
> Who is responsible for creating the file??/etc/nvme/hostnqn ?
> Maybe nvme-cli first installation ?

Yes, I would say that if nvme-cli is told to use /etc/nvme/hostnqn and the filename is not there for nvme-cli to create the file containing a generated hostnqn per NQN naming standard in the NVMe spec.

Jay

> 
> Thanks
> -gili
> 
> -----Original Message-----
> From: J Freyensee [mailto:james_p_freyensee at linux.intel.com]
> Sent: Monday, August 08, 2016 8:55 PM
> To: Sagi Grimberg; Gili Buzaglo; linux-nvme at lists.infradead.org
> Subject: Re: Configured nqn
> 
> On Mon, 2016-08-08@11:47 +0300, Sagi Grimberg wrote:
> > 
> > > 
> > > 
> > > Hi
> > 
> > Hey Gili,
> > 
> > > 
> > > 
> > > I was wondering if there's a special reason why not to support 
> > > preconfigured initiator nqn from a file.
> > > (similar to iscsi in /etc/iscsi/initiatorname.iscsi).
> > 
> > It was raised on the list just last week. I think Jay has (of was
> > :))
> > volunteered.
> 
> I was going to do it, but I was waiting for a reply on confirming what 
> behavior people wanted to see:
> 
> 
> "So is the request to have nvme-cli use whatever is in 
> /etc/nvme/hostnqn (if it exists) over the default in the nvme-fabrics 
> driver???So preference of hostnqn input would be:
> 
> 1. --hostnqn flag
> 2. /etc/nvme/hostnqn
> 3. default hostnqn generated by nvme-fabrics driver (ex: nqn.2014
> -08.org.nvmexpress:NVMf:uuid:a2d7752c-a31b-477a-a003-31a5e1c424a9)
> 
> (note 2. may contain a hostnqn different from the default)"
> 
> 
> Gili if you are itchin' to do it, go ahead and give it a shot, just 
> let me know (or I can still do it w/confirmation to above). ?The 
> discovery fix I added to nvme-cli:
> 
> http://lists.infradead.org/pipermail/linux-nvme/2016-July/005562.html
> 
> has been merged into the nvme-cli tool by Keith.
> 
> 
> _______________________________________________
> Linux-nvme mailing list
> Linux-nvme at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/linux-nvme

Configured nqn

[Sagi Grimberg]

> Hi
> I think the host nqn file should exist even before the first connection attempt, so
> That a management software will be able to first configure the storage to allow connection from this nqn(with ACL).
> The rpm spec can run a script during first install to take care of that.

Sounds reasonable...

Care to send a patch? ;)

Configured nqn

[Gili Buzaglo]

Probalbly yes, we'll see who's available ..
Thanks

-----Original Message-----
From: Sagi Grimberg [mailto:sagi@grimberg.me] 
Sent: Thursday, August 18, 2016 12:56 PM
To: Gili Buzaglo; J Freyensee; linux-nvme at lists.infradead.org
Subject: Re: Configured nqn


> Hi
> I think the host nqn file should exist even before the first 
> connection attempt, so That a management software will be able to first configure the storage to allow connection from this nqn(with ACL).
> The rpm spec can run a script during first install to take care of that.

Sounds reasonable...

Care to send a patch? ;)