NPIV

[listener]

When I enable NPIV and generate WWN pairs, should this info show up on the switch? Will I be able to zone out my VM?

Right now I am on a Brocade 200e with Fabric OS 3.5 and I don't see the new WWN. My VM is on too.

thanks...

[mcowger]

You'll also need an HBA that supports NPIV and probably need to get it to generate some traffic.

--Matt

[listener]

I am writing IO to the disk via dt.

I have a 4G Emulex card, LPe1102 that does NPIV.

So, should I see the WWNs on the switch or not?

If I don't, how is NPIV a useful feature?

[mcowger]

I suspect it depends on your switch.....

But the idea would be to create that virtulized wwpn, then use your zonng/masking tool son your storage and switches ot mask a lun to that wwpn....have you tried it?

--Matt

[listener]

I added the WWN from the VM to the switch. However, it is grayed out on the switch and I can't query the WWN for device info.

The physical HBA sees the storage but the virtual port doesn't.

It would have been nice to have a more in depth doc on this stuff....

[StephenTJ]

I suspect it depends on your switch's Fabric OS version. I run CISCO MDS 9216s and NPIV support is on SAN-OS versions 3.0 or later so I had to upgrade to v3.1. Also, NPIV had to be enabled on the switch after the SAN-OS was upgraded before it would work. I cannot say if its the same for Brocade's Fabric OS but its something to check on.

[listener]

For every port on the switch it says NPIV is enabled.

I looked in my vmkernel log for errors or information on the vport. All I found was:

Dec 17 03:02:34 vmshe-hp380-3 vmkernel: 0:00:01:18.624 cpu1:1041)SCSI: 861: GetInfo for adapter vmhba2, , max_vports=-1, vports_inuse=-1, linktype=0, state=0, failreason=5, rv=0, sts=0

I'm not sure if that's a good or bad thing for the HBA.

[StephenTJ]

Your HBA does not have enough resources, that is what the failure reason means.

Failure reason 5: HBA does not have enough resources, Check and update the Fibre Channel HBA firmware

Here is a good document for configuring and troubleshooting NPIV:

[listener]

Cool The doc was really useful. I have upgraded my firmware to the latest and greatest.

Now under /proc/scsi/lpfc/1 I see this so NPIV is definitely on.

SLI Rev: 3

NPIV Supported: VPIs max 127 VPIs used 0

RPIs max 512 RPIs used 16

However, there is still a vport problem. In the vmkernel log it tried to create it and then it failed. It says the rescan failed... Did I miss something?

Dec 17 08:53:31 vmshe-hp380-3 vmkernel: 0:00:11:28.211 cpu7:1074)WARNING: SCSI: 745: Failed to Create vport for world 1075, vmhba1, rescan failed, status=bad0001

Dec 17 08:53:32 vmshe-hp380-3 vmkernel: 0:00:11:29.211 cpu7:1074)SCSI: 117: Physical Path : adapter=vmhba1, channel=0, target=1, lun=25

[listener]

Anyone?

[Si-clone]

Does anyone have any further feedback here? I'm experiencing an almost identical situation, although with a Qlogic HBA. FC Switch and HBA's all support NPIV.

[Glen_Ansley]

Listener, I may very well be wrong but, I believe that NPIV functionality wasnt introduced until Fabric OS v5.1.x. It might pay to upgrade to the latest anyway, unless of course 3.5 was a typo as I couldn't find a v3.5.

[listener]

True, NPIV was introduced in version 5.1. I have 5.3

brocade_200e> version

Kernel: 2.6.14

Fabric OS: v5.3.0

Made on: Thu Jun 14 19:06:31 2007

Flash: Fri Dec 14 14:52:51 2007

BootProm: 4.6.4

Area Port Media Speed State Proto

=====================================

0 0 id N2 Online F-Port 50:0a:09:81:86:d7:cc:11

1 1 id N2 Online F-Port 50:0a:09:85:00:01:8d:56

2 2 id N2 Online F-Port 50:0a:09:81:86:07:d6:9a

3 3 id N2 Online F-Port 50:0a:09:85:00:00:d7:5a

4 4 id N4 Online F-Port NPIV public

5 5 id N4 Online F-Port NPIV public

[ironmannut]

Yeah i'm having the same issue. In VI i can see the generated vports but when i try to confirm them by going to /proc/scsi/lpfc/1 it does show any of the vports. I've tried both qlogic and emulex. I'm running a brocade 200e with FOS 6.0 with VMware 3.5. Has anybody got it to work?

[Si-clone]

From what I can tell after information was received from our Brocade support is that the Brocade 200e needs to be in Gateway mode. If it is in switch mode (default) then it will not work. I'm passing on some basic information here as I have not researched the matter further. Our development 200e doesn't have all 16 ports filled (a requirement of gateway mode I think) so I'll be running some tests on a new 200e in a few weeks. I'll report back here if nobody else has solved the issue in the short term.

[waynej]

From Brocade's spec sheet on the 200E:

Switch mode (default): 16 universal ports (max)

Brocade Access Gateway mode: 12 F_Ports and 4 N_Ports.

Port types:

Switch mode: FL_Port, F_Port, and E_Port
Gateway mode: F_Port and NPIV-enabled N_Port

So it does appear that you must be running in Gateway mode to have NPIV enabled. You also do need all 16 ports populated and FOS 5.3 or higher to use Gateway mode according to the 200E FAQ on the brocade website.

[sim]

I had the same problems. ("Failed to Create vport for world 1071, vmhba1, rescan failed, status=bad0001"). I traced the issue to be erroneous LUN mapping in the storage. I had mapped the storage LUN to LUN0 (as seen by the fabric/ESX Server) instead of LUN5 expected by the ESX Server.

After remapping the LUN everything works fine.

Sim

[listener]

NPIV does not require that you have a brocade running in gateway access mode. The feature will work just fine on a single brocade running in normal management mode.

Turns out the vport will get destroyed if you don't map the vWWNs to the storage group (aka igroup on netapp equipment).

[sim]

Yeah! I forgot to mention in my post that I disabled Access Gateway mode on the Brocade switch. It worked fine. It turns out that both our issues were due to storage Mapping with vWWN. In my case I had mapped the vWWN to the storage but to incorrect LUN number.

Thanks.