brunowolff
Contributor
Contributor

NPIV settings

Hello folks!
I need to allow a Linux virtual machine created on my vCenter enviroment acess a specific LUN storage.
First of all, we have a cluster with 4 physical hosts and HA enabled. I'm zonning this LUN to all physical hosts and created a RDM disk for our a Linux vm. After this, I set the NPIV feature on virtual machine options and checked the box "Generate new WWNs". I zonning this LUN to my virtual machine Linux using the WWN's informed by vCenter. The issue is that my fabric switch can't see these virtual WWNs.
My host has the following HBA: QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA and I had already set the NPIV feature on my fabric switch.
I checked on the physical host but there is no VPORT or virtual WWNs
~ # esxcfg-scsidevs -a
vmhba0  ata_piix          link-n/a  sata.vmhba0                             (0:0:31.2) Intel Corporation 31xESB/632xESB/3100 Chipset SATA Storage Controller IDE
vmhba1  qla2xxx           link-n/a  fc.2000001b329aa89b:2100001b329aa89b    (0:14:0.0) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba2  qla2xxx           link-n/a  fc.2001001b32baa89b:2101001b32baa89b    (0:14:0.1) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba3  qla2xxx           link-n/a  fc.2000001b329a4a9d:2100001b329a4a9d    (0:15:0.0) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba4  qla2xxx           link-up   fc.2001001b32ba4a9d:2101001b32ba4a9d    (0:15:0.1) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba5  qla2xxx           link-n/a  fc.2000001b329a8e9e:2100001b329a8e9e    (0:18:0.0) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba6  qla2xxx           link-up   fc.2001001b32ba8e9e:2101001b32ba8e9e    (0:18:0.1) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba7  qla2xxx           link-n/a  fc.2000001b329ae5a2:2100001b329ae5a2    (0:19:0.0) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
vmhba8  qla2xxx           link-n/a  fc.2001001b32bae5a2:2101001b32bae5a2    (0:19:0.1) QLogic Corp ISP2432-based 4Gb Fibre Channel to PCI Express HBA
Thanks for help!
Tags (2)
0 Kudos
5 Replies
alefestaedist
Hot Shot
Hot Shot

Humm from the FC switches what do you see? I mean what kind of Fabric switched you got (Brocade?), do you see any WWN(hosts,storage)?

eventually can you provide your zoning info (as you configured?)

0 Kudos
brunowolff
Contributor
Contributor

Hello! My switch is Cisco MDS 9513 NXOS 5.0(1a) and the storage is Hitachi USP-V.

I can see the host's WWNs and the storage's WWNs. Included I have a zone between hosts and storage. My problem is the physical hosts cannot see vm's WWN neither the FC switch.

0 Kudos
alefestaedist
Hot Shot
Hot Shot

Sorry for the delay.this document from brocade may help you. http://www.brocade.com/downloads/documents/white_papers/white_papers_partners/NPIV_ESX4_0_GA-TB-145-...

IMO I would check this:

1) may the host access to the datastore/LUN where the RDM is created?

2) now that the host is correctly configured inside the zone did you tried to regenerate the NIPV WWN insided the VM and check if the VM now appears?

3) do you have any error on the swtich?

Alex

0 Kudos
brunowolff
Contributor
Contributor

sorry for delay too

- my 4 hosts access the LUN where the RDM is created. (The compatibility mode of the RDM was set to virtual)

- I have configured correctly the hosts inside the zone and I had already tried to regenerate the NPIV WWN's (2 pairs) but my network admin said that he cannot see these virtual WWN's and he did not see any error on switch about this.

Another doubt, if I generate NPIV WWN's of my virtual machine through vCenter, may I check the virtual WWN's on its, respective host. Like: esxcfg-scsidevs -a?

0 Kudos
alefestaedist
Hot Shot
Hot Shot

Sorry for the delay..this time on my side.. I've did  some resarch and maybe got the solution for you.

Basically what I see is the what you did is all correct, as long I ask you just to check a couple of things before go what I wrote below, the real issue you got is the way the VM present its newly WWN to the Storage. So here's what  you have to check and note.

First of all the check:

inside the ESX type:

#ls /proc/scsi/qla2300

This will help you to find the right hba to search through. Once you got the right number (i.e. 6) type:


# cat /proc/scsi/qla2300/6

This will help you to see if there is traffic through the HBA, you should have an output similar to this:


Virtual Port 1 SCSI LUN Information:
( 0:10): Total reqs 10, Pending reqs 0, flags 0x0, 2:0:1000,

Retry the command once the VM is powerd on:
# cat /proc/scsi/qla2300/6

This time you should see something like this:


FC Port Information for Virtual Ports:
Virtual Port index = 1
Virtual Port 1:VP State = <ACTIVE>, Vp Flags = 0x0
scsi-qla2-port-3=500601609020fd54:500601601020fd54:a00000:1000: 1;
scsi-qla2-port-4=500601609020fd54:500601681020fd54:a10000:1000: 1;
Virtual Port 1 SCSI LUN Information:
( 0:10): Total reqs 10, Pending reqs 0, flags 0x0, 2:0:1000,

now if everything seems ok let's try this:

Download the vmx file from the VM and check the entry similar to these:

wwn.node = “25bb000c29000ba5″
wwn.port = “25bb000c29000da5,25bb000c29000ca5,25bb000c29000ea5,25bb000c29000fa5″
wwn.type = “vc”

these are you what you need to zone..

When the VM is powered on, before the virtual hardware POSTs, it scans the physical HBAs of the ESX host for the RDM mapping to SAN storage. If an HBA doesn’t support NPIV, it is skipped in the detection process. If ESX cannot see the zoned RDM LUN through an NPIV aware HBA, again the HBA is skipped in the detection process.

So first suggestion is to ask your storage admin to manually zone the wwn you just found on the fabric, IMHO I would create 2 different zone one for the host and one for the VM in order to avoid unusual traffic from and to the VM/Host.

ESX will assign a maximum of four NPIV Port WWNs during the detection process. What this means is that if you have four NPIV HBAs connected to four NPIV aware SAN switch ports which are in turn zoned to four SAN LUNs, all four will be NPIV activated. If you have only one NPIV HBA, you’ll only use one of the virtual Port WWNs. If you have six NPIV HBAs, only the first four will be activated with NPIV Port WWNs in the discovery process.

So in my lab test to get things work I tried to disable the zoning and check if everything worked..once I've got the correct WWN I re-zoned all.

I am aware that probably in your enviroment you cant do that easily so the only way is to catch through the storage switch screen if the wwn comes online or not (it require a great eyes since a VM comes online for HBA scan for less than 1 sec) or try with a manual zoning of the WWN created.

Last bu not least remeber that if you regenerated the WWN is better to unregister the VM from the inventory and re-add since Vic use to mantain information on a created WWN VM.

Hope this could solve your problem.

Alex

0 Kudos