doubling in FC HBA

[drehstrom]

Hi there,

I just stumbled upon a problem configuring the SAN for our ESXi (7.0). Our servers are equipped with two dual FC-cards each. So I'd expect to see 4 vmhba. But checking the adapter section I see 8 vmhba!? Checking deeper I discovered that each two vmhba report the same WWPN and WWNN. Same problem on both ESXi (servers are HPE DL 380 Gen10)...

Maybe someone can explain this!?

Thanks in advance
Stephan

[DavoudTeimouri]

Hi,

Let me know, the device part number and also may be you have CNA card as well and you are using CNA as NIC.

[drehstrom]

Hi Davoud,

thanks for helping me out here.
Adapters are all HPE SN1100Q 16Gb 2P FC HBA. We have some FCoE-capable NICs as well but we are not using them. They do show in ESXi separatly. I've attached a screenshot to make it clearer.
The following vmhba pairs show the same WWNN and WWPN:

vmhba1	vmhba72
vmhba2	vmhba73
vmhba76	vmhba78
vmhba77	vmhba79

And maybe you can explain to me the logic behind the random numbering 😎!?

[a_p_]

In case that this is an environment that has been updated to ESXi 7.0 Update 1d, then you may have hit the issue discussed in https://communities.vmware.com/t5/VMware-vSAN-Discussions/Strange-device-controller-phenomenon-after...
In case that's it, then - rather than trying to fix the issue yourself - consider to open a support case with VMware to have that fixed.
Support may also be able to answer your question regarding the vmhba numbering.

André

[DavoudTeimouri]

Hi,

If you didn't enable Virtual Function or something like that on the devices, check the below link and check the list of storage adapter via Shell and remove duplicate VMHBA, if necessary.

https://kb.vmware.com/s/article/2127274

[drehstrom]

We recently updated to 7.0.2 but I don't exactly know since when the doubling is present. Will definitly check out the link you provided.

Thanks a lot 😊.

[drehstrom]

I checked the link you provided and did exactly as suggested there. But this article is quite confusing. First they talk about /etc/vmware/esx.conf, then I should make changes to /etc/esx.conf which doesn't exist and I can't find anything like "vmhba" or "vmkdevmgr" in esx.conf as described...

EDIT: Don't know if it's significant or helps but looking into the storage controller details via "esxcli storage adapter list" I see a "Second Level LUN ID" remark under Capabilities at four of the eight vmhba...

[a_p_]

The path name in steps 4, 5 and 6 looks like  acopy&paste bug, i.e.  the "/vmware/" subdirectory is missing.
Anyway, the KB article is doesn't apply to vSphere 7.0. With vSphere 7.0 parts of what had previously been sored in the esx.conf file, is now stored in the configstore, a binary database.

André

[drehstrom]

Yeah, thought so.
I used the command suggested in the link you provided and its exactly like described.

Since this is a productive system I will not mess around with any esx-files if I#m not sure what I'm doing, so I will open a support case.
Will post here when we found a solution.

Thank you all so far for your support.

[HyunwooPark]

Hi

Is there any further details about this?

[e_espinel]

Hello.

HPE StoreFabric SN1100Q 16Gb 2-Port FC HBA (P9D94A)

The HPE SN1100Q 16Gb Fibre Channel Host Bus Adapters supports for N_Port IDvirtualization (NPIV) and industry standards-based Class-specific control (CS_CTL),provides end-to-end quality of service (QoS)

If the Adapter is the right one, you should install the following tool, in order to manage and configure the adapter.Another way is to enter the adapter through the UEFI (BIOS).

https://support.hpe.com/hpesc/public/swd/detail?swItemId=MTX_b8c15cc415e848fa957fc4abe9&swEnvOid=200...

[drehstrom]

Hi,

like I said, I opened up a support case about this.
In general there were two solutions:
1. We deactivated NVMe for we're not using it. So we got rid of half the double entries.
2. We deactivated FCoE for we're not using it either and now we see only what we want to.

But no help with the strange numbering. Support stated, that this is just a cosmetic problem and they're not able to fix it. It's not pretty but I can live with it.

[jayodele1234]

Hey, just to shed light on this. We recently had the exact same issue not too long ago. We also contacted VMware support and they first asked us to run this command and share the results with them "esxcfg-scsidevs -a"

Shortly after we received the message below on exactly how to disable NVMe support on our HBA's and this resolved our issue.

To workaround this "cosmetic" issue, disable NVMe support for your HBA and reboot the host (make sure it's in maintenance mode before) 
 
Set the parameter by running this command:
 
esxcfg-module -s 'ql2xnvmesupport=0' qlnativefc
 
Then reboot the host. This disables NVMe support.