How to disable an FC HBA alert on ProLiant DL380 G7

[tekns]

I Have HP Systems Insight Manager 7.1 Update 1 installed on a Windows 7 x64 system. I have discovered several VMWare ESXi 5.5 hosts running on HP ProLiant DL380 G7 servers. All of these servers were installed using the HP custom .ISO for VMWare ESXi 5.5, so they are running the HP agents.

I have WBEM credentials configured and it is correctly reading the hardware with the exception that it is throwing the following warning/error on all systems under the Health Status --> WBEM header on teh System Status page:

! FC HBA

If I click on the FC HBA link it generates the following detailed information:

FC HBA Ports
Port Status Port Name Controller ID Port Type Port Type Description Current State
Stopped Fibre Channel Port 1 QLogic QLE8152 Other Fabric There is a minor problem that is causing limited interference.
Stopped Fibre Channel Port 2 QLogic QLE8152 Other Fabric There is a minor problem that is causing limited interference.

FC Port Statistics
Controller ID Bytes Transmitted Bytes Received Pkts Transmitted Pkts Received CRC Errors Link Failures
QLogic QLE8152 0 Bytes 0 Bytes 35741347 24021341 1 0
QLogic QLE8152 0 Bytes 0 Bytes 35741562 24021616 1 0

I can see entries for the following under Storage Adapters in vSphere:

ISP81xx-based 10 GbE FCoE to PCI Express CNA
vmhba1
vmhba2

I assume this is the hardware causing the FC HBA events in HP SIM. I am looking for a way to stop getting these events or ignore them.

Possible solutions:
1. Filter/ignore the FC HBA within HP SIM (it is not an option under "WBEM Health Inclusion Status"
2. Disable the hardware on the VMWare ESXi 5.5 server so that it no longer registers with VMWare or responds to HP SIM WBEM query.

Anyone have steps to accomplish either of these or an alternate solution to accomplish stopping the FC HBA events?

[tekns · Accepted solution]

After 3 months of digging and researching, I was able to find my own answer to this.

What appears to be happening, as best I can tell, is that the VMWare ESXi kernel is detecting the QLogic QLE8152 10GB NIC as both being an iSCSI NIC device (loading system driver "qlge" for it) as well as being a Fibre Channel device (loading system driver "qlnativfc" for it). HP Systems Insight Manager sees the Fibre Channel driver loaded, but does not see any successful connectivity for it, so generates an alert (similar to what it does for NICs that aren't connected, but with no option in the "WBEM Health Inclusion Status" to Ignore it).

Solution:

Disable the "qlnativefc" system/kernel driver within VMWare ESXi.

I have tested the following against VMWare ESXi 5.5.0 (build 1331820), but suspect it will be the same/similar for other VMWare ESXi versions.

Detailed Steps

1. Enable SSH for the Host via vSphere (Configuration tab :: Security Profile  :: Click "Properties" on Services window :: highlight SSH :: click "Options" :: Click "Start" )

2. SSH connect to the VMWare Host using Putty (or your favorite SSH program) and login as root.

3. Run the following command to check the status of the QLogic kernel/system drivers:

esxcli system module list | grep ql

Normal output should look something like:

qlnativefc    true    true
qlge    true    true

4. Disable the qlnativefc system driver from loading on kernel boot by issuing the following command:

esxcli system module set --enabled=false --module=qlnativefc

5. Verify that the qlnativefc system driver is diabled by running the same list command from step 3. Output should look something like:
qlnativefc    true    false
qlge    true    true

6. Reboot the ESXi Server (will not load the qlnativefc system driver on the next boot...did not find a way to unload without a reboot):

7. Re-establish an SSH connection after the host reboots and verify that the qlnativefc system driver is both disabled and did not load by running the same list command from step 3. Output should look something like:

qlge    true    true
qla2xxx    true    true
qlnativefc    false    false

NOTE: Not entirely sure why it started loading a new qla2xxx driver after I disabled the qlnativefc driver...could be a lower detect order driver in the kernel that matches the same hardware id string as the qlnativefc driver. In any event, having it load did not cause the same HP SIM alerts for "FC HBA".

[Kerkeling]

Hi, good findings!

We had a similar problem today, but only one FC adapter on one host (out of 16 identical ESXi hosts, each with a dual port "ISP2532-based 8Gb Fibre Channel to PCI Express HBA") reported this error "a minor problem that is causing limited interference".

The difference in our case is, the FC HBAs are used for storage connectivity, so disabling the driver is not an option?!

And it's not just a false alarm that could be ignored, in vSphere Client it was also visible that one adapter had no connection, thus no redundancy to the storage anymore.

A reboot of the ESXi host fixed the problem so far, but it's not a solution if that happens again. Especially because in vShpere Client there was NO alarm (hardware status all green) in this case! Only HP SIM reported the disabled adapter!!