Teaming and failover status comes up as warning -- NON IP Hash mismatch

[lvaibhavt]

Hi All,

In my home lab I have created a Distributed Switch and enabled health check. I checked and found that the teaming and health status was coming up as unknown. Screen shot attached.

After 3 minutes I got an alert in the VI client for the host stating vds teaming mismatch status. Snapshot attached.

As I cleared the alarm the teaming and health status changed from unknown to warning. NON IP Hash mismatch. snapshot attached.

I have checked all the port groups and I am not running IP Hash on any port group. I have restarted services.sh but no go. I did check vmkernel.log file but could not find anything informative. I am running esxi 5.5 U1 and VC 5.5

I am running my environment in VMware Workstation 11

Any suggestion how to resolve NON IP Hash mismatch error.

Thanks

Vaibhav

[npadmani]

did you see this KB

VMware KB: vCenter Server 5.1 Teaming and Failover Healthcheck reports the warning: Warning - Non-IP...

[lvaibhavt]

Thanks for replying

I did see that article but it mentioned about 5.1 and not 5.5 so I was hoping if there is any resolution in 5.5

Thanks

[RanjnaAggarwal]

which teaming policy are you using on your vSwitch or it's port groups whereever you are facing this issue? if it not ip hash then it might be originating virtual port ID or mac hash or LBT. then check your physical network configuration that NIC teaming is configured or not because this should not be configured and check this KB

VMware KB: vCenter Server 5.1 Teaming and Failover Healthcheck reports the warning: Warning - Non-IP...

[Gortee · Accepted solution]

Thanks for asking your question to my blog sorry about the delayed response.  Health check is a little odd sometimes.  VMware recommends using to identify issue and not constantly.  Health check is not supported on any type of virtual nics on the platform see:

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=203479...

This includes HP FlexFab or UCS.   I would assume that the virtual nic in workstation may also be throwing this for a loop.   I would consider the following options:

- Check your switch teaming policy

- Check each port group teaming policy

--Ensure none are set for IP hash (For home lab source based mac is your best bet or load based teaming)

-If everything is not IP hash then I would turn off the feature and back on you might find new results

-In the end the fact that you are using workstation may be the cause of the issue because the ESXi hosts see a virtual nic.  

Notes on screenshots:

-First shot is normal this is the message while it's still doing the tests

-Second and Third are either related to a real mismatch or the issue in the KB above

That's my best guess... the truth is I have seen all kinds of crazy on anything with virtual nics.   I saw a situation where health check was running a CAM table out of entries (500,000 entries) with health check on UCS... it is really bad on UCS don't use it.

Hope it helps.

Thanks,

Joseph

blog.jgriffiths.org

[lvaibhavt]

Thanks for explaining in detail -- it helps