VM down alert, but Host is up

[SunshineM]

Hello,

We have 3 ESXi 5.0 hosts with DRS enabled. On one host, there are around 7 VMs. We received down alerts for all these VMs for that particular host only.

> I could not RDP and ping these VMs.

> However console was accessible

> Can ping that Host

> Datastores were in connected status

> Networking on that VM was ok, showing no red cross

I just vMotioned those VMs to other hosts, and all were able to RDP/ping.

But I want to know what this have should caused, how can I troubleshoot?

thanks in advance,

Mihir

[weinstein5]

Since all failed I would look at the common connections - If all the VMs were inaccessable and connected to the same virtual switich I would look at the network connections to that vSwitch

[SunshineM]

You mean physical connections of NICs (uplink) connected from vSwitch to physical switch ports?

I can ask my network enngineer for physical switch ports.

But if I want to check for vSwitch Networking. Any commands to see logs?

As I told just went through Host > Configuration > Networking and nothing was found abnormal. So how to check through commands?

[weinstein5]

You mean physical connections of NICs (uplink) connected from vSwitch to physical switch ports?

Yes - since you lost communication to all 7 at one is an indication that their network connections were dropped simultaneously and my experience that is a network issue -

[SunshineM]

I will check with network engineer. But let me know how to check by taking SSH of that host. I mean what commands for checking vSwitch and its networking.

[SunshineM]

Just found in the Networking Adapters section, there are 2 uplinks in use. Both uplinks are showing "UP". But for one uplink "Observed IP Range" is showing as "None".

Went to this article VMware KB: Observed IP range does not show network in ESX or ESXi but it does not say anything.

How to troubleshoot this?

Thanks.

[a_p_]

Basically I would expect the two uplinks to show similar values for the observed IP ranges. Depending on the physical switch you could click the speech bubble next to the vSwitch to see how the CDP information (in case of Cisco switches) looks like. I f this doesn''t help, run esxtop from the command line (either the ESX Shell or through SSH) and press "n" for Network, this will show you the VMs, which uplinks they use and whether there's traffic on it.

André

[SunshineM]

Yes, both uplinks were showing observed IP ranges earlier.

CDP is not showing for vmnic that "NONE" Observed IP range.

ESXTOP > n shows 4 vms through vmnic0 (which is working OK) and there is 1 VM showing connection to vmnic1 (which is showing "NONE" and that cannot be pinged from outside)

Please suggest.

[a_p_]

With the link showing as active, but no CDP information (I assume the other vmnic shows CDP!?) this could be a software/driver issue. I saw something like this in the past with ESXi 4.1 (and I think Broadcom NICs) and the only way to make the uplink work again was to reboot the host. I can't tell you for sure whether this is the case, but since you have vMotion/DRS it may be worth a try!?

André

[homerzzz]

I have seen this issue with 5.0 and to resolve without a reboot you can remove the vmnic from the uplink, then re-add it to the uplink. The observed ranges should show. I have not seen this issue again since upgrading to 5.1

[SunshineM]

Thanks Guys for suggestions. I upgraded the Uplink drivers to the latest version, and this solved the issue.

[JagadeeshDev]

It should an issue with the networking. (check the VM network port group or vswitch)