host IP address not responding

[notRoman]

I hope this isn't a stumper, but it's weird. I started up my ESXi 4.1 host one day and could not connect to it. Investigating further, I noticed that while I can set an IP address and even get one automatically with DHCP, it doesn't work. Even at the management console in tech support mode, this host can't ping it's own IP or the default gateway. It can ping 127.0.0.1. Also, virtual machines that were set to start up automatically do and I can connect to them just fine over the network. So, obviously the hardware and network are working fine or in any case, since I can't ping its IP address from itself, then I guess that means that something in the networking software is broken. !? I suppose I could have done something to cause this, but I don't know what. It had been working fine before I shut it down a couple days earlier as far as I know. Any clues?

[weinstein5]

How many physical NICs are in the host?  If VM networking is functioning I think the networking si working but the issue is on the physical layer I would look at the physical NIC that  managemnet port uses -

[VirtuallyMikeB]

Good day!

First and foremost, try restarting the managemnet network from the DCUI.  It will be good to know a bit more about the configuration as already mentioned, but first, restart the mgmt network.

Thanks,

Mike

http://VirtuallyMikeBrown.com

https://twitter.com/#!/VirtuallyMikeB

http://LinkedIn.com/in/michaelbbrown

[notRoman]

Thanks for the replies!

I have one NIC and I have restarted the machine at least once as well as switched from DHCP to static IP.

[VirtuallyMikeB]

This is certainly a weird situation.  The management VMkernel port *can* receive DHCP traffic - which means it's sending DHCP broadcasts and receiving unicasts from the DHCP server, possibly across VLANs (assuming the DHCP server resides on a separate VLAN).

But the managmenet VMkernel port *cannot* ping its own IP or the default gateway.

Is this a network at work or a home lab or...? What kind of setup are we talking about?

Has the physical switching changed at all? Do you have visibility into it? If so, can you post configs? Are there VLANs involved?

Are you sure the physical NIC is supported by VMware?  Follow the second half of the following Knowledge Base article where it calls out

Obtaining Network card driver and firmware information

http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=102720...

Then verify the NIC is supported by looking it up on VMware's HCL.  Maybe this can shed some light on it.  If it's a newer, namebrand server with onboard NICs, you may not have an issue, but let's make sure.

Cheers,

Mike

Message was edited by: VirtuallyMikeB (forgot the link to the KB!)

[notRoman]

Ok, the setup is a PC connected to a DSL router/ ethernet/wireless switch with a few things plugged in. It is a home test bed basically. I don't remember exactly what is on VMware's supported hardware list if any of it, but it has been working fine for several months.

I was messing with some networking stuff a bit, but I think it was mostly trying to figure out what's going on after the problem and 1) I looked at the router and can't think of what might be wrong and 2) if networking works well enough to get packets in and out of the machine, how could it be a networking problem especially when 3) I can't ping my own IP either from the management console menu thing or from a command line (tech support mode)? I guess I was thinking maybe someone would have a procedure to rebuild/reinitiate something that gets saved on the machine or knows what logs to look at. I thought I had posted something, but now I don't see it here. It was something in one of the logs ("messages" i think) about the heartbeat not connecting to any IP. I'll have to go back and take another look to get the exact wording. But that just seems to tell me what we already know, that IP binding isn't quite working.

[dtracey · Accepted solution]

Hi there,

Couple of basic things to check - make sure there's nothing else on the network that has come up with that IP address.  Also check that the NIC is enabled for management traffic through the DCUI. and that no VLAN has been set for the NIC.

If all else fails then you can backup your VMs (if they are on local storage) and reset the host configuration (again through the DCUI).

Good luck,

[notRoman]

Ok, well, I reset the system configuration at the DCUI and that worked. I can get back in. Now I just need to see if I can add all my virtual machines back and get them to work. I've added one and it looks ok.

Thanks for the help everyone.

[notRoman]

Now that I'm rebuilding everything, I remembered one unusual change that I made to the vSwitch which is that I allowed promiscuous mode. Also, I set monitor_control.restrict_backdoor = true on two virtual machines. One VM was a clone of the other. Setting promiscuous mode at the vSwitch apparently is necessary to allow the machine to use the new MAC address that it gets or at least it's necessary to allow communication with these VMs. That's because the OS installed on them is ESXi 4.1. Now, this all did work for at least a few days; so, that could be totally insignificant.