Guest VM (Win2003R2 SE x64) cannot connect to network

[chopstick74]

Hello,

I just ran into this issue where all my Guest VM's suddenly could no longer access the network. This is happening to both my servers and all the VM guests on them. Here is my current setup:

Servers: Dell PowerEdge 2950 + 2 Quad Core Xeon processor + 32 GB RAM

VMWare: ESX 3.5 Update 2 (standalone server with no other VMWare infrastructure components)

ESX IP Setup: DHCP Client

1. of Guest VM on each server: 2

Guest VM OS: Windows Server 2003 R2 Standard Edition x64

Guest IP Setup: DHCP Client

I have no problem running with the above setup for the last few months. All of a sudden, all guest VM can no longer ping even the default gateway. After rebooting the guest OS, I couldn't even receive an IP from the DHCP server. The odd thing is that ESX server has no problem receiving an IP from the DHCP server, but none of my VMs could. It's like the virtual switch became isolated and no traffic could go beyond the ESX switch. I have made no changes to the network. Since I am beginner with VMWare, all network settings are default settings.

Another strange behavior is that if I leave the VM on for a long period (over night), it will eventually get an IP (and all other info like DNS/WINS/etc) from the DHCP server, but I still cannot ping the default gateway or anything else outside of the ESX server. Anything in the ESX server, I could PING them. So, VM 1 could ping VM 2 within the same ESX server.

I am suspecting 1 of two possible issues:

1) Something happened to our network and some how all guest VM's are affected and cannot receive an IP from the DHCP server

2) There is a bug within ESX 3.5 Update 2 that because effective as off Nov 21, 2008.

Any help anyone could provide, would be greatly appreciated.

Thanks,

Ben

[weinstein5]

Welcome to the forums - I would personally err on the side of a network issue - how are your virtual switched configured? Do you have the VMs and the Service Console connected to the same virtual switch? If you hard code the IP address in your VMs can they ping each other? Are you sure of the gateway is still valid? Can you ping the gateway form the service console or another machine on the network?

If you find this or anyother answer useful please consider awarding points by marking the answer helpful or correct

[RParker]

Is this U3 of ESX? I would first try logging into the console of those, do : service mgmt-vmware restart

That will restart the services on that machine. Failing that, put the VM's in suspend, reboot the host, see if you can connect at that point. You can also try setting up a new network / Virtual Switch the same as the one you have already but on a different Network port. Maybe the driver hung the NIC.

[chopstick74]

Hi weinstein5,

Thank you for your response.

There is only 1 vswitch. This is the default vswitch created during installation of ESX. I assume both VM and console is using the same switch. I have attached a picture of the diagram from VMWare Client tool.

When I power up 2 VM's, they will each get a (somewhat) random IP (DHCP client), but the same subnet (255.255.0.0). Between these 2 VMs, they could ping each other.

At this point, the default gateway is blank. Hence, no pinging to any host that's outside of the ESX server.

I did a few more tests on my server that indicated I may be encountering a network issue. I have unplugged my server from the corporate network and plugged it directly into a standalone router (with DHCP service). Within this standalone network, the VMs have no problems with receiving an IP.

Very strange.... any thoughts on what I may be encountering?

Thanks,

Ben

[chopstick74]

I am currently running ESX 3.5 U2.

I have restarted the VM and ESX. No difference.

I am not sure how to create a vswitch. I have created a new network for the VM.

Would you recommend testing ESX 3.5 U3?

Also, on my previous post, I did mention that after further testing with a different network (standalone), the VM have no issue with receiving an IP from the DHCP Server. This makes me believe that I may be experiencing a network environment issue.

Thanks,

Ben