Virtual Machine Network ConnectivityProblem Within New VI3 cluster

[sfortuna74]

Hello -

We completed the build of a brand new ESX 3.5 cluster last week, which is comprised of 4 identical HP blade servers. The problem which we have been seeing is as follows: In at least three instances in the past 2 days, a virtual machine has lost it's connection to the network, and the only fix seems to be VMotioning the VM to a different ESX host. The problem is not specific to any one on the host machines, as we have seen it happen on at least 3 of the ESX servers so far.

Has anyone seen this problem before, and could anyone point us in the right direction towards working towards fault isolation?

Thanks very much,

Steve

[Texiwill]

Hello,

Verify that your VM does not have any snapshots. Also verify by looking at your Guest OS system logs what is happening to the network. Lastly, did you install VMware Tools?

Best regards,
Edward L. Haletky
VMware Communities User Moderator, VMware vExpert 2009
====
Author of the book 'VMWare ESX Server in the Enterprise: Planning and Securing Virtualization Servers', Copyright 2008 Pearson Education.
Blue Gears and SearchVMware Pro Blogs -- Top Virtualization Security Links -- Virtualization Security Round Table Podcast

[thehyperadvisor]

I am not sure if your using passthru modules, cisco 3020's, etc for network or how your vswitches are configured? But you should check to make sure that all switches and ports used for your esx servers are configured the same.

Verify the nic order when you have the issue, if changing the nic order fixes the issue it's possible a port config on the switch.

I have seen the drivers cause issues in my HP cClass blade environment so make sure that your using the latest psp (8.1) and drivers for blades.

When the issue occurs is it with only one vm guest or all the vm guests on the host that are affected? And does it work then stop working for no known reason?

If none of this helps send your physical and virtual configuration setup/layout.

hope this helps - thehyperadvisor.com

[sfortuna74]

Thanks very much for the feedback - Our network architect just informed us that he's seen some errors on the Cisco switch which the blades are connected to, and he's about to open a TAC. Nevertheless, we'll keep your comments in mind with specific regards to the problems you've seen with drivers on the HP Blades.

[sfortuna74]

Also I need to update the information I originally posted and add that I subsequently found yesterday that vmnic3 on all 4 vswitches attached to each ESX server is not listed as seeing the necessary VLAN. This would seem to be in line with our Network team seeing issues on the Cisco switch.

Thank you again for the counsel.

[sfortuna74]

Out network team has identified an error condition on our Cisco core switch.

[thehyperadvisor]

Glad you found the problem.

hope this helps - thehyperadvisor.com

If you found this or other information useful, please consider awarding points for "Correct" or "Helpful".