I have been recently experiencing issues with some... - VMware Technology Network VMTN

VMware Cloud Community

my cluster is 16 x bl460c g7 blades in a c7000 enclosure & is usually v stable. esxi 5.1

I have reconfigured HA on all blades.

12 Replies

Are you sure the VM Networks are allocated to all the network ports for each host?

It sounds like the vlan that the VM is on is not presented properly to the host it was vMotioned to. The portgroup may be setup properly but if the vlan isn't presented correctly on the port(s), that could be an issue.

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

Hi Ben, thanks for reply. Yep, all hosts are identical, have been in situ for nearly 3 years. First noticed this a few weeks ago, finally got stung overnight with a major vm going down & causing production outage.

Have you had your network team check that each port on the back end is setup correctly?

I've seen this issue before and we found out that one port was missing a vlan in the trunkgroup.

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

When you login to the OS on the actual VM, there is no nic present, yet the two boxes are ticked in the network boxes when you edit settings "Connected" & "Connect at power on." Simply vmotioning the vm to another host recifies this. the blades in question have other vms running requiring the same vlan & are functioning fine.

What NIC driver do you have selected? E1000 or VMXNET3? I've seen issues like that with the E1000 driver.

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

flex10 embedded nics.

be2net driver

No, what NIC driver do the VMs use?

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

Just checking, I think its the E1000

I can't find the KB Article that I've read on this in the past, but at my last job we had issues similar to what you're seeing. After rebooting, the VM would come up with no network configuration. We had to go through the painful process of adding a new NIC with the VMXNET3 driver and swapping the IP onto the new NIC and then remove the E1000 NIC.

For a start, I recommend trying this with one of the affected VMs and performing a vMotion to see if that fixes the issue.

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

found this kb http://kb.vmware.com/selfservice/microsites/search.do?language=en_US&cmd=displayKC&externalId=100180...

Just to muddy the waters, of the two vms that had the problem, one of the vms has a nic type of flexible & the other is set to E1000. Thanks for your help so far.

Flexible is very old and I've seen many issues with the E1000 driver (as I've said)... I recommend switching to the VMXNET3 driver.

Ben Liebowitz, VCP vExpert 2015, 2016, & 2017 If you found my post helpful, please mark it as helpful or answered to award points.

HI,

As an idea, check if you haven't exceeded the ports assigned to your vSwitch. vMotion will migrate the VM on the host without giving you any warning that there are no available ports in that vSwitch .

Kind regards.