VirtualAerospac
Contributor
Contributor

ESXi update taking VMs off network

Hello,

I recently updated our ESXi 5.0.0 blades from 768111 to 821926.

I applied this update using update manager to 15 hosts.

Out of the 15 hosts we are now having issues with 2 of them. The hosts are on the network and in vSphere 5.0 client ok but any virtual machines running on them appear off the network.

Has anyone see this issue? or anything like it?

Thanks

Austin.

0 Kudos
12 Replies
milton123
Hot Shot
Hot Shot

0 Kudos
VirtualAerospac
Contributor
Contributor

That doesnt really seem to help me out.

All my virtual machines are up to date and the other hosts that I updated work. It happens to be just these 2 hosts.

0 Kudos
UmeshAhuja
Commander
Commander

Hi,

As said by Milton try to upgrade Hardware version and check , if it does'nt work then have a look to the below link where its says that sometime there might be the vmnic configuration problem.

http://vmtoday.com/2012/02/vsphere-5-networking-bug-2-affects-management-network-connectivity/

Thanks n Regards Umesh Ahuja If your query resolved then please consider awarding points by correct or helpful marking.
0 Kudos
VirtualAerospac
Contributor
Contributor

Hi - thanks for getting back to me.

Im still getting a bit confused though on the responses. My virtual Machine hardware is fully up to date and the host it sits on.

We have 15 hosts and just over 50 virtual machines that can vMotion between them. the VM is live on all hosts but when i move it to host 2 the ping drops, but host 2 is still online and working. Once I vMotion it to another host the ping comes back up and the VM works.

All virtual machines and hosts are on the latest versions.

I have attached a screenshot of our network config and as you can see the server sits on the "Production vlan18" (not to sure if it would help though)

0 Kudos
a_p_
Leadership
Leadership

5000, 2000, 1000, ...

Do you use HP blades with Virtual Connect? Did you already check whether there's an issue with the configuration/settings?

André

0 Kudos
UmeshAhuja
Commander
Commander

Hi,

Is all your hosts having same network configuration  i.e There is no differenct in the Names define to them ?

Any error you are getting if so pls share the same.

Thanks n Regards Umesh Ahuja If your query resolved then please consider awarding points by correct or helpful marking.
0 Kudos
VirtualAerospac
Contributor
Contributor

Yes we do use virtual connect

Check all settings, virtual connect and more but everything seems grand.

Im just wondering if its possible that an update got messed up during installation. The same updates applied to 15 but 2 get affected.

Is there a roll back feature?

0 Kudos
VirtualAerospac
Contributor
Contributor

There is no other errors coming up anywhere.

There isnt even errors on the 2 hosts themselves, they look to be fine.

0 Kudos
UmeshAhuja
Commander
Commander

Hi,

Can you cross check the naming convention for all the network setting for all hosts might be something difference for the host you are having problem

Thanks n Regards Umesh Ahuja If your query resolved then please consider awarding points by correct or helpful marking.
0 Kudos
a_p_
Leadership
Leadership

Im just wondering if its possible that an update got messed up during installation.

That's certainly something you can never rule out.

Is there a roll back feature?

To roll back to the previous build, press SHIFT-R while the ESXi host boots. This will let you revert to the previous state.


André

0 Kudos
homerzzz
Hot Shot
Hot Shot

I have experienced this, but with dvSwitches. On the hosts that experience the issue,  go into the vswitch properties and remove and re-add each vmnic one at a time. This worked in my case and VMs were once again pingable on the hosts.

0 Kudos
colin_graham
Contributor
Contributor

Are your NICs on vSwitch3 available to both VM Port groups? Are they configured as Trunk Ports on the physical switch connection. Have had an issue with a client in the past after an update on vSphere 4 where the NICs on the host enumerated in a different order and the physical NICs were tagged in different VLANs on the physical Switch. All of a sudden, the VM port Group had servers on the wrong VLAN so wouldn't respond to pings.

0 Kudos