Has anyone seen this problem? VM kernel port that uses for Vmotion change itself to Virtual Machine port after I turn on HA/DRS in the cluster. Or this is not a problem? Can someone explain this???
Here is the set up:
I have 3 ESXi 3.5 Embedded configured. All three hosts are in a cluster. Each of the host at Vswitch 0 have two VMkernel ports.
Before I turn on HA/DRS, the network configurtion look like this:
ESX 1 -- Vswitch 0
VMkernel port (Mgt network)
VMkernel port (Vmotion) and VLAN tag 101
ESX 2 -- Vswitch 0
VMkernel port (Mgt network)
VMkernel port (Vmotion) and VLAN tag 101
ESX 3 -- Vswitch 0
VMkernel port (Mgt network)
VMkernel port (Vmotion)and VLAN tag 101
After I turned on the HA/DR, the VMkernel ports at two of the hosts changed to Virtual Machine port and the configuration changed to this:
ESX 1 -- Vswitch 0
VMkernel port (Mgt network)
VMkernel port (Vmotion) and VLAN 101
ESX 2 -- Vswitch 0
VMkernel port (Mgt network)
Virtual Machine port (Vmotion) and VLAN tag 101 (I went in to check the properties and it still have Vmotion enable, its just changed the Port Group type)
ESX3 Vswitch0
VMkernel port (Mgt network)
Virtual Machine port (Vmotion)
and VLAN tag 101 (I went in to check the properties and it still have
Vmotion enable, its just changed the Port Group type)
The Vmotion VMernel port uses the Mgt VMkernel port default gateway. (I figured since they are the same port group type within in the same Vswitch, it will only allow one default gateway).
Vmotion still works, I have no problem migrating machines between hosts.
All inputs are greatly appreciated. Please help. Thanks
Could you post a screen shot of the VI Client - because you can only enable VMotion on a VMkernel Port - there is not option to do that in a VIrtual Machine Port Group
Sorry, I haven't been back to the customer site. I'm still trying to duplicate the problem in my home lab.
Did you ever get a resolution to this? We are seeing the same behavior on a 6 node HA/DRS cluster. Two of the nodes the VM Kernel port used for VMotion changed itself to a Virtual Machine port in the infrastructure client GUI. We actually saw the 2nd one when it happened, it changed when it came out of stand-by mode.
If you run esxcfg-vmknic -l at a command prompt, the port is listed properly as a VM Kernel port, and VMotion, DRS, HA, etc all work fine on those two nodes, so it is definately an issue with the GUI not showing it properly.
ESX is at version 3.5.0 158869, so it's pretty current.
nbhms
I had the same here (again) this morning
I have a two machine ESX 3.5u4 Cluster where this never happens and a three machine ESXi3.5u4 embedded Cluster where this happens on more or less regular basis
The two clusters are managed by a single VirtualCenter 2.5u4 with an SQL 2005 Database backend
When I arrived here this morning I had two of the three ESXi's showing the Vmotion Vmkernel Port as "Virtual MAchine Port Group"
Vmotion was still operational , so it was a pure GUI problem .
A refresh on the networking page of the concerned hosts solved the issue (had to leave the netowring tab and return to see the update) , but I'm quite sure to see the same behaviour again shortly
As everything keeps working , it's not a drama , but to say the least , it's not very pretty
Cheers,
Claude
Just came accross the same condition.
One of our ESX 3.5 U4 hosts started getting errors when trying to migrate VM's off of it. " Unable to migrate from vmfbst4.xxx.xxx.to vmfbst3.xxx.xxx: The VMotion interface is not configured (or is misconfigured) on the destination host 'vmfbst3.xxx.xxx'.
Checked the portgroups and found the IP address on the VMkernel portgroup was missing. Vlan information and nic teaming info still there, just lost ip. the host is part of a 6 host HA/DRS cluster that has been running for quite some time. they were upgraded to 3.5 earlier this year (feb/mar) and have all current patches. All other host are running fine and the VMkernel port group on them still shows ip info.
Added new VMkernel portgroup and was able to migrate VM's once again. Tried to remove broken Kernel portgroup but recieved error : Unable to delet portgroup"VMkernel1", for the following reasons: VM Kernel NIC"
Left running with 2 VMkernel portgroups(1-working, 1-not working) overnight. This morning the new portgroup has lost it's ip information also, and vmotion is broken again.
Anyone come up with a cause/solution for this yet?
Have exactly the same problem here with ESXi U4. VMkernel ports changes to a virtual machine prorrt and looses ip information when adding the host to a hA/DRS Cluster. Does anyone have a fix?
I found this is just a GUI problem, had the same issue, but when i refreshed the network configuration, it came back to the VMKernel group.
How to fix this GUI issue, any ideas?
Anbu
Restart managment service on ESXi .
1. Log on to ESXi host command line (tech support mode or ssh)
2. Type command " /sbin/services.sh restart "
if possible restart your virtual center server once.
I experienced the same problem myself. I thought I was seeing things when I was reviewing my network configuration on my ESX hosts. I was getting ready to create a new VMkernel port group for vmotion after seeing the problem, so I renamed the faulty port group. As soon as I hit apply, the port group changed back to VMkernel. I renamed it back to the original name, and it remained a VMkernel port group. I haven't tried vmotion yet, but had done so the previous day without any error.