VCSA 6.7 - Lost Network Connectivity When Migrating VCSA Guest to Another Host

[Punkrulz24]

Good afternoon all,

I am at wits end with this issue. To start, I am definitely a linux noob, so if there are commands you'd like me to run please let me know.

We embarked on upgrading our VCSA/ESXi from 6.0 to 6.7 the other day. The upgrade of the appliance was successful. This morning I was preparing to upgrade one of the hosts to ESXi 6.7, and I decided that I was going to vMotion the VCSA 6.7 Appliance to the host not currently being upgraded. As soon as I had done that, I had lost network connectivity and have not been able to communicate with vCenter since. Fortunately I can connect to the web interface of the host I had moved it to, and have been able to poke around the CLI of the VCSA appliance in order to troubleshoot.

I have been attempting to compare and troubleshoot this issue but have been completely unsuccessful. Any time I go to ping the interface, or ping out from the CLI I get destination host unreachable. Can someone help in establishing what may have gone wrong here, and how it can be fixed? I will get you any information you need.

Thanks!

[msripada]

Hi,

From the description, it sounds like the nic card used by VCSA or possible issue with physical infra.

If you do not want to test on VCSA , create a new VM, assign same subnet ip address and assign same portgroup as VCSA VM

To isolate, move the VCSA (or create a new VM with same portgroup and same subnet ip as VCSA) to impacted ESXi host

Take ssh to ESXi host - > esxtop - > press n - > identify the VCSA/impacted VM name and the nic against the VM using it.

Try to uncheck and recheck the network card - validate if the nic starts pinging, if yes then the nic might be faulty or having incorrect configuration at may be switch level(guessing).

You can compare the same with other ESXi host or nics on the same host which is working fine

Thanks,

MS

[IRIX201110141]

Additional notes to what MS already sayed

It the destination PG and vSwitch on the ESX Host use multible vmnics go and move all except one to unsed and try each vmnic one after another to find the faulthy one.

We always blame our network guys for misconfigurate a pSwitch port. Rule #1: the network guys a always guilty

Regards,

Joerg

[Punkrulz24]

Good morning all!

So this is some interesting information, and to start I wonder: Does this somehow play into the fact that when I launch the old VCSA server the new one reports as orphaned?

When looking at esxtop, I do NOT see the new VCSA host at all here (even though the VM apparently resides on my second host where I'm running the command from). Also, we have two networks listed in the switching (vSwitch 0 and 1), and have functioning hosts on both of them... so I haven't been inclined to believe there was a problem with the NIC or not.

[Punkrulz24]

So I learned that this is definitely due to the networking configuration on the host. We have a network called DMZ, and whomever configured the host still called that network VM Management.

My next question: Is there any way to update the proper network to be used via command line on the esxi host? I ask as any changes we make to the VM when consoled into the host report that it can't save changes to the VM. This may be because it became orphaned partially through the vMotion to the other network.

[JuanRubio]

The same thing happened to me on a esxi6.0 host, I lost connectivity on the VCSA after migrating to that host, I disabled Traffic Shaping policy and voila!