0

I have an existing CentOS 6.9 VM running in Hyper-V on Windows Server 2008 R2. I want to migrate this machine to a new Windows Server 2016 installation. This is the last VM that will allow me to shut down the old server.

My problem is the eth0 network adapter is not found on the new server. The issue is further complicated because the VM then stalls during boot trying to mount a cifs share. Checking Google shows other complaints about the missing network adapter drivers, but the normal work-arounds are not helpful to me because of the boot problems.

I waited more than 8 minutes for the cifs mount to timeout and the VM to continue booting. At that point I was running up against the end of my maintenance window, and so I had to give up and re-start the VM on the old server.

I'm afraid I'm going to need to wait until I can plan a longer maintenance window that will allow time to disable the cifs mount on the existing VM, so I can then troubleshoot the network adapter on the migrated VM before re-enabling the cifs mount. It could be several months before I window like that opens to me.

Any ideas for a fix I could accomplish during a weekly 1 hour window, remember I'll need to allow about 20 minutes of that for just copying the 180GB vhd file to the new server?

  • 2
    Take a copy of the VM and go experiment with it while you have lots of time. – Michael Hampton Feb 1 '18 at 17:04
1

The solution for the network was in two parts:

  1. In /etc/sysconfig/network-scripts/ifcfg-eth0 I commented out the line that starts with HWADDR on the original VM (I had done this before).
  2. I also needed to edit /etc/udev/rules.d/70-persistent-net.rules before shutting down the original VM. I simply commented out the line with my network card, so that udev can re-detect the NIC and give it the same name that matches my ifcfg-eth0 file. Now I can re-boot the original if I need to, but the network will also come up correctly when the vhd file is copied to the new host.

I also fixed my /etc/rc.d/rc.local file, where I had an entry that was blocking and would not allow me to use the console to fix things.

Unfortunately, I still have a problem where the network becomes unavailable again a few seconds after boot up. I can start pinging the machine and it responds as soon as the network is available during boot, but then it stops responding again after 15-20 pings and will not come back no matter what I do. I can still ping the local IP address from the VM's console (ie: not the loopback/home address), but no other address. But the network did come up first and I have console access, so I'll troubleshoot this new issue separately.


As an update, I was also encountering this bug:

https://bugzilla.redhat.com/show_bug.cgi?id=655855

For which the "fix" was to disable the irqbalance service:

sudo chkconfig irqbalance off

One more reason for me to get this server to a newer distribution.

  • Instead of commenting out those lines, couldn't you just have set a static MAC address for the VM's NIC on the new host? – GregL Mar 12 '18 at 18:00

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.