Azure w/ ExpressRoute.

We have a group of maintenance machines that shut down at night to reduce resources. Everything running fine for months. Yesterday, we came in to find several of the machines unreachable. They can be reached from within their Azure subnet, but not from elsewhere on the network.

After some troubleshooting, we noticed that they have lost all the BGP routes that normally show on their network interface. They only have their UDR for their local subnet.

We have a case open but it's getting bounced around in the finger-pointing game currently.

All working machines have their routes from BGP. It just got more urgent because a Production machine rebooted last night and now it is the same thing.

The NIC shows connected to the VNet, and the other NIC's connected to that VNet are working and effective routes show their routes from BGP.

We've tried to Redeploy the VM's thinking it was possibly an Azure host issue to no avail.

  • What's your question? – Andrew Schulman Jan 25 '18 at 16:12
  • Looking for ideas on what may cause BGP routes to not re-populate to a VM after a reboot when they have been working fine this way for months. – BriKP Jan 25 '18 at 16:13

Issue resolved.

Microsoft finally found & admitted that non-communicated updates night before last had introduced a NAT issue on 5 clusters that were preventing BGP routes to propagate to certain hosts.

Redeploy, new NIC's, etc. did not work as we were moving to different hosts or changing NIC's but in same affected clusters.

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.