How can we improve Azure Virtual Machines?

Fix or allow automated cleanup for orphaned Microsoft Hyper-V Network adapters on Azure VM's

We have identified an issue with our Azure VM environment.

We use an Automation script to stop and start our VM's every night - however it seems that every time the VM's start up a new virtual network adapter is created, leaving potentially hundreds of old orphaned ones on the system (showing up as Microsoft Hyper-V Network Adapter #123).

In itself this is merely an inconvenience but we have twice experienced an issue where after a certain number of these orphaned adapters have built up the VM's fail to create a new working adapter with the effect that network communications between our VMs fail with a "network location cannot be reached" error.
Additionally none of our staff were able to log into out Azure Remote Desktop hosts - because we are loading user profiles for staff that are stored on another Azure VM - and this network failure means that the profiles couldn't be loaded.

Manually deleting the old orphaned network adapters and stopping and starting the VMs again solved the problem immediately, but it took several days for us to find the cause of this problem in the first place. (not sure if it is significant but all our VM's seemed to stop at adapter #309)

I suggest that the issue is fixed - such that the orphaned network adapters are not allowed to build up - or at the very least this potential issue needs to be flagged and a simple mechanism for customers to manage them is provided.

regards
-adrian

3 votes
Sign in
(thinking…)
Sign in with: Microsoft
Signed in as (Sign out)

We’ll send you updates on this idea

Adrian Mitchell shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

1 comment

Sign in
(thinking…)
Sign in with: Microsoft
Signed in as (Sign out)
Submitting...
An error occurred while saving the comment
  • Robin commented  ·   ·  Flag as inappropriate

    Had the same issue. My company got crippled for 3 weeks because of this! Even MS support team couldn't figure out what the problem was. We ended up rebuiling our cloud servers.

    This is clearly a bug you'd expect MS to patch through the standard updates, not a mere suggestion for a new, *optional* feature.

    Like so many other MS users/clients, I find it very frustrating that MS doesn't provide a place where to report bugs. At the very least, MS staff should escalate what they can clearly identify bugs and give you (as the originator) a case number so you are kept informed of development.

Feedback and Knowledge Base