I’m doing some testing of CHR in Hyper-V, but experience a problem with the CHR hangs or stop responding when running certain commands.
The problem only occurs when I have added more than one network adapter. The first network adapter is connected, and the other ones are currently disconnected. I don’t use the legacy adapter.
If I run the command “interface print” or “ip address print” the router stop responding for 10 minutes.
After the 10 minutes wait it output “action timed out - try again, if error continues contact MikroTik support and send a support file (13)” to the console. I can’t connect to the router on http or winbox.
I have tested with both the “bugfix” and the “current” version. It’s tested on both Hyper-V in Windows Server 2012 R2 and Windows 10 (1607).
If I use legacy network adapters, I don’t see this problem. But I would prefer to use new/synthetic network adapter.
I’m pleased to know that you are aware of the issue. I have downloaded 6.38rc31 (release candidate) and sinned a new VM up.
If I add a second NIC and type “interface print” in the console the router output:
Rebooting…
Stopping services…
failed to stop ipsec: std failure: timeout (13)
failed to stop ppp: std failure: timeout (13)
failed to stop wireless: std failure: timeout (13)
failed to stop route: std failure: timeout (13)
When it start up again I see this error in the log:
system,error,critical open /dev/panics failed
After another reboot I can type interface print again, and see both interfaces right away. But if an add a 3 NIC, then I can’t use the router. It keeps writing “Rebooting..” and nothing happens until I do a power off. I can’t use the router afterwards.
zyflex
Most annoying issue refers to live migration in Hyper-V cluster. I wrote about it here.
And the second issue I found is incomplete auto-negotiation on ethernet ports.
I hope that we will see a much better CHR for Hyper-V with version 7. Better multicore and a better Linux kernel for Hyper-V with correct tools.
My guess regarding Live migration issue is to blame the old tools. As you said in that post this was an issue that could occur on older Linux and tools.
To solve that issue you could add netwatch job that pings the hyper-v host or something else close every second. This should do the trick. Don’t solve the problem but it’s a workaround
Test the new 6.38rc34 release. This is running for me now and all these hangings problems are gone. If we are talkning about production I would wait for next current release which may contain these fixes.