Hyper-V host connectivity lost but VMs still communicate
-
So, the reason for my rant about the iDRAC console needing java is because I lost connectivity to a Hyper-V Server 2012 R2 host this morning.
All of the guest machines are still functioning.
The machine has 2 NICs and they are teamed.
The vSwitch is pointing to the Team.
There is not any separate network card for management.Anything to look at once I get Java going?
-
Are you certain the VM's are using the team, and not an individual NIC?
-
@DustinB3403 said in Hyper-V host connectivity lost but VMs still communicate:
Are you certain the VM's are using the team, and not an individual NIC?
.......
I specifically listed the details in the post.
-
I see that, but I've never seen a server lose its management interface with team'd nics and not also lose access to the VM's.
-
Does the host have a Static IP? Perhaps that has been taken by something else hence it cannot communicate but the VMs can?
-
@Jimmy9008 said in Hyper-V host connectivity lost but VMs still communicate:
Does the host have a Static IP? Perhaps that has been taken by something else hence it cannot communicate but the VMs can?
It does. This is a small office and the DC is running on the host.
-
Check if the DHCP server has leased the IP which was statically assigned. Check if DNS has something else listed for that IP address. Sounds like the IP address for the server is taken, hence no response, but the VMs have different IP addressed and work as they don't have duplicates.
-
The only time I've seen something similar is on XenServer if XAPI is hung for one reason or another. The VM's will still operate, but accessing the management console is down.
-
@DustinB3403 said in Hyper-V host connectivity lost but VMs still communicate:
I see that, but I've never seen a server lose its management interface with team'd nics and not also lose access to the VM's.
This is the second time this has happened, and is now something I need to troubleshoot.
Once I honestly could care less about because it could have just been some freak thing. I did check a few logs then, but nothing in detail, and I did not see anything off hand.
The NICS are standard Dell Broadcoms. The team was made via powershell from the Hyper-V console back when it was all setup.
-
The Hyper-V dom0 is available locally, but I am not. Hence wanting iDRAC console.
When this happened last time I stepped a user through shutting down the VMs and rebooting the server.
But it was also at 4:30pm, not 9am.
-
This is the only thing I can find that might be the culprit.
Take a look
-
This is the workload on the host
- Old SBS 2008 finally decommisioned 3 months ago, just not deleted
- Current DC
- local exchange 2013 /sigh
-
@DustinB3403 said in Hyper-V host connectivity lost but VMs still communicate:
This is the only thing I can find that might be the culprit.
Take a look
I found that one too, but I am not using the broadcom teaming. I am using native Hyper-V teaming.
-
Just to ask, did you see those event logs on the host at all (I know I know you're using the hyper-v teaming)
-
So this says at the VM level the intermittent connection issue, but people here have reported it affecting the host as well..
Edit: Update the drivers to the NICs
-
@DustinB3403 said in Hyper-V host connectivity lost but VMs still communicate:
Just to ask, did you see those event logs on the host at all (I know I know you're using the hyper-v teaming)
No there are no even logs at all related to networking. In fact, the host still says it has internet access even though it cannot ping out.
Also this is Server 2012 R2 +GUI + Hyper-V Role. I forgot about this client using this. I am so used to everything I have setup being Hyper-V Server.
The Errors are all related to the Veeam 4am backup crashing.
-
There is also the recommendation to disable "Virtual Machine Queue" on both the host and the VM's to address this issue.
-
@DustinB3403 said in Hyper-V host connectivity lost but VMs still communicate:
There is also the recommendation to disable "Enable Virtual Machine Queue" on both the host and the VM's to address this issue.
Always and forever disabled...
-
I just rebooted a CentOS 7 VM on the host and it lost network connectivity, cannot even ping the local gateway.
So it looks like networking is shitting itself, but online VMs are still working.
But it pulled a DHCP address...
-
I bet if you rebooted the host the issue goes away (at least for a short while) what version drivers are installed for these NICs?
