Hi,

Lately a Hyper-V hosts exhibit problems restarting/shutting down certain VMs.

The guest OS seems to work OK, chkdsk and sfc /scannow don't give me any errors.

However when I shutdown or do a restart of the VM it stalls - I have to turn off the VM (shutdown doesn't work at that stage), which is not a good thing.

When later restarting the VM it sometimes refuses to start (maybe it got damaged by turning it off?).

The way I got it working again was attaching a repair ISO (not the Windows disk) and checking boot records and disks from outside the guest OS. This did work but it doesn't make sense to me: we're talking VM's here, so not sure how the 'physical' concepts of boot partition/record etc exactly map to the 'virtual' ones in Hyper-V.

I'm worried because with all the updates the hypervisor need rather frequent rebooting so this problem could easily get worse.

So finally my question: how does one best handle these kinds of booting/restarting problems in Hyper-V?

Is it advisable to create a new VM (with exactly the same parameters) and attach the VHDX to that VM, getting rid of the old VM in the process (and hopefully getting rid of these 'external' problems).

Sorry, no expert in these matters. Any and all help much appreciated.

Best,

Bavo

PS: I also ran some checks on the hypervisor itself (chkdsk, sfc, dism etc) and it seems OK

I have a 3-node 2012R2 Hyper-V cluster and have updated the cluster with 2 new servers. I want to remove the last old server from the cluster but it is always the current host server. When I patch and move the role off of that machine for the reboot it just goes back to that node. I was thinking that if I remove the old server from a potential owner of the quorum the it couldn't be the host server but don't know that that is correct. I didn't want to evict the node if it is somehow preferred in the system and an eviction would leave no preferred server. Couldn't find an answer in the forums.

As always appreciate your help.

eburch@lasertel.com

Hi everyone,

I purchased Windows 10 Pro upgrade from Home edition so I can have access to Hyper-V to develop with Docker. I know; I could've just made a virtual Linux machine or installed a Linux OS, but I nonetheless decided to stick with my base Windows OS because I thought it'd be convenient at the time. Basically this is what happened:

1. I purchased an upgrade to Pro for $99
2. For some reason I ran into an error in the activation settings. I forgot what the error was, but I was having trouble getting my upgrade activated.
3. In my poor judgment, I instead searched for a Pro product key online and used that to activate.
4. So now, I'm stuck with my Pro edition but still unable to use Hyper-V and other Pro features (it fails to update on restart likely due to duplicative ownership issues with the key).

Is there any way to fix this or anyone I can contact? MS's default online support was not very helpful.

Thanks.

Hi All,

I've just setup a Server 2016 with (2) NICs running Hyper-V role and connected to an HP ProCurve.  Here's how I set it up:

On the HP Procurve I setup the ports like this: trunk C21-C22 trk5 lacp

On the server it's setup like this.  Do you see any problems with this setup?  We only have one switch so please don't comment on redundancy.  Just setting up the two 1GB NICs teamed.

Good afternoon.  I am in the process of creating a Microsoft Server 2016 VM with Hyper Visor.  I came across some guidance from the following link:  https://solution-soft.com/sites/default/files/wysiwyg/Configuring%20Microsoft%20Hyper-V%20Virtual%20Machines%20to%20use%20a%20Static%20MAC%20address.pdf

which states,

NOTE: Microsoft Virtual Machines static MAC addresses must be within the following range:              

Minimum Mac Address = 00-15-5d-01-80-00              

Maximum Mac Address = 00-15-5d-01-8F-FF

So, must I? 

The reason I ask is that we are converting a physical server into a VM and retiring that physical server.  One of the programs on that server has licenses that are tied to that server's MAC.  I'd like to reuse that MAC.  As long as we avoid network collisions on our small network (about 20 physical and virtual machines) which is a closed loop system (i.e. not connected to the internet), I don't understand the above must statement.

por que me sale el mensaje

   "No se pudo obtener informacion del disco.

     El almacenamiento donde se encuentra el disco duro virtual no admite el uso compartido de discos duros virtuales"

Hello,

I am trying out Hyper-v to compare to KVM.  I installed a bare-metal server with the free Hyper-V 2019.  It is not on any domain and it has a working IP address.  I am able to RDP into it.  I have a Windows 10 desktop to manage it.  I cannot seem to get the Hyper-v manager from windows 10 to connect to it.  What are the exact steps that are needed?  I have created a guest vm on my desktop and the hyper-v manager can connect to and manage it just fine.  Help would be greatly appreciated!

Hi all

I have a hyper v cluster of 5 hyper v all running Windows 2016 Build 1607 core mode.

I have noticed in one of them there is a c:\programdata\Microsoft\Windows\Hyper-V\Planned Snapshots folder with huge vmrx files and a smaller vmcx file with the same guid.

When I open these vmcx file in notepad I can see some of the VM and disk mounted. However these VMs do not exist on this particular hyper v

May I know what this folder actually I does?

The VMs are stored on a separate San storage

Morning all,

We've recently trialled an in-place upgrade of our Cluster servers from 2016 to 2019 Datacenter edition. It all went very smooth and everything seemed to be working.

However - a week later and suddenly the cluster is failing.

A validation report shows that the three nodes cannot talk to each other on any network via UDP Port 3343.

I've opened the necessary ports on the firewall - but it's still happening.

It is generating a lot of errors (some critical) in the event log for Failover Cluster Manager.

I'd be grateful if anyone could help or offer advice.

Thanks

Gareth

Hello community, i'm after some licencing clarity please.

We have a customer who has an existing Server 2012 Standard, with HyperV role and runs 2 VM's. This host OS was purchased as HP ROK.

Now the customer wants to replace this server with new. So, we would need to licence the new server. Am I right in thinking that I would need to purchase a server 2019 standard open licence, use downgrade rights and downgrade the host to Server 2012 to match the 2 server 2012 vm's as they are technically not licenced?

Cheers

I had virtualized a physical server2k last year, since the actual physical hardware was coming up on 10+years of service.  This server has a custom written windows service, used for RFID scanning; the service records the comings and goings of products flowing through our facility, to a ms-access database.  We're in the process of having this custom windows service, upgraded, to work from a server 2016 64bit, but it's looking like the re-write won't be for at least another 6 months.

Last year, after successfully converting the physical server to virtual, I had expected to see an increase in performance from the virtualized server2k, since the retired physical server had 100mb network, 1gig ram, old disk drives and the new hyper v host had 64gig ram, raid10 ssd, gigabit networking.  I've configured this VM to have 2 virtual processors, 4gig ram (max for server 2k); the vm's network status reports 1.4gbs speed; users accessing the shared access database from this virtualized server, over the network, experience slow performance when querying the database.  There seems to be no increased performance, since going virtual (performance is same as before).  Does anyone have idea's, as to how I could increase performance for this vm?  The same access database shared from a server 2016 hyper v on the same hyperv host, is much faster, but of course, the server 2016 vm settings are configured with 32gb ram and 8 virtual processors...

I'm new to hyperv, so any advice would be awesome.

Robert

Hi All

I am using Hyperv 2016, i have a Virtual Machine and this virtual machine has a snapshot created during backup by backup tool. Now I am not able to delete the snapshot as the option is grayed out, i cannot shutdown the virtual machine as it is very critical. How do i delete this snapshot. Experts help me on this.

Mouse Not Captured In Remote Desktop Sesion" in Hyper-V

is there something about a windows 201 vm on 20 hyperv host

Hi all,

I have enabled Hyper-V-VmSwitch logging on a Hyper-V host (Server 2012R2), however no log entries seem to appear.

The same counts for: Microsoft-Windows-Hyper-V-VMMS-Networking

Other Hyper-V (networking) logs seem to work fine, for example: Microsoft-Windows-Hyper-V-SynthNic-Admin is returning log entries.

What am I missing, do I need to enable a setting on the Hyper-V side?

Thank you in advance!

Regards,

Mervin

So, maybe third time's the charm posting on the forums (hopefully the best forum for the question this time):

I'm copy/pasting this from the regular MS forums as suggested by a moderator.  I view this as more of a host OS issue than a Hyper-V issue personally, but defer to the experts.

So, it's taken me a while to track down exactly what process was causing this.  I've attached the first evidence I've found as to what process is causing this problem.

Since my account on this side isn't verified, the screenshot can be viewed here:

https://answers.microsoft.com/en-us/windows/forum/windows_10-performance/w10-1903-in-vm-hanging-with-system-process-using/0c30370e-823a-41f1-bdfc-37c685ee001b?tm=1572380956433

This is a virtual Windows 10 Pro 1903 installation on a Server Datacenter 2016 Hyper-V host.

Host has a single Xeon E3-1275 v6 quad-core processor.  64GB RAM.  Host is installed on an SSD while the VMs and their storage are kept on a pair of traditional HDDs in RAID1.  There are a total of 8 VMs running on this host.  One is the Linux-based virtual appliance for our on-premise Bitdefender security product, six are additional Windows Server VMs for file servers, print server, WSUS, a redundant DC, and some internal web-based applications.  The last is this problem child Windows 10 VM, and it is the only one experiencing issues.  All of the Windows-based VMs are Generation 2 and have access to a single vCPU.  Memory has been assigned based to each VM on demonstrated need, at least initially.

Around 3 months ago (not sure exactly, but certainly not long after we approved the update to 1903) this VM began not checking in to WSUS and when I would look into the issue, Hyper-V manager would report the VM was using about 12% of the processor, which translates into 100% in the virtual environment.  Requests to remote in would time out, and connecting through Hyper-V Manager would show the lock screen which was marginally interactive, but it was unable to proceed through logon.  The only way to get it to respond is to reset the virtual machine, after which it starts up normally.  However, it could freeze again in as little as 15 minutes, or it could take over a day, without any discernible pattern.

This system is used to host the only Java installation we have on campus, which it needs to access HVAC controls for our campus, as well as monitor some of our wireless access points.  Other than the browser, and Java, there is no additional software (except for that which I've downloaded for troubleshooting since the beginning of this issue).  The system had been working fine for over a year with only regular updates being done.  As it's a VM, there's no physical hardware for which drivers could be out of date.  The VM had been initially assigned 2GB of memory, however, that has since been incrementally increased to 8GB with no success to mitigate this issue.

Since the time between system startup and hang appear to be somewhat random, it's not usually possible to notice when it hangs.  An individual in another forum asked about licensing, and, while we are licensed for this scenario, it would seem to be a strange method by which to enforce licensing in the first place.  In the case of the screenshot attached, it froze minutes after I left my desk for the day, but before the problem VM's lock screen came on.  I left screen sharing software running in the hopes of catching it in the act and finally succeeded.

Any suggestions for this one?  Certainly it's possible to just reinstall from scratch, but if a solution exists, I'd rather find that than just keep nuking and paving the VM every time it decides to try to turn into a space heater.

Best wishes,

Roy

Today is the third time in five months I've had this issue, and any help is appreciated.

Hyper-V Host HV1 has two VMs, DC1 and APP1.  DC1 and APP1 use a Hyper-V switch to connect to the LAN.  Host and VMs are on the same subnet.  

Sporadically, the guests lose the connection, cannot ping each other, or the host, or their default gateway.  

The only solution I have found that works is to shut down the VMs, remove the virtual network adapter, delete the Hyper-V Switch, then build a new Hyper-V Switch using the same NIC on the host (the host only has two on board NICs, plus iDrac.)  Then assign that to each VM and boot the guests to their OS.  That works, but it then crumps again a few weeks later.  

Tidbits: DC1 is Gen1 VM, while APP1 is Gen2 (DC1 is Server 2003, if that matters... Don't ask why, it's a long story... and APP1 is Server 2012 R2 Standard).  
Host is Server 2012 R2 Standard
HV Manager is 6.3.9600.16384
Host is up to date on firmware from vendor website
Host is fully patched with OS updates
Dell's built in hardware diagnostics show no problems.

I can get log files, if you wish.  

Any help or ideas is _greatly_ appreciated.

Best, 

Jeff

Hi Support,

I have just installed a new server Windows 2019 as a host and as a VM.  This is a new HP DL380, 32 GB ram and raid 5, 15k drives, server and everything is updated. everything looks good in the VM, I have dynamic memory and 10 core cpu in the settings.  the issue is any file transferred between the vm and other machines is extremely slow.  the issue does not exist for the host.  The server does not use the Broadcom network card. There are no applications only an RMM agent, Bit Defender antivirus and Datto backup agent.  The size of the VM is 800 GB.   I appreciate any help. 

thanks,

Jamshid

We have two functioning Hyper-V clusters running Windows Server 2016 Datacenter. Each cluster is at a different site. While each cluster uses two different server types and different number of cluster nodes, all other hardware, switches, storage, etc. is the same.

In June we performed normal updates on both clusters and all VMs hosted on those clusters. Shortly after, the process of pausing a node to drain/stop it would fail with another node going into isolation. This also caused CSVs to go unavailable for a time and the VM workload on that node to be stopped and restarted. After a month or so, the host isolation events disappeared, however CSV volumes would still fail to completely go offline or come back online while drain/stopping a host node. Again, this caused the VM workload on the host to be stopped and restarted.

It would seem that the process of VMs live migrating and the CSVs moving at the same time caused the CSVs to hang on occasion. This occurred randomly, and was not guaranteed to cause the problem each time. However, live migrating the VMs manually and then moving the CSVs or vice versa would allow the node to be paused without issue. As such, finding the cause of the problem proved to be very difficult.

In the end, we built up a separate cluster on much different hardware, but again, the same supporting storage, switches, etc. We were able to reproduce our problem and make it occur more frequently by increasing the IO workload on the CSV volumes while pause/draining a host node. When the failure happened, we often got Microsoft-Windows-FailoverClustering Event ID 5142 - Cluster Shared Volume 'CSV5' (CSV5) is no longer accessible from this cluster node because of error '(1460)'. But not always. For whatever reason, CSV volumes were not fully transitioning when being moved to a new host node while live migration was being performed.

After attempting to rebuild the cluster with an older update, resetting the switch firmware and a number of other trouble shooting steps, we found the answer to be Windows Defender. It appears the Realtime Monitoring feature would cause this problem and disabling it using either of the two steps fixed our problem completely (on Host cluster nodes):

1. Start -> Settings -> Update & Security -> Windows Defender ->Real-time protection (set to off)
2. From an administrative PowerShell command: Set-MpPreference -DisableRealtimeMonitoring $true

This can be verified with the Get-MpPreference command and checking the value ofDisableRealtimeMonitoring. Or you can check the status of the WdNisSvc (Windows Defender Network Inspection Service) service. This service should go to the stopped state when Realtime Monitoring is disabled.

This is extremely surprising since Microsoft's own documentation states that when specific roles are installed exclusions are automatically added to Defender.

https://docs.microsoft.com/en-us/windows/security/threat-protection/windows-defender-antivirus/configure-server-exclusions-windows-defender-antivirus#list-of-automatic-exclusions

Additionally, these automatic exclusions generally match the recommended exclusions for Hyper-V.

https://support.microsoft.com/en-us/help/3105657/recommended-antivirus-exclusions-for-hyper-v-hosts

Again, using the Get-MpPreference command, one can check the value ofDisableAutomaticExclusions to see if they are being used.Even if you manually add the recommended exclusions, it appears as if Windows Defender ignores these. Only disabling Realtime Monitoring fixes the problem we were experiencing.

I was surprised that I was unable to find anyone else experiencing this issue since we are not doing anything out of the norm for our clusters. Everything is pretty much a basic Windows installation and configuration. Any additional help about what we might be doing wrong to cause this problem would be great.

Thanks.

Hi.

Our network guys recently reported that they're observing flapping MAC addresses between physical switches.

After analysis, we found out that the reported MACs are VMs' MACs on various standalone Hyper-V hosts. We use SET vSwitches (Switch-embedded teaming with Dynamic distribution) on all hosts. The flapping MACs are only from some VMs on some hosts and in case also MACs of the pNICs on the host.

Each Hyper-V host is physically connected using 2 (or 4) pNICs with 1 (or 2) connections to every physical switch (total of 2).

According to documentation (https://github.com/MicrosoftDocs/windowsserverdocs/blob/master/WindowsServerDocs/virtualization/hyper-v-virtual-switch/RDMA-and-Switch-Embedded-Teaming.md#mac-address-use-on-transmitted-packets) this shouldn't be happening and Hyper-V should be replacing the MAC when communicating over the non-affinitized member.

Any ideas what could be wrong? Some misconfiguration or bug?