VLAN Tags and Hyper-V Switches

April 15, 2013, 3:02 pm

≫ Next: VM Connection screen and hyper-v manager not in sync (black screen)

≪ Previous: NIC Teaming and Powershell Errors

Does the Hyper-V 2012 Virtual Switch support forwarding VLAN-tagged packets to a guest OS with the VLAN tags intact? In other words, can I have a single virtual NIC handle multiple VLANs by doing the VLAN filtering inside the guest OS?

I would like to run a guest OS that sits on multiple VLANs, and while I could create and delete virtual NICs which are assigned to a single VLAN, it would be much more flexible in my environment to have Hyper-V simply forward frames with the VLAN (802.1q) tags intact so that the guest OS can see the tags and deal with them appropriately. (looking at running a virtual router that sits across multiple VLANs).

I can't see any obvious way to do this. I thought that leaving the VLAN tag for the guest off would cause packets to be forwarded unfiltered, but that appears to not be the case. Does anyone know how to enable forwarding tagged frames through a virtual switch/NIC to a guest OS?

Thanks!

↧

VM Connection screen and hyper-v manager not in sync (black screen)

April 15, 2013, 4:20 pm

≫ Next: Find a physical machine for zombie vm

≪ Previous: VLAN Tags and Hyper-V Switches

Hi,

First time poster, please be gentle...

I have a problem which is similar to the black screen issue, my presentation is a little different though. I find that the HV Manager seems to "get stuck". For example, I make a HV connection to a shutdown machine, see the "blank" screen as expected. If I then start the machine in HV manager, the machine starts OK and comes up in the VM Connection. But in HV manager, nothing seems to be happening, as if the machine has not started. However if I go into the settings on the guest when I know it's started (sometimes after a little clicking around as it seems stuck at "Help" on the right-click) HV Manager then seems to wake up and starts to display stats and the thumbnail screen.

I did start using RSAT on Win7 and that seemed much more reliable and this issue has been more noticeable since I moved to Windows8. I've also started moving some disks to VHDX. As an aside, I noticed that the first time machines are started in the Windows8 console, they seem very slow to start but I'm not sure if that's coincidental. After the first time startup, they seem OK.

If anyone can offer advice/information as to whether these are known behaviours/issues, I'd appreciate that.

Martin

↧

Find a physical machine for zombie vm

April 15, 2013, 9:47 am

≫ Next: cannot complete proper settings for the creation of a virtual machine

≪ Previous: VM Connection screen and hyper-v manager not in sync (black screen)

Hey folks,

I have been given the task of finding the physical machine of a zombie VM .The problem here is we have don't have any records of this zombie machine computer account in active directory .We can ping the zombie vm but can't connect . we don't know the physical machine of that zombie . we have around 1000 servers mixture of pyhsical and virtual .we are using Hyper-v

Ask . how can we find the zombie vm's physical machine based on ip address

↧

cannot complete proper settings for the creation of a virtual machine

April 14, 2013, 2:40 pm

≫ Next: Storage LUN not visible on Windows 2012 VM using NPIV method

≪ Previous: Find a physical machine for zombie vm

This is a Windows 8 pro 64bit notebook with all of the hyper-v features enabled Created a virtual machine a number of times successfully but cannot connect to the virtual adapter. Additionally, I have pointed the settings to my DVD drive with the windows XP install disc but nothing is advancing. I am configuring with an administrator account.

Setting the virtual adapter lead to a command line where the DHCP server could not resolve. Images of what I have done are uploaded to a skydrive folder at: https://skydrive.live.com/redir?resid=52B21D2F3F75A51A!3770
How can I troubleshoot these issues?

↧

Storage LUN not visible on Windows 2012 VM using NPIV method

April 15, 2013, 12:10 am

≫ Next: All VM's locked up in Server 2012 Cluster!

≪ Previous: cannot complete proper settings for the creation of a virtual machine

Hello,

Few times, we observe that when we map (present) a Hitachi LUN on a Windows 2008 R2 or Windows 2012 VM (created on Windows 2012 Hyper-V) using NPIV method, LUN does not get visible on the VM system. When we reboot the VM system, LUN get visible on the VM system.

This happen approximately half of the times. In other cases, as soon as we map/present the LUN on the same VM, it get visible automatically.

I tried to find whether there is any known issue with Windows 2012 NPIV method but not able to find any official note about it.

Does anybody else also observing same behavior or any idea what could be the issue.

Thanks,

Nitin

↧

All VM's locked up in Server 2012 Cluster!

April 14, 2013, 2:02 pm

≫ Next: Hyper-V 2012 VM Placement by SCVMM 2012 SP1

≪ Previous: Storage LUN not visible on Windows 2012 VM using NPIV method

Hi,

About 2 hours ago this evening, we lost communication to all 20 VM's on a 8 Node Cluster and they took 25 minutes to recover! Fibre Channel shared storage is being used throughout the cluster through 2 redundant routes from each node. The host OS was still working fine. I appreciate this points to the storage going down but I really don't think it had, there are no errors on either FC switch and no error on either storage device. We are using Server 2012 on HP DL360's.

The Get-ClusterLog returns this error at the time they all went down:

00000370.000010d8::2013/04/14-20:12:25.115 ERR [RHS] RhsCall::DeadlockMonitor: Call TERMINATERESOURCE timed out by 6 milliseconds for resource 'SCVMM VM1'.
00000370.000010d8::2013/04/14-20:12:25.115 INFO [RHS] Enabling RHS termination watchdog with timeout 1200000 and recovery action 3.
00000370.000010d8::2013/04/14-20:12:25.115 ERR [RHS] Resource SCVMM VM1 handling deadlock. Cleaning current operation and terminating RHS process.
000006dc.000011b4::2013/04/14-20:12:25.115 WARN [RCM] HandleMonitorReply: FAILURENOTIFICATION for 'SCVMM VM1', gen(3) result 4/0.
000006dc.000011b4::2013/04/14-20:12:25.115 INFO [RCM] rcm::RcmResource::HandleMonitorReply: Resource 'SCVMM VM1' consecutive failure count 1.
00000370.000010d8::2013/04/14-20:12:25.115 ERR [RHS] About to send WER report.

Any help would be appreciated.

↧

Hyper-V 2012 VM Placement by SCVMM 2012 SP1

April 14, 2013, 2:12 pm

≫ Next: Unable to Partion on Unallocated Space

≪ Previous: All VM's locked up in Server 2012 Cluster!

Hi,

I'd like to understand the logic behind VM placement in SCVMM 2012 SP1 and a 8 Node Hyper-v 2012 cluster. Each Node has 32GB RAM and each is a completely identical HP DL360.

Here is the current placement as performed by SCVMM

Node 1: 1 VM

Node 2: 2 VM's

Node 3: 6 VM's

Node 4: 1 VM

Node 5: 2 VM's

Node 6: 7 VM's

Node 7: 1 VM

Node 8: 1 VM

There are 2 x 16GB VM's, the rest of between 2GB and 8GB. Each of the overloaded hosts has a 16GB VM running on it.

I have a cluster reserve set to 1 Node failure. Why has it decided to favour Node 3 and 6? I appreciate the Maths to work out Node suitability are very complex but this seems rather odd.

Thanks in advance.

Eddie

↧

Unable to Partion on Unallocated Space

April 15, 2013, 11:55 pm

≫ Next: IBM Blade HS21 support for Hyper-v 2012

≪ Previous: Hyper-V 2012 VM Placement by SCVMM 2012 SP1

Hi,

I have a server with Raid 10 configured for max drive space of 3500.00 gb available. During the installation of WINDOWS SERVER Hyper-V Core 2012, I created an OS partition C: (120 GB) and rest sholuld be for VM's 3,4 TB of space.

What I discovered through MMC Disk management is the following;

System Reserved 350mb
c:\ 120 GB
Unallocated: 3,4Gb

I created one partition almost 2 TB and now I have unallocated space 1,3 TB but I can't du nothing with it. When I try to create partition I have only 2 options. Properties and help. The unalocated space has no available options to utilize it.

Is there any reasonable solution to make use of this space without having to re-install the OS?

↧

IBM Blade HS21 support for Hyper-v 2012

April 15, 2013, 11:57 pm

≫ Next: Hyper-v 2012 Failback

≪ Previous: Unable to Partion on Unallocated Space

Hi ,

I was looking for a compatibility Matrix for IBM's Blade HS21 and hyper-v 2012 but couldn't find it .

Does anyone know if it's officially supported and where can I find the Matrix ?

thanks

↧

Hyper-v 2012 Failback

April 15, 2013, 3:50 am

≫ Next: Hyper-V in Server 2012 - Strange NIC issues

≪ Previous: IBM Blade HS21 support for Hyper-v 2012

I have built a Hyper-V Replica LAB with 2 Windows 2012 Hyper-V Hosts in a Workgroup environment and referred the below article to build it.

http://jsmcomputers.biz/wp/?p=360

My Servers are

HV01.domain.local- Primary Server

HV02.domain.local - Replica Server

I have successfully tested and performed a "Planned Failover and Failback" But When I am trying to perform an Unplanned Failover ,I am unable to failback to the Source.

I used the Powershell command "Complete-VMFailover" on the Replica VM . Thereafter I got lost of correct guide or steps in order to do the below

- Replicate the changes in the Replica VM back to the Primary VM (which is online now)

- And resume normal replication process from Primary --> Replica

Can anybody provide the exact steps to be followed during the Unplanned Failover and Failback

Regards

Muralee

↧

Hyper-V in Server 2012 - Strange NIC issues

April 16, 2013, 5:25 am

≫ Next: Hyper-V 2012 IDE controller + 4k sector VHDX - Windows 2012 Standard won't install

≪ Previous: Hyper-v 2012 Failback

I'm going to describe the issue as best as I can, but it's extremely odd to say the least...

We have 2 new servers, HP DL380 G7s. that are going to be the start of a new Hyper-V cluster. I built the first one up with minimum fuss, but the second has NIC issues as soon as I add the Hyper-V role.

From the four on-board NICs, two are in our domain VLAN 1, and the other 2 in their own VLAN 18. They will be teamed, but for the sake of simplicity I'll keep teaming out of this for now. Before adding the Hyper-V role I can ping out of every single NIC, after adding the Hyper-V role 1 of the NICs no longer respond to ping.

At this point I have not created any virtual NICs, nor changed anything in the physical NIC adapter settings

I remove the Hyper-V role and they all start to work again, I re-add the role and a different NIC stop working.

The driver and firmware versions are exactly the same as the server that currently works. Also, NICs 1+3 go to one of our DC switches, and 2+4 go to another, this is to add redundancy once they get teamed. However for testing I have now patched in the problematic server directly into the same ports the working server was using.

To add to the strangeness, if I create a virtual switch using the NIC that doesn't work I can then ping out of that.

Any ideas welcome :-)

Andrew France - http://andrewsprivatecloud.wordpress.com

↧

Hyper-V 2012 IDE controller + 4k sector VHDX - Windows 2012 Standard won't install

April 12, 2013, 2:52 pm

≫ Next: Re build old Hyper V 2008 server from scratch

≪ Previous: Hyper-V in Server 2012 - Strange NIC issues

Windows 2012 and the new VHDX format are stated to have full support for native 4k sector drives.

However if I create a VHDX file on hyper-v 2012 with native 4k (logical + physical sector sizes set to 4k), a windows 2012 Standard guest OS will not install, it gives the error:

Windows cannot be installed to this disk.

This computer's hardware may not support booting to this disk.

Ensure that the disk controller is enabled in the computer's BIOS menu.

however if I set it to a 512e VHDX where the logical sector size is 512k but the physical is 4k then it works fine.

... is the issue that the IDE controller in hyper-v 2012 doesn't support 4k sector drives?

I have read there is a negative performance impact when using 512e so I was hoping to take full advantage of Windows 2012 and native 4k sector drives.

thanks

↧

Re build old Hyper V 2008 server from scratch

April 16, 2013, 5:50 am

≫ Next: Hyper-V recommendations

≪ Previous: Hyper-V 2012 IDE controller + 4k sector VHDX - Windows 2012 Standard won't install

My org has an Hyper V 2008 r2 Hyper v cluster that is a mess. Licensed incorrectly, over provisioned, no HA or LM configured correctly. There are only 4 NICs no teaming, no dodicated CSV, LM, VM networks all cables patched to one switch etc

In order to allow for growth and configure the required resiliency that I insist upon I have recommended commissioning a new host and rebuilding the cluster from scratch using the new host server as Cluster Node one then decommissioning a node in the old cluster rebuilding and adding to the new cluster and then again with the third node decom , rebuild and add. So at the end My org is left with a highly resilient fault tolerant Cluster.

can i provision a new cluster then add a new host and add the OLD storage to the new cluster / host? I have been advised that additional space is required for the cluster / new host. Can the existing CSV storage not be added? and then the VMs migrated to the new Host so that the two clusters are effectively using the same CSV?

It must remain on 2008 r2 as the SAN is not certified on Hyper V v3

↧

Hyper-V recommendations

April 16, 2013, 1:46 pm

≫ Next: internal network, can't ping parent partition.

≪ Previous: Re build old Hyper V 2008 server from scratch

Howdy,

Could you all please recommend a good book that covers the basics and advanced options in Hyper-V?

Thanks.

↧

internal network, can't ping parent partition.

August 15, 2009, 9:48 am

≫ Next: Error on live migration

≪ Previous: Hyper-V recommendations

Hi,

I'm using Hyper-V Server R2 (not Windows 2008 with the Hyper-V role)
I create a new internal network to allows my VM to communicate with my host.
but I'm not able to ping the host IP of this internal network.
but the host can ping the VM ip address!!!

from hyper-v:
ping 192.168.253.2 (ip address of the VM) works

from the VM:
ping 192.168.253.1 (ip of the hyper-v network switch adapter) doesn't works!

any idea?

↧

Error on live migration

March 21, 2013, 3:12 pm

≫ Next: Using Integration Components in a mixed environment

≪ Previous: internal network, can't ping parent partition.

Hello,

I have been doing a live migrations with several machines and I am having an issue with only one of the VM's. It will not do a live migration and gives me the following error.

Error (12700)
VMM cannot complete the host operation on the (ServerName) server because of the error: Storage migration for virtual machine 'SERVER' (2C4C28E1-791A-4474-A74C-1D4E7436D134) failed with error 'Data error (cyclic redundancy check).' (0x80070017).

Migration did not succeed. Mirror operation failed for the source VHD file 'I:\SERVER\SERVER.vhdx' to the destination file 'D:\IMAGES\SERVER\SERVER.vhdx': 'Data error (cyclic redundancy check).'('0x80070017').
Unknown error (0x8000)

Recommended Action
Resolve the host issue and then try the operation again.

Anyone have any suggestions? the server is running fine and there are no other errors that can lead me to the problem?

Thanks,

Gavin...

↧

Using Integration Components in a mixed environment

April 16, 2013, 10:50 am

≫ Next: Hotfix kb977837

≪ Previous: Error on live migration

Can I use Integration Components in a mixed environment?

Examples:

Hyper-V Server 2012 Host running Guest Machines with OLD (win2008, R2) hyper-v integration services version installed

Hyper-V Server 2008 R2 Host running Guest Machines with NEW (win2012) hyper-v integration services version installed

There are problems? Bugs?

↧

Hotfix kb977837

April 16, 2013, 6:17 am

≫ Next: Hyper-V Server 2012 Local Storage Provisioning - Best use of multiple arrays?

≪ Previous: Using Integration Components in a mixed environment

Hi Everyone,

Just would like to check if anyone has used the hotfix http://support.microsoft.com/kb/977837

We are encountering the error mentioned in this article but we just want to be sure that this can be uninstalled if the need arises.

Thanks everyone

Regards,

Eman

All the best, Eman Lacuata

↧

Hyper-V Server 2012 Local Storage Provisioning - Best use of multiple arrays?

April 16, 2013, 12:08 pm

≫ Next: Snapshot gives unknown error

≪ Previous: Hotfix kb977837

All,

I'm planning a deployment of Hyper-V Server 2012 for a home development environment to be used for a combination of prefessional and hobby software development projects Right now, I'm looking at my storage configuration options. Virtualization storage is a vast subject, and I believe my scenario should be relativley straightforward simply because of the scale. So, I thought I'd ask for some community advice.

I'll be using local storage on a Dell PowerEdge 2950 III host, which has a PERC6/i RAID Controller, and 6x1TB HDDs.

I'm considering a deployment of Hyper-V to a UFD. It seems to me that is a reasonable middle-ground between having the host O/S and VMs on the same array (adversely affecting performance) and dedicating a separate array to the host O/S (wasting unused disk space).

I'm planning a 2TB RAID 10 array (4x1TB HDDs) as the primary data store for my VMs. This leaves 2x1TB HDDs unallocated, which I would put in a separate RAID 1 array. The question I'm left with, is what are potential good-uses for the 1 TB RAID 1 array? Some of my initial thoughts on this are:

1. Expose a Virtual Disk/LUN for host paging - the UFD deployment procedures for Hyper-V disable paging to reduce ware on the UFD. So, would it make sense to create a virtual disk/LUN forhost paging?

2. I might like to tinker with adding pass-through disks to a VM running SQL Server for database log and data files. So, again I could configure the PERC6/i to present dedicated virtual disks/LUNs from the RAID 1 array for this purpose. Would that seem appropriate? I know it's not as ideal as using separate arrays for SQL VM, data files, and log files...but I'll only have two arrays to work with.

3. Any other suggestions to make good-use of the RAID 1 array?

Thanks in-advance!

↧