none
2012 R2 HV - SCVMM 2012 R2 UR2 - Linux Guest OS - Problem with Shutdown from SCVMM RRS feed

  • Question

  • Hello guys,

    I'm facing a problem with shutdown from inside scvmm 2012 r2 (ur2), after installing the Guest OS (CentOS 6.5 /w built in the newer LIS) and installed the scvmmagent x86_64 on the Guest OS i tried to shutdown the VM from scvmm console but i'm taking the error id 1735 about virtual guest services.

    Have you see it ? Is there any solution?

    Regards,

    ankso

    Wednesday, April 30, 2014 2:03 PM

All replies

  • Could you give us a little more detail on the exact sequence of events? Was the scvmmagent installed in your Linux image *before* you deployed the Linux VM with SC VMM? Or did you take some other path to get the VM created?

    And can you give us the exact and complete text of the error message you are getting?

    Thanks!


    Michael Kelley, Lead Program Manager, Open Source Technology Center

    Wednesday, April 30, 2014 3:18 PM
    Moderator
  • Hello Michael,

    Steps:
    1. Create the VM and then installed the CentOS 6.5 via iso
    2. Installed the scvmmagent with WinSCP -> upload the install.sh and scvmmagent.latest.version.tar
    3. chmod +x install
    4. ./install scvmmagent.latest.version.tar
    5. Trying to Shutdown via SCVMM 2012 R2 (UR2) console
    6. Error: The virtual machines cannot be shut down because Virtual Guest Services are not installed on the guest operating system (centos65-templateVersion1) or because the installed Virtual Guest Services do not support shutting down the guest operating system that is running on the virtual machine.
    To shut down, you must install Virtual Guest Services that support shutting down the guest operating system.
    ID: 1735
    7. That tried to do it before save it to the scvmm library and make it a template
    8. I've done also a shutdown -r now just to make sure and tried again the Shutdown but nothing happened

    Regards,

    ankso



    Wednesday, April 30, 2014 4:10 PM
  • The error message indicates that somehow your CentOS 6.5 is not running with the Hyper-V drivers. Within your CentOS VM, run 'lsmod | grep hv' and see if the Hyper-V modules are showing up as expected.  CentOS 6.5 has the Hyper-V drivers built-in, so if that aren't showing up, we need to go back to the installation process and figure out why not.

    If the Hyper-V modules do show up, then could you try running the standalone Hyper-V Manager console against this Hyper-V instance, and do a shutdown using Hyper-V Manager?  If it works with Hyper-V Manager, but not with SC VMM, the the problem is related specifically to SC VMM.


    Michael Kelley, Lead Program Manager, Open Source Technology Center

    Thursday, May 1, 2014 1:14 AM
    Moderator
  • I am having a similar problem, I have build 6 CentOS 6.5 VMs on a customers HyperV 2012 (not R2) platform.

    The tools problem is not constant across the VMs and some of the VMs work perfectly and management tools report the status and allow the VM to be shutdown properly but sometimes the same VM will just not report status at all.

    When I log into a working and none working VM I can't see any difference in the modules loaded or the services running, and comparing the dmesg logs between different VMs doesn't show any differences.

    I have remote shell access to both a working and a none working VM if there are any diagnostics I can extract from the HyperV tools on the VMs.


    Thursday, May 1, 2014 8:21 AM
  • Hello Michael,

    The problem solved (i still don't know why) when i re-created the VM from inside the SCVMM Console and not from the Hyper-V Manager. When i installed CentOS 6.5 and the scvmmagent all done well. Also i cannot figure out why i'm facing this problem.

    Regards,

    ankso

    **Update No2**
    I can't believe it, i tried now to re-create again a VM from SCVMM Console and after installing the scvmmagent.tar and trying to shutdown from scvmm console, i'm taking the error i said above.

    **Update No3**
    On the EventViewer i'm seeing these 3 Errors:
    1. Networking driver in CentOS65 loaded but has a different version from the server. Server version 5.0  Client version 3.2 (Virtual machine ID D0391E21-3A2C-424E-976B-B87AD62827D0). The device will work, but this is an unsupported configuration. This means that technical support will not be provided until this problem is resolved. To fix this problem, upgrade the integration services. To upgrade, connect to the virtual machine and select Insert Integration Services Setup Disk from the Action menu.
    2. Hyper-V Heartbeat failed to connect to virtual machine 'CentOS65' because the version does not match the version expected by Hyper-V (Virtual machine ID D0391E21-3A2C-424E-976B-B87AD62827D0). Framework version: Negotiated (0.0) - Expected (3.0); Message version: Negotiated (0.0) - Expected (3.1). To fix this problem, you must upgrade the integration services. To upgrade, connect to the virtual machine and select Insert Integration Services Setup Disk from the Action menu.
    3. Device 'Microsoft Synthetic Display Controller' in 'CentOS65' is loaded but has a different version from the server.  Server version 3.3 Client version 3.2 (Virtual machine ID D0391E21-3A2C-424E-976B-B87AD62827D0). The device will work, but this is an unsupported configuration. This means that technical support will not be provided until this problem is resolved. To fix this problem, upgrade the integration services. To upgrade, connect to the virtual machine and select Insert Integration Services Setup Disk from the Action menu.

    My Hyper-V hosts are with the latest cluster and hyper-v 2012 r2 updates and also all the important Windows OS 2012 R2 updates like the both of the SCVMM Nodes 2012 R2 UR2.

    From the linux viewpoint:

    [root@centos65]# cat /etc/centos-release
    CentOS release 6.5 (Final)

    [root@centos65]# lsmod | grep hv
    hv_netvsc          23702  0
    hv_utils                9149  0
    hv_storvsc         11323  2
    hv_vmbus         144850 5 hid_hyerpv, hv_netvsc,hv_utils,hyperv_fb,hv_storvsc




    Monday, May 5, 2014 5:49 AM
  • Hello, You can ignore the above warnings as they are extraneous. Please refer the following KB article:

    http://support.microsoft.com/kb/2956569/en-us

    The one regarding the Heartbeat driver is new and we are looking at it. However, I do not think that should be a problem.

    As per Michael's suggestion, have you tried to shutdown the VM using the Hyper-V manager? What is your observation there? Please let us know.

    Thanks,

    Abhishek

    Wednesday, May 7, 2014 1:49 AM
  • When i try to shutdown the Guest OS from inside SCVMM console i'm taking the error about VGS (virtual guest services). As i said i'm running CentOS 6.5 that have built-in LIS and should play fine after installing the scvmmagent inside the centos.

    Ok about ignoring the above errors but the problem with shutdown exists. Aslo shutdown from the hyper-v manager works.

    Regards

    Wednesday, May 7, 2014 12:36 PM
  • OK -- this is key information. If it works from Hyper-V Manager, but fails from SC VMM, then we will go looking for a problem in SC VMM. I'm contacting the SC VMM guys now to see if they have any insight.   It sounds like the Linux Integration Services part is working correctly and doing the shutdown, but SC VMM isn't doing something right.

    @Tim JD Fletcher:  Are you seeing the problem with SC VMM, or with Hyper-V Manager?


    Michael Kelley, Lead Program Manager, Open Source Technology Center

    Wednesday, May 7, 2014 5:47 PM
    Moderator
  • Hi - I am experiencing similar issues wherein not able to shutdown guest VM (CentOS 6.4) from SCVMM Console whereas it works well when I am shutting down from HyperV Manager.

    Pl do share if you do have any resolution for the captioned issues.


    Monday, May 12, 2014 12:59 PM
  • @Michael Kelley and @Tim JD Flecher do we have any update about this because it's kind of high critical for us to make it work.

    If you need any additional information do not hesitate to contact me.

    Regards,

    ankso


    Monday, May 12, 2014 1:32 PM
  • I have attached a VMM log file here http://www.speedyshare.com/zqBWq/3NEWdbgview.rar

    As you can see  in the log "Waiting for VM state to change from Running to PowerOff", but this never happens.

    I'm using powershell to create a new centos 6.5 VM.I didn't have this problem with centos 6.4.

    Thanks.

    Tuesday, May 13, 2014 6:28 PM
  • @tasoss:  Can you identify the name/id of the VM that has the issue (in that log file)?  It would help identification of the problem in the log file that you provided.

    For all who are experiencing this issue, can you identify the following:

      • Guest VM operating system impacted (is it just Linux, just RHEL/CentOS?)
      • VMM Server version - is this specific to 2012 R2 UR2? 

      For any impacted VMs, please report the following:

      1. Were these VMs created after upgrading to UR2, or did they exist before?
      2. For these VMs, what is the value of their HasVMAdditions property? (From PowerShell run the following command: Get-VM <name> | fl HasVMAdditions)
      3. Let us know if the following 2 workarounds work. If only one VM is experiencing the issue, try the workarounds in this order. If more than one, please try them on separate VMs.
    1. Try to shutdown the VM directly from PowerShell: Get-VM <name> | Shutdown-VM
    2. Try to Refresh the VM from the VMM console, once the refresh succeeds, try to shutdown from the VMM Console

    Thanks,


    Kris


    www.operatingquadrant.com


    Wednesday, May 14, 2014 5:08 PM
  • Hello and thanks for the reply.I had to log again because i couldn't remember the vm's id and name.

    So i have uploaded http://www.speedyshare.com/ab5eb/CentosLog2.rar

    It contains the log file and the get-scvirtualmachine command for this VM.

    The name is "foobar" and its id "30a71df9-c825-4cf8-b666-a19ab8ca057c".

    Thanks again!

    Wednesday, May 14, 2014 9:13 PM
  • @tasoss

    Thanks for the information.  Can you try these workarounds on that VM?

    Let us know if the following 2 workarounds work. If only one VM is experiencing the issue, try the workarounds in this order. If more than one, please try them on separate VMs.
    1. Try to shutdown the VM directly from PowerShell: Get-VM <name> | Shutdown-VM
    2. Try to Refresh the VM from the VMM console, once the refresh succeeds, try to shutdown from the VMM Console

    www.operatingquadrant.com

    Wednesday, May 14, 2014 9:32 PM
  • Thanks again for replying.
    I should have mentioned that when i shut down the machine from the hyper-v then the operation is completed(i logon to centos and then shut it down).Moreover if i use the gui of hyper-v manager and choose to shut it down then it also works.

    So about the workarounds

    1)I get http://pastebin.com/Hsb9gYkm

    2)Those actions are disabled in the VMM console.So i can't perform them.

    It happens on every single VM i create.


    ps:One more thing.After shutting down the vm from hyper-v,the job is completed in vmm console too and the VM is created.The problem is that i can't login to this new VM with the password that i have passed via the powershell.I have to put the old password.This means that some things didn't work during the creation of the VM(i suppose).
    • Edited by tasoss Wednesday, May 14, 2014 10:07 PM more info
    Wednesday, May 14, 2014 10:03 PM
  • @Kris Bash: I'm having the problem that i cannot shutdown the VM from SCVMM but i can operational shutdown it from hyper-v manager.

    1. Guest VM OS = CentOS 6.5 (built-in latest integration services BUT i want to mention that from the 6.5 Kernel missing the libvirt and hypervkvpd packages that are responsible for the IP Injection stuff. So i manually add them to work properly with IP injection on CentOS 6.5
    2. VMM Server Version = System Center Virtual Machine Manager 2012 R2 UR2 and latest windows server 2012 R2 updates.
    3. The VMs is after upgrading to UR2 they didn't exist before because then we started testing this.
    4. About Get-VM -VMName <name> | fl HasVMAdditions I don't take any output just empty lines.
    5. The workaround works properly, after the Get-VM -VMName 'testvm' | stop-vm and then refresh it on SCVMM console i saw in the tab VM Additions to say Detected and after that also Shutdown from VMM works properly.

    But this is not good workaround for us and for all the linux VMs that i wanna make. What is the next step @Kris?

    Regards,

    ankso


    Thursday, May 15, 2014 6:41 AM
  • @Kris: Kris we are working together with aksouzafeiris(i'm working as a developer and he is a sys admin).

    I'm sure that the problem has been existing since before UR2.

    Thanks!

    Thursday, May 15, 2014 7:57 AM
  • I also have this problem - but there is a lots of other input.

    First: CentOS 6.5 created with SCVMM 2012 R2, with (or without) scvmmagent,

    Linux centos65.localdomain 2.6.32-431.17.1.el6.centos.plus.x86_64 #1 SMP Thu May 8 02:18:11 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux

    Tested with kernels:

    2.6.32-431.11.2.el6.centos.plus.x86_64
    2.6.32-431.17.1.el6.centos.plus.x86_64
    2.6.32-431.5.1.el6.centos.plus.x86_64
    2.6.32-431.el6.centos.plus.x86_64
    2.6.32-431.el6.x86_64

    Loaded modules:

    [root@centos65 boot]# lsmod|grep hv_
    hv_balloon             12019  0 [permanent]
    hv_netvsc              23702  0 
    hv_utils                9149  0 
    hv_storvsc             11323  2 
    hv_vmbus              144850  6 hv_balloon,hid_hyperv,hv_netvsc,hv_utils,hyperv_fb,hv_storvsc

    Second:

    Few standalone Hyper-V hosts (2012 R2), all upgrades up to 2014-050-18, different hw (1 CPU, and 2 CPUs)

    Dynamic memory for Linux hosts.All integration services offered.

    Linux (CentOS) machines after few hours lose connectivity with hypervisor. HV Manager shows: Heartbeat: No. After restart machine is back to normal (Heartbeat shows: OK, No Application Data)

    The only solution I was able to find is to remove and reinsert hv_utils module from kernel:

    # rmmod hv_utils.ko --force

    # modprobe hv_utils

     And it works for a few hours.

    Tuesday, May 20, 2014 6:06 AM