VirtuallyAware

Experiences in a Virtual World

Archive for December 2007

VirtuallyAware.net

leave a comment »

You can now get to my spaces site by just typing http://www.virtuallyaware.net.  I may host this on my own web server soon, but for now, it will point to where you are now.  Anyway, thanks for stopping by whichever way you got here.

Written by VirtuallyAware

December 11, 2007 at 3:42 PM

Posted in Uncategorized

Update – KB941125 – The Virtual Server service may stop responding when the service is starting if one or more of the virtual machines are configured to automatically start when the Virtual Server service starts

with 3 comments

*****************************************************************************************************************************************************

Update 12/24/2007

After some further testing, the problem with Virtual Server 2005 SP1 continues. 

Notes:

KB941125 did help some.  It eliminated the complete freezing of the web console, but the guests still get hung up and are unresponsive to any commands.

Setting the resource allocation to the default 100 does not seem to be any help.  All of my guests have been at the default 100 and I have seen the re-occurrence of the guest freezing. 

We continue to test one of our HP DL585 G2 systems with VS 2005 R2 SP1, but have returned the other one back to VS 2005 R2 without SP1.  That server has not skipped a beat since the switch.  Since the older version knows nothing about hardware virtualization, that might be the road to look at in troubleshooting the VS 2005 R2 SP1 system. 

Other options tried:

AMD Opteron™ Processor with AMD PowerNow!™ Technology Driver Version 1.3.2.0053 for Windows XP and Windows Server 2003 (x86 and x64) -  This was to solve the problem with time shift between processor cores that would give negative ping times and would show event id 1054.  Below are some articles that outline this.

http://www.amd.com/us-en/Processors/TechnicalResources/0,,30_182_871_9033,00.html

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=110&prodSeriesId=463637&prodTypeId=12169&objectID=c01075682

http://support.microsoft.com/kb/938448/en-us

*********************************************************************************************************************************************************

Let me start out by saying that my experience with Virtual Server and all the subsequent versions that followed have been relatively bug free.  This is the first one that really stopped me from progressing to Virtual Server 2005 R2 SP1.  At first I thought what I had was a corrupt install, but after a few rebuilds of the systems and looking at firmware and drivers everything looked fine.  The main symptom I saw was that every time I put the servers into saved state either manually or with a script, when restoring to running state, the web interface would not be accessible nor would the VMRC or VMRCplus or any type of communication.  The guest servers would not ping or be able to be shutdown with a script.  The whole Virtual Server COM API was not accessible.  The only way to fix this would be to kill the vssrvc.exe process, restart the virtual server service and then discard the saved states of the guests.  If you tried to restore the saved state of the guests, you would go into the inaccessibility again.  Upon starting the guest servers back up from an off state, everything would work just fine, until the next saved state which in our case is sometimes during our scripted backups or on a shutdown script we have in place or when I felt like save stating any guest.  We do have 2 other virtual hosts running Virtual Server 2005 R2 SP1 that seem to be working just fine.  The 2  hosts that were showing this problem are DL585G2 systems with AMD-V.  Not sure if this is linked to the overall problem, but the DL585G1(AMD without AMD-V) and DL380G4 (Intel Xeon without VT) have not experienced this problem at all.  Lets hope this is the last of the bugs for now.  At least until Hyper-V.

 

*****************************************************************

Attached is the news group string.

So, do you have to install the patch and do one of the workarounds or
"should" the patch be all that is necessary?  Do you have any experience
with the effectiveness of KB941125?
Rob McShinsky
"Benjamin Armstrong [MSFT]" <benarm@online.microsoft.com> wrote in message
news:uMg6X8eNIHA.3516@TK2MSFTNGP02.phx.gbl
> Yup.
>
> Cheers,
> Benjamin Armstrong
> ============================
> Windows Virtualization Program Manager
> Blog: http://blogs.msdn.com/Virtual_PC_Guy
> Book: http://tinyurl.com/ysxcbm
>
> This posting is provided "AS IS" with no warranties, and confers no
> rights. You assume all risk for your use.
>
> Microsoft News Groups wrote:
>> Do you know anything about KB941125?  Sounds alot like what you are
>> outlining.
>>
>> Rob McShinsky.
>>
>> "Benjamin Armstrong [MSFT]" <benarm@online.microsoft.com> wrote in
>> message news:O8tUZheNIHA.484@TK2MSFTNGP06.phx.gbl
>>> First:  I would recommend that you contact Microsoft Product Support so
>>> that we can get formal cases open against this.
>>>
>>> Second:  Try creating a blank virtual machine (with no VHD) and connect
>>> it to all of your virtual networks.  Configure that virtual machine to
>>> start first, and then have the other virtual machines start with a delay
>>> after that.
>>>
>>> Cheers,
>>> Benjamin Armstrong
>>> ============================
>>> Windows Virtualization Program Manager
>>> Blog: http://blogs.msdn.com/Virtual_PC_Guy
>>> Book: http://tinyurl.com/ysxcbm
>>>
>>> This posting is provided "AS IS" with no warranties, and confers no
>>> rights. You assume all risk for your use.
>>>
>>> Microsoft News Groups wrote:
>>>> Hey Ben;  I am seeing that you are also seeing this from other people
>>>> as well 9/5/2007 "CGI Timeout – VS 2005 SP1 Admin Site".  We are seeing
>>>> this also when we do a startup script to restore saved guests to a
>>>> running state (We do a shutdown script prior to put them in saved state
>>>> and create a log file the startup script reads).  Seems as more and
>>>> more people are popping up with this issue with VS 2005 R2 SP1.  What
>>>> is your recommendation?
>>>>
>>>> Thanks
>>>>
>>>> Rob McShinsky
>>>>
>>>>
>>>> "Microsoft News Groups" <Please@UseGroup.com> wrote in message
>>>> news:eb7zYe8MIHA.1188@TK2MSFTNGP04.phx.gbl
>>>>> I will certainly check this out on my system.  I made another
>>>>> adjustment today after I say a single server "freeze" today.  I still
>>>>> had to shutdown the other guests and then kill the vssrc services to
>>>>> get control again. This was with all the guests with the same, default,
>>>>> values.  So what I did was remove all guests from the console, then
>>>>> delete the VMC files for all of guests, and then did a Create again to
>>>>> re-establish the vmc again.  I did this because I was doing a little
>>>>> comparison between VMC files that were created for guests native on
>>>>> box, vs. guest that were migrated over from another SP1 box  but
>>>>> without  AMD-V or from a machine that is not running SP1 or AMD-V.  It
>>>>> seemed that the VMC (xml) files were structured different. They have
>>>>> not been running long enough to determine if this will help, but it
>>>>> seems a little strange.  Maybe Ben might have an idea on if this could
>>>>> potentially be causing the problem.
>>>>>
>>>>> Here’s to more testing.  Let me know if you have any further progress
>>>>> or want to try some different testing scenarios.  I have 2 DL585 G2
>>>>> systems to test on. 1 built an the other not.  I moved to the current
>>>>> testing host because I experienced the same problem on the other host.
>>>>> I thought it was a corrupt install.  The same guests that were on that
>>>>> guest are now on the new problematic host.  My hopes is that one of
>>>>> the guests is just messed.
>>>>>
>>>>> Rob McShinsky
>>>>> Dartmouth Hitchcock Medical Center
>>>>> Lebanon, NH
>>>>>
>>>>>
>>>>> "Stephan Kaufmann" <StephanKaufmann@discussions.microsoft.com> wrote
>>>>> in message news:791DF571-457B-4089-8CF0-519AFDE173ED@microsoft.com
>>>>>> Hi Rob
>>>>>>
>>>>>> Very intresting!
>>>>>>
>>>>>> I have 6 VMs (2 Windows Server 2003 =SV000x  and 4 Win XP=VP000x) on
>>>>>> my
>>>>>> DL385 G2 with 10GB RAM. If I set the following values for "Resource
>>>>>> Allocation", I always get the freezing problem as soon as one of the
>>>>>> VMs with
>>>>>> lower priority boots:
>>>>>>
>>>>>> 1st value: Relative Weight
>>>>>> 2nd value: Reserved Capacity (% of one CPU)
>>>>>> 3rd value: Maximum Capacity (% of one CPU)
>>>>>>
>>>>>> SV0001: 100, 5%, 100%
>>>>>> SV0002: 100, 5%, 100%
>>>>>> VP0001: 50, 0%, 80%
>>>>>> VP0002: 50, 0%, 80%
>>>>>> VP0003: 50, 0%, 80%
>>>>>> VP0004: 50, 0%, 80%
>>>>>>
>>>>>> All VMs have the hardware-assisted virtualization active. As nearer
>>>>>> the
>>>>>> values are on the default, the problem appears seldom.
>>>>>>
>>>>>> It would be great, if  you can try it on your machine with these
>>>>>> values!
>>>>>>
>>>>>> Stephan
>>>>>>
>>>>>> "Microsoft News Groups" wrote:
>>>>>>
>>>>>>> I am trying this now.  I have 7 guests (all with additions 13.813)
>>>>>>> on this
>>>>>>> host for testing.  Before I saw this post, I had three of them with
>>>>>>> a
>>>>>>> relative weight of 99 and the other 4 with a relative weight of 100.
>>>>>>> All of
>>>>>>> them had the default maximum capacity of 100% and no reserve.  This
>>>>>>> is
>>>>>>> usually our default.  development or test servers get a relative
>>>>>>> weight of
>>>>>>> 99 or less and production servers all get 100.  Since I have changed
>>>>>>> all of
>>>>>>> the guests over to a relative weight of 100, 2 days ago, I have not
>>>>>>> seen the
>>>>>>> problem, but I have seen the machine go days without experiencing
>>>>>>> the
>>>>>>> problem.  I used to think it had something to do with our backup
>>>>>>> scripts
>>>>>>> that happen everynight, but the combination of the "freezing" not
>>>>>>> lining up
>>>>>>> with backup time period and this only happening with the new DL585
>>>>>>> G2
>>>>>>> systems with AMD-V makes me think the problem is somewhere else.  We
>>>>>>> have 3
>>>>>>> other DL585 G1 systems without AMD-V technology that are running
>>>>>>> 2005 R2 SP1
>>>>>>> and the same scripting options that have never had any problems.
>>>>>>>
>>>>>>> I will keep you up to date on any future freezing.  I am relectant
>>>>>>> to load
>>>>>>> the hosts with any more guests since the remediation of the
>>>>>>> "freezing"
>>>>>>> problem is to basically crash the guests.   If you want or need any
>>>>>>> information offline, please let me know.  I have 3 of the DL585 G2
>>>>>>> systems
>>>>>>> that I am eager to get stable so will try just about anything to get
>>>>>>> the
>>>>>>> service level to the same as the other 8 hosts I have.
>>>>>>>
>>>>>>> Thanks
>>>>>>>
>>>>>>> Rob McShinsky
>>>>>>>
>>>>>>>
>>>>>>> "Benjamin Armstrong [MSFT]" <benarm@online.microsoft.com> wrote in
>>>>>>> message
>>>>>>> news:O%23k8WNhMIHA.5988@TK2MSFTNGP02.phx.gbl
>>>>>>>> Rob -
>>>>>>>>
>>>>>>>> Can you try this and see if it affects your system?
>>>>>>>>
>>>>>>>> Cheers,
>>>>>>>> Benjamin Armstrong
>>>>>>>> ============================
>>>>>>>> Windows Virtualization Program Manager
>>>>>>>> Blog: http://blogs.msdn.com/Virtual_PC_Guy
>>>>>>>> Book: http://tinyurl.com/ysxcbm
>>>>>>>>
>>>>>>>> This posting is provided "AS IS" with no warranties, and confers no
>>>>>>>> rights. You assume all risk for your use.
>>>>>>>>
>>>>>>>> Stephan Kaufmann wrote:
>>>>>>>>> Hi,
>>>>>>>>>
>>>>>>>>> I also have a HP DL385 G2 with 2 CPUs and Virtual Server 2005 R2
>>>>>>>>> SP1 x64
>>>>>>>>> installed on Win Server 2003 Enterprise x64. I have the same
>>>>>>>>> problem and
>>>>>>>>> found a solution with active AMD-V!
>>>>>>>>>
>>>>>>>>> All VMs must have the default values in the "Resource Allocation"
>>>>>>>>> config.
>>>>>>>>> As soon as one or more VMs have other values and the first VM
>>>>>>>>> boots up,
>>>>>>>>> it freezes. If you set all VMs to the default values, all VMs work
>>>>>>>>> fine
>>>>>>>>> with AMD-V active in BIOS and also "Enable hardware-assisted
>>>>>>>>> virtualization if available" active in the "General Properties" of
>>>>>>>>> the
>>>>>>>>> VM.
>>>>>>>>>
>>>>>>>>> If you really want a VM with other "Resource Allocation" config,
>>>>>>>>> you have
>>>>>>>>> to disable "Enable hardware-assisted virtualization if available"
>>>>>>>>> of this
>>>>>>>>> single VM and it works too.
>>>>>>>>>
>>>>>>>>> Can anyone confirm this?
>>>>>>>>>
>>>>>>>>> Thanks.
>>>>>>>>>
>>>>>>>>> "Microsoft News Groups" wrote:
>>>>>>>>>
>>>>>>>>>> Ben;  From what I am seeing, since this happened again last
>>>>>>>>>> night, all
>>>>>>>>>> systems become unresponsive through what ever remote capabilities
>>>>>>>>>> are
>>>>>>>>>> out there.  Looking at the event log, all systems went into a
>>>>>>>>>> saved
>>>>>>>>>> state right before t his happened.  The website is inaccessable,
>>>>>>>>>> and it
>>>>>>>>>> cannot be reached with the VMRC client to go to the
>>>>>>>>>> administrative
>>>>>>>>>> display.    We have 8 other hosts running the same set of scripts
>>>>>>>>>> with a
>>>>>>>>>> mix of SP1 and non-SP1 VS 2005 R2.  All of those are working
>>>>>>>>>> perfectly.
>>>>>>>>>> The only difference is that these are DL585 G2 servers with the
>>>>>>>>>> AMD-V
>>>>>>>>>> technology.  We  have some DL585 G1 systems that are running very
>>>>>>>>>> well.with no problems.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> This is a snipit right before the Host went out to lunch.
>>>>>>>>>>
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:25 PM Virtual Machine "DH350" was saved.
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:25 PM Setting Change The setting "hardware/bios/time_bytes"
>>>>>>>>>> for the
>>>>>>>>>> virtual machine configuration "DH350" was changed.
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:25 PM Setting Change The setting
>>>>>>>>>> "settings/configuration/saved_state/path/absolute" for the
>>>>>>>>>> virtual
>>>>>>>>>> machine configuration "DH350" was changed from NULL to
>>>>>>>>>> "D:\B\ServersE\DH350\DH350.vsv".
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:25 PM Setting Change The setting
>>>>>>>>>> "settings/configuration/saved_state/path/relative" for the
>>>>>>>>>> virtual
>>>>>>>>>> machine configuration "DH350" was changed from NULL to
>>>>>>>>>> ".\DH350.vsv".
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:13 PM Setting Change The setting
>>>>>>>>>> "settings/shutdown/quit/was_running" for the virtual machine
>>>>>>>>>> configuration "DH344" was changed from true to false.
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:13 PM Virtual Machine "DH344" was saved.
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:13 PM Setting Change The setting "hardware/bios/time_bytes"
>>>>>>>>>> for the
>>>>>>>>>> virtual machine configuration "DH344" was changed.
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:13 PM Setting Change The setting
>>>>>>>>>> "settings/configuration/saved_state/path/absolute" for the
>>>>>>>>>> virtual
>>>>>>>>>> machine configuration "DH344" was changed from NULL to
>>>>>>>>>> "D:\A\ServersD\DH344\DH344.vsv".
>>>>>>>>>>  11/20/2007
>>>>>>>>>> 5:28:13 PM Setting Change The setting
>>>>>>>>>> "settings/configuration/saved_state/path/relative" for the
>>>>>>>>>> virtual
>>>>>>>>>> machine configuration "DH344" was changed from NULL to
>>>>>>>>>> ".\DH344.vsv".
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> Looks liike it  protected the guests well before it died.  A
>>>>>>>>>> funny thing
>>>>>>>>>> is however, after a reboot I can get to all consoles and all
>>>>>>>>>> looks fine.
>>>>>>>>>> I can see the systems in saved state.  When I try to restore one
>>>>>>>>>> of
>>>>>>>>>> those systems from saved state, the server hangs again and does
>>>>>>>>>> not
>>>>>>>>>> allow me any access to the web or vmrc.  The server allocates RAM
>>>>>>>>>> which
>>>>>>>>>> I can see in task manager, but the guest never truely comes
>>>>>>>>>> online. The
>>>>>>>>>> only way to get the system to come back it to reboot one more
>>>>>>>>>> time or
>>>>>>>>>> kill the vssrvc.exe process, restart the service and discard all
>>>>>>>>>> previous save states for those systems that were saved during the
>>>>>>>>>> hang
>>>>>>>>>> time.
>>>>>>>>>>
>>>>>>>>>> Anythoughts
>>>>>>>>>>
>>>>>>>>>> Rob McShinsky.
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>>
>>>>>>>>>> "Benjamin Armstrong [MSFT]" <benarm@online.microsoft.com> wrote
>>>>>>>>>> in
>>>>>>>>>> message news:%239qN2JWJIHA.5468@TK2MSFTNGP05.phx.gbl
>>>>>>>>>>> For number 1 – are there any entries in the event log that might
>>>>>>>>>>> indicate what is happening here?
>>>>>>>>>>>
>>>>>>>>>>> For number 3 – are they using the web site or the VMRC client?
>>>>>>>>>>>
>>>>>>>>>>> Cheers,
>>>>>>>>>>> Benjamin Armstrong
>>>>>>>>>>> ============================
>>>>>>>>>>> Windows Virtualization Program Manager
>>>>>>>>>>> Blog: http://blogs.msdn.com/Virtual_PC_Guy
>>>>>>>>>>> Book: http://tinyurl.com/ysxcbm
>>>>>>>>>>>
>>>>>>>>>>> This posting is provided "AS IS" with no warranties, and confers
>>>>>>>>>>> no
>>>>>>>>>>> rights. You assume all risk for your use.
>>>>>>>>>>>
>>>>>>>>>>> TheWall wrote:
>>>>>>>>>>>> Hi, 1.When i click on reset nothing happens. Doesnt reset, just
>>>>>>>>>>>> keep
>>>>>>>>>>>> on running.
>>>>>>>>>>>> 2. Connection to the server has been lost
>>>>>>>>>>>> 3. Client being the platform being logged onto when we need to
>>>>>>>>>>>> do
>>>>>>>>>>>> testing (Virtual Machine) Thanks
>>>>>>>>>>>>
>>>>>>>>>>>> "Benjamin Armstrong [MSFT]" wrote:
>>>>>>>>>>>>
>>>>>>>>>>>>> Some questions:
>>>>>>>>>>>>>
>>>>>>>>>>>>> 1. After reseting a virtual machine what do you have to do to
>>>>>>>>>>>>> get the
>>>>>>>>>>>>> website back?
>>>>>>>>>>>>> 2. What problems / errors do you see from the 32-bit client.
>>>>>>>>>>>>> 3.  What do you mean by ‘client’ in this context?
>>>>>>>>>>>>>
>>>>>>>>>>>>> Cheers,
>>>>>>>>>>>>> Benjamin Armstrong
>>>>>>>>>>>>> ============================
>>>>>>>>>>>>> Windows Virtualization Program Manager
>>>>>>>>>>>>> Blog: http://blogs.msdn.com/Virtual_PC_Guy
>>>>>>>>>>>>> Book: http://tinyurl.com/ysxcbm
>>>>>>>>>>>>>
>>>>>>>>>>>>> This posting is provided "AS IS" with no warranties, and
>>>>>>>>>>>>> confers no
>>>>>>>>>>>>> rights. You assume all risk for your use.
>>>>>>>>>>>>>
>>>>>>>>>>>>> TheWall wrote:
>>>>>>>>>>>>>> I am running a 64Bit virtual Server.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>> I have a few issues: 1. In the server admin Website it does
>>>>>>>>>>>>>> not
>>>>>>>>>>>>>> respond if i try to reset or save Machines
>>>>>>>>>>>>>> 2. Cant connect to server from 32 bit client
>>>>>>>>>>>>>> 3. Cant add excisting clients created in Virtual server
>>>>>>>>>>>>>> 32bit.
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>>>>>>>>
>>>>>>>
>>

Written by VirtuallyAware

December 7, 2007 at 1:01 PM

Follow

Get every new post delivered to your Inbox.

Join 276 other followers

%d bloggers like this: