Run VMmark 2.0 Failed (VM Resources Check)

[sfchang0731]

Hi All,

Have any one run the VMmark2.0 completed?

I have preparing the infrastructure for VMmark2.0 test already.

When I start the benchmark from client0 (STAXMon.jar)

1. Reading configure file : ok

2. Configuring workload: ok

3. Starting  environment setup: ok

4. Time sync: ok

5. Get Error "Failed: VM Resources Check"

I tracking the issue: The configure file (xml file) call the "VMmarkPScli.ps1" to run. And faild in the function "CreateClusterVMMapping".

I always can not solve the issue, Help~~~~ (Detail issue log file: ClusterVMmapping.txt)

Thanks

[Joshua Schnee]

Hi,

Based on the attachedments you've provided it appears your environment is likely missing one or more required components. 

What version of Powershell is installed?  Make sure version 2.0 is installed.

What version of VMware PowerCLI is installed?  Make sure the latest downloadable version is intalled (currently build 332441)

Please review the benchmarking guide and double check that you haven't missed a step.

Thanks,

Joshua

[sfchang0731]

Hi jpschnee,

Thanks for your help~~

I checked my all software version again and all the same with VMmark2.0 doc.

I re-install powershell and powercli but still failed.

Have any idea?

Sam

[Joshua Schnee]

Hmm.

Can you send me the following output?

From a vSphere PowerCLI command prompt:

>$host

as well as :

>Get-VIToolkitVersion

What version of Windows are you running from?  32 or 64bit?

Thanks,

Joshua

[sfchang0731]

Hi,

The information for your refer as below list,

From a vSphere PowerCLI command prompt:

>$host

Name             : ConsoleHost
Version          : 2.0
InstanceId       : 28837841-81a1-4416-993b-5627341924f7
UI               : System.Management.Automation.Internal.Host.InternalHostUserI
nterface
CurrentCulture   : en-US
CurrentUICulture : en-US
PrivateData      : Microsoft.PowerShell.ConsoleHost+ConsoleColorProxy
IsRunspacePushed : False
Runspace         : System.Management.Automation.Runspaces.LocalRunspace

as well as :

>Get-VIToolkitVersion

PowerCLI Version
----------------
VMware vSphere PowerCLI 4.1 U1 build 332441
---------------
Snapin Versions
---------------
VMWare vSphere PowerCLI 4.1 U1 build 332441

What version of Windows are you running from?  32 or 64bit?

Windows Server 2003 R2 Ent SP2 32bit

Thanks

Sam

[Joshua Schnee]

Hi Sam,

That all seems right.  When you reinstalled, did you uninstall the previous version or just install on top of what was already there?  If the later, I would try uninstalling, rebooting, and the reinstalling.

What version/build of ESX or ESXi is on your 192.168.0.200 host?

Also, from a PowerCLI command prompt, can you do the following and attach the output?

>Connect-VIServer <your VIServerName>

>Get-VM -Location 192.168.0.200 | Out-File VMlist.txt

>Get-VMHost 192.168.0.200 | Out-File HostInfo.txt

>Get-VMHost 192.168.0.200 | Get-Member | Out-File -Append HostInfo.txt

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

Sorry to response too later!

I try the uninstall the Plwershell and PowerCLI than restart and install again, but still failed.

The attachment are output file from PowerCLI.

And I build the ESX4.1 in the 192.168.0.200.

Thanks for your help!

Sam

[Joshua Schnee]

Hi Sam,

As far as I can tell, this also looks fine.  I'm really not sure what's going on in your environment.

Can you run the following command and attach the results?

>.\VMmarkPScli.ps1 -f createreport -vc <VCName> -c <ClusterName> -u <VC User Name> -p <VC User Password>

It might also be worthwhile for you to log into your vCServer, switch to the "Hosts and Clusters" view (Ctrl+Shift+H), and then take a screen shot of your environment as vCServer sees it.

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

I have attach the report and screen-shot file already.

And I just set the one tile to run the test at first time. it's ok?

Thanks

Sam

[Joshua Schnee]

Yes it's fine for it to be set to 1 tile.  I'll keep digging around but do you have another host to add to your cluster?  I'm wondering if that might be what's causing the issues here.  VMmark 2 has a minimum requirement of 2 hosts.

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

"VMmark 2 has a minimum requirement of 2 hosts"?

I can not find the this requitment for VMmark 2 in the Guide.

What's requirement and function setup in the second hosts?

Thanks

Sam

[Joshua Schnee]

Hi Sam,

I'm not sure it specifically calls out the two host requirement (I'll get with our tech writer on that), but if you look at the hardware requirements section on page 35, you'll see that "The VMmark 2.0 benchmark requires all ESX hosts used in each test to be vMotion compatible and to use shared storage".  In, other words, to do all of the various work required by the benchmark, you need two or more hosts that can vMotion any/all VMs to each other. 

BTW I'm not a 100% this is your issue, but if you have another host to try, it would be great to rule it out.

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

First, Thanks for your help. I add the second host in the cluster and passed the "VM Resources Check Failed" issue.

Today, I run the test with all workload for VMmark, but faild in the vmotion_function.xml (STAXPythonEvaluationError signal raised. Terminating job)

So, I just run the only standby$Mailserver and Delopy. It's OK.

But run workloads - vmotion and SVmotion, terminating job again!

Do you know this issue that what reason to made~

The attachment are screen shot for error, my test environment maps and vmmark2.conf file.

1. 1 tile, 1 client0

2. Share storage: 2 LUNs "nfs1" "nfs2" (all VMs put in the nfs1)

3. 2 hosts (ESX 4.1)

Thanks

Sam

[Joshua Schnee]

Hi Sam,

Glad adding the second host to the cluster fixed your previous issue.  The issue your experiencing now with sVMotion is likely due to your naming of the Standby0 VM.  It's case sensitive, so "Standby0 != standby0".  Try changing the name to "Standby0" in all places in your environment, especially within vCServer.

In the files you uploaded, I didn't see any reference to vMotion errors.  If you are having issues with that workload please upload your results folder and screenshots (thanks!) when you can.

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

Yes, you are right.

It's case sensitive to the STAXPython run failed and terminate job.

Now I have run the all workload complete and ending normally.

But script "tilescore.pl" generate a score failed. (File name: Score_1_Tile_Test.out)

=============================================

cygwin warning:
MS-DOS style path detected: C:\vmmark2\tilescore.pl
Preferred POSIX equivalent is: /cygdrive/c/vmmark2/tilescore.pl
CYGWIN environment variable option "nodosfilewarning" turns off this warning.
Consult the user's guide for more details about POSIX paths:
http://cygwin.com/cygwin-ug-net/using.html#using-pathnames
Computing_results_for_test_with_tile_count: 1 ...
Referencing_tile: 0.
Argument " " isn't numeric in addition (+) at C:\vmmark2\tilescore.pl line 936, <$FileInput> line 2.
Undefined times at 180
Run_start 1295542680  : Fri Jan 21 00:58:00 2011
Start_time 0 : Thu Jan  1 08:00:00 1970
End_time 1295553420 : Fri Jan 21 03:57:00 2011
Run_end 1295553420 : Fri Jan 21 03:57:00 2011
Duration_in_minutes : 21592557.00
Steady_start_start 1800 : Thu Jan  1 08:30:00 1970
Steady_state_end 9000 : Thu Jan  1 10:30:00 1970
Phase_0_begin 1800 : Thu Jan  1 08:30:00 1970
Phase_1_begin 4200 : Thu Jan  1 09:10:00 1970
Phase_2_begin 6600 : Thu Jan  1 09:50:00 1970
Error: could not resolve start-time. Results data missing from one or more *.wrf files or time on clients not synchronized

========================================================================================

Can you help me?

I have attachment the zip file for all result folder and screen shot for STAF application.

Thanks

Sam

[Joshua Schnee]

Hi Sam,

I've looked over the results you attached.  You seem to have numerous setup issues with all of the workloads except for vMotion and Standby.  Likely your issues are due to hosts file mistakes and template setup.  I'd recommend that you go back through the benchmarking guide and make sure that you've followed all of the instructions for configuration, especially those instructions for setting up an environment from templates.  Although they are prepackaged for you, you'll still need to modify the VMs for your environment.  Then test each workload in isolation, possibly reducing the runtime to aid with debugging.  Once you get a bit further along, feel free to ping me again.

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

Thanks your suggestion!

I have reset the test environment (All VMs and Client0) that follow the guide.

But still failed in the get the score for tile.

I found the some error message in the event viewer. Do you see the message before?

The message not only happened in my client0, other clients the same.

Thanks

Sam

[Joshua Schnee]

Hi Sam,

Based on what I found on the internet, these appear to be generic .Net issues.  Did you remember to follow step 9, on page 126 of the benchmarking guide and install Windows Updates?  I would start there.  If that doesn't resolve your issue you may need to look at several of the kb articles on "support.microsoft.com", for work arounds.

Hope this helps,

Joshua

[sfchang0731]

Hi Joshua,

I have solved the "crypt32" error.

I installed the all update package for windows 2k3, but ".Net Runtime 2.0 Error" still happened during test running!

I will check issue support in the microsoft web site.

Thanks

Sam

[Joshua Schnee]

Hi Sam,

I thought about this a bit further, and you might also be seeing this as a result of a configuration issue.  Can you provide your latest results directory?  If you're seeing the issue on all tiles and all instances of DVDstore (WebA,B,and C), can you also provide the hosts file for your prime client, DS2DB0 VM and DS2WebA0 VM?

Thanks,

Joshua

[sfchang0731]

Hi Joshua,

I attachment the results and hosts file for your reference.

I just run the one tile in my test environment, and results are run the DS2 workload only.

The ".Net runtime 2.0 error" for ds2webdriver.exe sitll exist on the event viewer.

Thanks

Sam

[Joshua Schnee]

Hi Sam,

In looking at the hosts files for your DS2WebA, B and C VMs, I see that you didn't follow the guidelines for configuring these VMs.  In the template provided you should have seen the following :

# begin snip

127.0.0.2 DS2-WEB.eng.vmware.com DS2-Web

# Change to the IP of DS2DBn, do not change the HOSTNAME

<some ip>   DS2-DB

# end snip

In your hosts files, you changed both the ip address and the hostname.  For all of the web server stuff to work, each web front end must be able to resolve DS2-DB, which yours cannot.  Please modify your DS2Web VM hosts files so that they can resolve DS2-DB and try again.

Ex:  192.168.0.106  DS2-DB

Hope this helps,

Joshua