vSphere API: get cannot correct Cluster performance statistics

Is it the right place to ask specific questions on vSphere for extraction of Performance API? If this isn't the case, please let me know where I should post instead.

Overview

I do something wrong, because at least the value of mem.consumed.AVERAGE is different from what shows the GUI vSphere, that bothers me.

In addition, the list of statistics available (from vimPort.queryAvailablePerfMetric ()) for a ClusterComputeResource resourcePool is smaller that what (I think) the documentation [1] says it should be.

What I am doing wrong? I hope that these two problems are related...

I use the SDK 5.5 against a 5.1 vCenter.

Details: mem.consumed Clusters

Here, I have access to a ClusterComputeResource resourcePool, and get the statistic of mem.consumed, assuming it's the same metric as the statistical 'consumed' GUI for a cluster. But the API values are lower than the values in the GUI. I don't understand why. Do you have?

Maybe related, our vCenter facility also has some stand-alone ESX hosts. They also have a resourcePool (under a "plain" ComputeResource). The HostSystem both the ComputeResource resourcePool have "consumed" statistic, but also here the ResourcePool "consumed" is that lower to of HostSystem metric consumption.

Metric of usagemhz of the bunch is excactly the same GUI, so at least a part of what I do seems to be just.

How can I get the values for "consumed" statistic a cluster that shows the GUI?

Details: Measures available for clusters

My own code gives the same results as com.vmware.performance.PrintCounters of SDK/vsphere-ws/java/JAXWS/samples/com/vmware/performance/PrintCounters.java in the file SDK.zip:

$ JAVA_HOME = / usr/lib/jvm/java-6-Sun/jre. run.sh com.vmware.performance.PrintCounters - url https://Gaius/SDK -basic-connection - username "xxxx" - password "yyy" - entitytype ResourcePool - entityname resources - filename /tmp/resource.xml

$ grep called InfoNom-key /tmp/resource.xml

> key called InfoNom < usagemhz < / called InfoNom-key >

> key called InfoNom < vmmemctl < / called InfoNom-key >

> key called InfoNom < consumed < / called InfoNom-key >

> key called InfoNom < generals < / called InfoNom-key >

> key called InfoNom < cpuentitlement < / called InfoNom-key >

> key called InfoNom < mementitlement < / called InfoNom-key >

But where is mem.totalmb? or cpu.totalmhz? Looking at the GUI to Performance vSphere / advanced / table Options... / CPU / last day, totalmhz statistics is here, and I can see values for it in the GUI. But why not on the API?

When you look at the documentation for vSphere for memory counter PerformanceManager (5.1 [1], 5.5 [2]), I do not understand what it means that the entity column is empty. This means that it should or should not be available for a ResourcePool?

Links

  1. 5.1 memory counters
  2. 5.5 meter memory

I discovered what was going on. I asked resourcePool of ClusterComputeResource and ClusterComputeResource not directly. Once I asked the ClusterComputeResource directly, I had access to all statistics, as I expected and it had the values that I expected.

Tags: VMware

Similar Questions

  • How to generate more than one hour performance statistics in vSphere Client 4.0?

    We have installed a trial version of vSphere ESX 4.0 (on processor Intel 8 with Fedora Core 12) and vSphere Client (on Windows Vista 32-bit) and that it provides only 1 hour of performance statistics.  Then apply a license valid, it still generates only 1 hour from CPU graphics and logs.  Does anyone else have this problem?

    Performance logs are exported from vSphere Client by either: a)-> report file-> performance (recorded as file .xls for all virtual machines from the host), or b) right-click on the individual VM-> report Performance... (for the individual VM).  Whatever it is, will appear .xls files contains only statistics of CPU, memory, storage, network and disk of the last hour.

    In addition, the "Graphics Options" in the performance tab will not let you change the interval of time either, although there are options for the 'end '.  They are disabled. Please see screenshot.

    Help, please.  Thank you!

    The configuration is done on the host itself, not on the virtual machine
    Maish

    Architect of virtualization & Systems Administrator

    http://technodrone.blogspot.com

  • Get performance statistics of a VM method for the last 2 hours

    Need a scipt for toolkit VI to obtain performance statistics of a VM method for the last 2 hours

    What metrics (cpu.usage.average, disk.usage.average...) are you looking for?

    Have you tried one of the preconfigured metric collections (Cpu, memory, disk, network) which are available in the cmdlet Get-Stat?

    An example with the scope of the Cpu

    get-vm  | Get-Stat -Cpu -Start (Get-Date).addhours(-2) -IntervalMins 2 -MaxSamples 60
    
  • Question about performance statistics

    William-

    I was looking on the 'doc' House to see if there was reference to the period of time represented in the performance data given in the output of the script.

    Can you tell me what we actually with the host / cluster (CPU, memory) performance statistics?

    Also - I was wondering why "ESX/ESXi Hardware Configuration" includes CPU utilization and memory - that's just how the API works?  It just seems odd to include these data with summary material...

    Forgive the noob questions (and anything that might follow) I (FINALLY) movement in a project that will allow me to work much more consistent with the API, and this script makes so much to make my life easier.  We put the spotlight on performance and the ability of prediction, so I'll try to learn what can offer the API (overlooking DOCS 9840 & 10665 as time lost...)

    Thank you!
    Don

    These are just rough summary level a time interval counters, if you are looking for granular performance information, you will need to look in performanceManager which is the entitity wherein you will get more details.

  • Options for monitoring or recovery of C-Series system performance statistics, series EX codecs

    Hello - I've been looking around to see if there is an option available today to recover the performance statistics of the system (for example, load CPU, memory, etc...) of TAA endpoint via the command line interface or dynamic XML Poll (http 'getxml'), of an endpoint.  I understand that Cisco has today a company MIB available for these endpoints and the API is needed to pull information from the codec, however, I want to be able to pull the CPU load and memory of these endpoints dynamically load.  Is it possible to do today?

    Thank you in advance,

    -Andrew

    Hello, Andrew!

    In general, it's something I wouldn't really worry about and be more skeptical if the

    follow-up itself generate an additional burden troubled =.

    As the systems are based on linux, you have command line tools and proc interface

    You can use to get stats on the endpoint. I'd rather watch if there are problems

    rather than trying to do something dynamic, I think that it is better accomplished by the endpoints.

    Example, login as root and try w or free

    [Martin-Koch-EX90 - ATEA-Vaas: ~] $ w

    00:17:02 14 days, 23 min, 1 user, load average: 0.00, 0.02 and 0.05

    ATS to [email protected]/ * / IDLE JCPU PCPU WHAT

    root pts/0 00:03 0.00 0.31 s s s 0.02 w

    [Martin-Koch-EX90 - ATEA-Vaas: ~] $ free

    total used free shared buffers cached

    MEM: 510912 309832 201080 0 25768 156616

    -/ + buffers/cache: 127448 383464

    Swap:            0          0          0

    proc interface is also interesting:

    [Martin-Koch-EX90 - ATEA-Vaas: ~] $ cat/proc/uptime

    1210174.87 1193259.96

    [Martin-Koch-EX90 - ATEA-Vaas: ~] $ cat/proc/loadavg

    0.00 0.05 0.01 25905 1/131

    [Martin-Koch-EX90 - ATEA-Vaas: ~] $ cat/proc/meminfo

    MemTotal: 510912 kB

    MemFree: 210312 KB

    Stamps: 25580 kB

    Cache: 156016 kB

    SwapCached: 0 kB

    Active principles: 70892 kB

    Inactive: 115896 kB

    ...

    You should be able to add a key ssh and auto login and get the information you need and

    process via a script and then use this data in your monitoring tool.

    and of course you have the xstatus which you can also access by tsh or the xmlapi http, you can find more information

    in the guide of the api. As you wrote, there isn't that much out via snmp.

    The intention itself is right though, so maybe you could file a feature request for the traps

    and better pro active monitoring, great thing miss me the most on C/TC is the lack of traffic/network/loss reports.

    Please response rate!

  • Get-journal authorized to perform this operation by using read-only user role, was refused.

    Hi people,

    What is the minimum role to get the newspaper my whole army ESX?

    I use the following script oneliner:

    Get-VMHost | Get-Log messages | Select -ExpandProperty Entries | Where {$_ -like "*error*"}
    

    but it always ends in failure result:

    Get-Log : 05/07/2011 9:15:11 AM    Get-Log        Permission to perform this operation was denied.    
    At C:\Temp\a5440b5a-6f5a-4fe2-9027-e432ba87c869.ps1:3 char:21
    + Get-VMHost | Get-Log <<<<  messages | Select -ExpandProperty Entries | Where {$_ -like "*error*"}
        + CategoryInfo          : NotSpecified: (:) [Get-Log], NoPermission
        + FullyQualifiedErrorId : Client20_RuntimeDataServiceImpl_GetAvailableLogList_Error,VMware.VimAutomation.ViCore.Cmdlets.Commands.GetLog
    
    
    

    Any suggestion please?

    Permission can be derived from the API method used by the cmdlet Get-Log under the covers.

    In the description of the BrowseDiagnosticLog method, you will see that it takes the Global.Diagnostics .

    The default read-only role does not have this privilege!

    You will need to create a new role.

    The other point to watch is the location in the structure of vSphere where give you permission.

    I feel that this specific one should be given to the root of the vCenter to allow access to newspapers.

  • Linked Clone with vSphere api and VIClient 2.5

    I'm trying to create Clone bound using vSphere Web services api (c#) and it always end like full clone.

    My question is, is it possible to create Linked clone using vSphere api with VIClient 2.5.

    I know that vShpere supports version 2.5 VIClient. vShpere has the ability to create Linked clone.

    I try to know why it does not create Linked clone from my environment.

    I am ready to explain if you need more details.

    Thank you.

    Yes, you can vMotion Linked Clones... sVmotion may work, but I have not tested, probably not a good idea to move them.

    Don't forget to assign points if you found useful answers

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

    William Lam

    VMware vExpert 2009,2010

    VMware scripts and resources at: http://www.virtuallyghetto.com/

    Twitter: @lamw

    repository scripts vGhetto

    Introduction to the vMA (tips/tricks)

    Getting started with vSphere SDK for Perl

    VMware Code Central - Scripts/code samples for developers and administrators

    VMware developer community

    If you find this information useful, please give points to "correct" or "useful".

  • Cannot correctly display the default Microsoft Web sites text only

    Since "redevelopment" of Windows 10, Firefox cannot correctly display web pages of MSDN, Technet, or most other Microsoft Web sites. All I get is text based websites, no visuals at all.
    There was once a way to have Firefox Web sites meet and pass off as internet explorer or Chrome... has this capacity disappeared or is it the fault of the web programmers at Microsoft screwing around with their code?

    Found the problem...

    EFF privacy Badger blocked MSDN. I reported to EFF the problem, please help everyone in this.

  • VSphere Web Client cannot connect to the server vCenter Single Sign On.

    I'm running the virtual appliance of the trial 5.5.0.20400 build 2442330 on ESXi 5.5.0, 2068190

    While I try to log on to the Web Client, I get this error.  VSphere Web Client cannot connect to the server vCenter Single Sign On.

    I put fallow the steps to disable SSO by changing the webclient.properties line add file and ad sso.enabled = false .    Then on the vCenter Server Appliance, restart the vSphere client service by typing service vsphere-client restart .

    I enclose the reference files.

    All ideas will be useful


    This answer was simple, all I had to do was remove the # in front of the statement in the file.   and SSO has been disabled after the restart of the service.

  • VSphere APIs and compatibility ESX

    I've implemented the backup of a VMWare VM and application of recovery with vSphere API and VDDK. The development has been done entirely on 5.0 ESX servers. The following discussion provides an overview of how I implemented the restoration.

    http://communities.VMware.com/message/2130483#2130483

    When another person has tested my application, he noted the rollback feature does not work. When I've debugged, I found parameter spec unit to CreateVM_Task behaves differently 5.0 Esx to Esx 4.0. In the case of Esx 5.0, it works perfectly, where as in the case of Esx 4.0, none of the devices that I fill before the CreateVM_Task is created and the restore fails because of this.

    How can I understand the compatibility of the Esx and the vSphereAPI versions? I assumed that the SDK should be retro-compatible with Esx Server? Can someone shed some light how this problem can be solved?

    Thank you

    . / Siva.

    Hi Siva,

    That sounds good, maybe, we'll get you going even faster than I expected:

    As already noted William

    lamw wrote:
    but in general, the more recent editions of vSphere will support previous editions of vSphere. vSphere 5.1 supports vSphere 4.x and 5.0

    the only thing being that if new properties are added to a data in 5.1 object then you can not use in 4.x or 5.0.

    It seems to me that your code is based on 5.0 WSDL generated by gsoap stubs. It is ultimately you generate SOAP requests that are supposed to work with 5.0, 5.1, 6.x one as well now. Yet, when he delivers the SOAP messages to 4.x, it is likely that you include a property that is not defined in the WSDL of 4.x.

    Attached is a diff of the definition WSDL VirtualMachineConfigSpec of 4.0 to 5.0 Quick. Look at line 27 and following

    
    27              
    ====
          !>      
    ====
    28              
    

    Here, you can see that numCoresPerSocket property was not available in 4.0, but was added in 5.0. It is also what we said the documentation centre: for vSphere API 5.0.

    You declare that numCPUs is affected but memoryMB not. Between the two numCoresPerSocket has been inserted to 5.0. It seems that the 4.0 Server reads your VirtualMachineConfigSpec until numCPUs then encounter numCoresPerSocket and abandoned.

    So far my theory.

    What I don't understand is that this seems to happen in silence as you don't declare any exception. Yet, says William

    wlam wrote:
    As you know, there will be new properties in new versions of vSphere such as 5.1, that if you try to assign to an older version of vSphere, it will throw an unknown error...

    As you can learn from the attached diff we serveral more properties that have been added to the 5.0. Might well want to comment on all those in your current code and re-run CreateVM_Task against server 4.0.

    Perhaps this might solve the problem.

  • Question so that the migration of the VM using the vSphere API

    Hello

    I am having some problems during the migration of a virtual machine using vSphere SDK.

    I have install the vSphere environment as there are two hosts esxi installed on different physical machines with multiple virtual machines above them.

    Now I want to migrate one say the virtual macines VM1 fron host1 to host2. I did it via vCenter uses the vSphere client and it worked fine. Now, I have to perform the same task using the vSphere API. For this I use the template of "VMotion" located in the com/vmware/vm directory. I run the command as follows:

    targethost $ run.bat url username password, shows targetpool

    I'm not sure of the targetpool argument and I need advice on what I need to specify a name of pool of resources here. Acorrding to:

    http://www.VMware.com/support/developer/VC-SDK/visdk400pubs/ReferenceGuide/Vim.virtualmachine.html#migrate

    If the pool parameter remains disabled, current pool of the virtual machine is used as target the pool. I tried specifying the pool of resources such as:

    [- targetpool]

    but it gives the error:

    Error: List of resources target not found.

    Looking forward to all the advice. Thank you.

    Even if you do not have resource pools, it is actually a 'root' list of resources"which will be still exists. Try to specify 'Resources' as the name of the targetpool and allow you to migrate your VM.

  • ESX performance statistics

    I need help...

    After watching an interesting info from VMware on PowerCli session, I realized that I could get IOPS / s (and other performance statistics) via PowerCli ESX servers.

    I had a script that goes to each VMhost (ESX) and request the disk.commands.summation for each VM and the data store.

    Then I could summarize these take medium and gives me a report on how would IO.

    I had to go directly to the ESX rather than through vCenter, as the standard vCenter level stats had disk.commands.summation in real time; and the ESX retains at least a day.

    Later, I noticed that I could upwards the level Stats and get the OPS are / s of vCenter for a week...

    But trying to get the statistics of vCenter (via PowerCli) I get nothing.

    I did a Get-VMhost | Get-StatType... and I see the stat... but when I do Get-VMhost | Type get-Stat - disk.commands.summation returns nothing...

    Can someone please... I would like to know the OPS are / s for each VM and the data store for the past week, so I can properly balance my data stores.

    OK, let's try since last week.

    Get-VMHost  | Get-Stat -Stat disk.commands.summation -Start (Get-Date).AddDays(-7)
    

    What is this returns anything?

    One last thing, you are connected (connect-VIServer) to the vCenter and not the host ESX (i)?

    ____________

    Blog: LucD notes

    Twitter: lucd22

  • Detcts iPod but ITunes "cannot correctly identify" uninstalled iTunes and then re-installed. Restarted laptop. Folder deleted Apple Lockdown, rebooted, still could not "identify correctly.

    ITunes detects the iPod, but 'cannot correctly identify' uninstalled iTunes and then re-installed. Restarted laptop. Folder deleted Apple Lockdown, rebooted, still could not "identify" during each step, also disconnected iPod USB port.

    Try the following (in that order).  Looks like you're using Windows (not a Mac), so I'll make this assumption...

    Shut down (power off) the computer, disconnect all the USB devices.  Plug the laptop power adapter, so that it does not work on battery power.  Do this Reset (reboot), the iPod

    Learn how to reset your iPod - Apple Support

    Start computer, run iTunes and connect the iPod to a USB direct on the laptop by itself.  If there is no difference, put the iPod in Disk Mode

    Put the iPod in disk - Apple Support Mode

    Then connect to the laptop again, always by himself.  If iTunes picks it up, he may ask you to do a restore with a message on the "recovery."  If you're OK with deletion of the iPod, do the restore.  See if the iPod is "correctly identified" after restoration.

    If the problem persists, you said you uninstalled and reinstalled iTunes, but may not have uninstalled iTunes and associated software components.  Follow the steps in this document to uninstall an iTunes installation components, in the order described all listed

    Remove and reinstall iTunes and software for Windows 7 and later versions - Apple Support components

    Then, reinstall the latest version of iTunes

    http://www.Apple.com/iTunes/

  • How can I get the correct product key for my sleekbook want PC4

    How can I get the correct product key for my sleekbook desire 4 PC

    Hello

    As mentioned in your other post you can get it from Microsoft with your money:

    http://h30434.www3.HP.com/T5/notebook-operating-systems-e-g-Windows-8-and-software/i-purchase-envy-notebook-PC-4-1100se-how-can-i-get-the/m-p/2853225

    Kind regards.

  • How to get the sub cluster item caption?

    Hi all

    I have a big cluster, which contains several groups of void. They are all type def.  My queation is,

    How to get the sub cluster item caption?

    Thank you.

    Anne

    In this nugget , you will find many images that illustrate how to get to items within a cluster. The ideas are the same, you will search for the legend.

    Have fun!

    Ben

Maybe you are looking for