Very bad performance on a virtual machine

Similar Questions

• Slow startup and performance of the virtual machine

  Hello

  We used with ESX 3.5 U3 and VC 2.5 Build 119801 for quite awhile. We organize two virtual computers W2k3 running Citrix XenApp 4.5 (last patchlevel). One of these two machines have a meaning very, very bad performance 100% cpu and memory. The other virtual machine works (ago implemented by cloning each other) fine.

  We migrated virtual machine causing the problem to another host ahead has a bottleneck of the esx host. But if there is only the one virtual machine on an esx host, the problem remains. When starting the virtual machine, the W2k3 bootscreens hung for at least eight minutes. We followed the vmware.log in the datadir of the virtual machine. The vmware.log says:

  10:05:27.209 Mar 27: vcpu-0 | SVGA: Registration MemSpace 0xf8000000 (0xf8000000) and 0xf4000000 (0xf4000000)

  We have no memory of overcommitting, no action, no DRS, without limits and without reserves. I just deleted a snapshot as I thought writing the delta file could slow the operation.

  All other virtual machines seemed fine.

  Any of you have seen this problem before?

  Thanks in advance,

  Daniel Volkerts

  OK, then once the machine starts, run the real-time performance monitor, add counters for cpu, mem, disc and see if it goes pic... also Task Manager control which process takes more cpu or mem...

• When you look at the performance of a virtual machine, when you make a graphic of 3 months it appears correctly for the first two, but the last month is compressed.

  When you look at the performance of a virtual machine, when you make a graphic of 3 months it appears correctly for the first two, but the last month is compressed.

  Hello

  When you say that the last two months are compressed. I do not have what exactly you mean by that. Is there a chance you can attach a screenshot of the same thing? Also let me know if there are any made VC-database activity recently?

• Hosts with active Hyperthreading - best performance of the virtual machine?

  After having been involved in a recent thread on Hyperthreading activation on your hosts: -.

  http://communities.VMware.com/message/1567821

  In this thread Jason comments that when the HT is enabled he saw much better performance of virtual machines on this host, we have, I think I've also seen elsewhere.  We saw a clear improvement of the performance of our Oracle server when its on a host that has enabled HT, so much so that we have now disabled for this virtual machine DRS, so that it doesn't move off them.

  If anyone has any other given that a marked improvement?

  The specification of hosts within our farm are essentially the same, and new hosts with HT could actually could be called lower spec than our other guests so its interesting, it makes such a difference.

  In any case, interesting to see what other people have found.

  Yes, class processors Nehalem hyperthreading performance of vSphere one little, doubling the number of execution threads will make Scheduler ESX easiers so many jobs. I have a little stand alone ESX hosts with a single Quad Core Nehalem processor for applications Oracle and performance is very good.

• Performance of the virtual machine on ESXI 5

  Hello

  I have a DELL R710 server and I would like to install about 18 VM servers. almost all of them are Centos Linux 64-bit and 2 Win XP and Win7.

  I have 48 GB of RAM and 2 processors with a total of 12 cores.

  I would like to know if this can not work?

  should I worry about more power (cores) CPU or RAM?

  I'm pretty new to VMWare ESXi and I not change any configuration on the ESXI system.

  that I should?

  I guess I work in a mode of sharing...

  So basically, I need to know if I would have the same ratio of VM CPU and memory, as I do with physical machines CPU and Memoery?

  Thank you very much!!

  Dothan.

  adotan wrote:

  I use about 4 GB of RAM for each virtual computer.

  With 18 virtual machines, it will be 72 GB of RAM, and if you have only 48 physical RAM in the host it will be little tight. There is a lot of memory reappropriation of the techniques used by the Vmkernel, but this might be a little too much.

  I agree with Weinstein that you should start with a lower amount of vRAM for each virtual computer and increase if necessary.

• get data on the performance of the virtual machine

  Hello community,

  I am trying to obtain performance data from a virtual machine using the following code:

  //create start & stop time 
  var end = new Date(); // now
  var start = new Date();
  start.setTime(end.getTime() - 3600000); // 1h before end
  //create a querySpec for one entity
  var querySpec = new Array();
  querySpec.push(new VcPerfQuerySpec());
  querySpec[0].entity = VM.reference; //set entity of workflow VM
  querySpec[0].startTime = start;
  querySpec[0].endTime = end;
  //creteate PerfMetricID for one metric
  var PM = new VcPerfMetricId();
  PM.counterId = 2;
  PM.instance = "";
  var arrPM = new Array();
  arrPM.push(PM);
  querySpec[0].metricId = arrPM; //assign PerfMetric to querySpec
  querySpec[0].intervalId = 20;
  querySpec[0].format = "csv";
  
  var CSV = VM.sdkConnection.perfManager.queryPerf(querySpec);  // query PerformanceManager
  System.log (CSV);// show if type is OK
  //show properties
  System.log (CSV.entity);
  System.log (CSV.value);
  System.log (CSV.sampleInfoCSV);
  System.log (CSV.dynamicProperty);
  

  The workflow is valid and functional. But there is no data. Here's the log produced by the workflow:

  [2010-11-12 12:12:53.742] [I] DynamicWrapper (Instance) : [VcPerfEntityMetricCSV]-[http://class com.vmware.vim.vi4.PerfEntityMetricCSV|http://class com.vmware.vim.vi4.PerfEntityMetricCSV] -- VALUE : com.vmware.vim.vi4.PerfEntityMetricCSV@beb35d25
  [2010-11-12 12:12:53.742] [I] undefined
  [2010-11-12 12:12:53.742] [I] undefined
  [2010-11-12 12:12:53.742] [I] undefined
  [2010-11-12 12:12:53.742] [I] undefined
  

  I also tried .format querySpec [0] = 'normal '; -same result.

  Where is my fault? I'm on the right track?

  Pls support me in obtaining data on performance, Thx.

  -

  ---

  Kind regards, Andreas Diemer

  visit http://www.vcoteam.info & http://mighty-virtualization.blogspot.com/

  SERVUS Andreas!

  I think I got to the next step:

  perfManager.queryPerf (...) returns a table, not the CSV values itself. Therefore, pop() data in the result table.

  A change in your code that returns a large amount of data in my lab:

  var CSVArr = VM.sdkConnection.perfManager.queryPerf (querySpec);  query PerformanceManager

  System.log (CSVArr) ;// show if the type is OK

  view properties

  var CSV = CSVArr.pop ();

  System.log (CSV.entity);

  System.log (CSV.value);

  System.log (CSV.sampleInfoCSV);

  System.log (CSV.dynamicProperty);

  But I do not know why the System.log (CSVArr) does not show the array as a type... (maybe because there is only one item!)

  Edit: found the answer:

  ... - VALUE: com.vmware.vim.vi4.PerfEntityMetricCSV@229650a8 versus

  ..... VALUE: com.vmware.vim.vi4.PerfMetricSeriesCSV@5d7f26

  "Once again learned!"

  BTW: I found the idea with the table on the slide 59 of this presentation:

  http://communities.VMware.com/servlet/JiveServlet/download/1371233-29453/vSphereAPI_PerfMonitoring.PDF

  Hope this helps

  See you soon,.

  Joerg

  PS: How can you add not formatted 'verbatim' sourcecode-style here in the forums?

• Poor performance of the virtual machine

  I used VMware Server for a few years and thought I would give an ESXi, but I have problems.

  Here's the story backwards.

  One of my friends has a server which is used as its SMB company's intranet.

  This server is on its last breath and ready to die.

  He has no access to code custom or even how to reinstall this application to a new server.

  So I thought that ESXi would be a good fit.

  I had an older server (2.7 ghz Xeon dual processor HyperThreading 3 GB of RAM) laying around.

  I know that is not the latest and the greatest, but it uses a double 500 mhz with 1 GB of RAM now.

  I set up ESXi and then used VMware Converter to virtualize the system.

  Everything went well and the conversion worked as expected.

  The application worked.

  Now a serveral people hit the virtual machine, it works like a dog.

  The old dual 500 mhz Server works best.  Although on the surface, it is 5 x more speed and 3 x more RAM.

  Ping on the virtual computer will range from 1ms to 1000 + ms not very regularly.

  However, the ip address of managing pings under 1ms and is on the same physical adapter.

  Any suggestions for what to look for?

  Am I missing something?

  I thought my configuration to the top was pretty basic.

  How many virtual processors did give the virtual machine? If I would try a simple vCPU - double the amount of memory is affected and how much the server have?

  If you find this or any other answer useful please consider awarding points marking the answer correct or useful

• Accident of ESX 3.5 but virtual Center did not perform vMotion on virtual machines

  Hello

  We have 3 ESX servers on Virtual Center 2.5. One of the ESX Server suddenly restarted due to hardware failure. There were 5 virtual machines on this particular ESX host.

  When we check all the VMs who resided on the ESX host touched, he has just restarted and didn't vMotion for the rest ESX servers

  The question is why don't the vMotion VMs to the rest of the other servers. The rest of the ESX host have more than enough CPU and RAM to accommodate all 5VMs

  My understanding of vMotion would be when an ESX host is down, all the virtual machines will be migrated to other ESX hosts on the cluster.

  When an ESX host is down, the migration of all virtual machines to other ESX hosts in the cluster is not a VMotion functionality, but this is a feature of VMware HA. The vSphere availability Guide:

  "VMware HA offering high availability for virtual machines by pooling their and the hosts they reside on in a cluster." Cluster hosts are monitored and in case of failure, the virtual machines on a failed host are restarted on alternate hosts. »

  Therefore, enable HA of the cluster to achieve the desired result.

  Best regards, Robert

  Post edited by: RvdNieuwendijk

• Very bad performance in windows 8

  Hello

  I just moved from windows 7 to windows 8 and in extremely poor performance.
  Windows 7 my system is very fast and responsive. I7 2600 k, 8 GB Ram and 2 x GTX480 in SLI so much grunt but runs windows 8 so slow that my computer is useless limit.
  Hangs regularly just for no reason and it is very annoying, especially after spending $50 for Win7. It is also a clean installation of full format for win8 should work at its best.
  Any help would be appreciated. If this is the way, does anyone know if I can get my money back.
  Thank you

  Hi KmxRz,

  What is the brand and model of your PC?

  Thank you!

  Mine is a customized version, but it seems that my problem is solved.
  My pc just prompted me to update with keyboard and microsoft mouse drivers (I thought they would be generic as part of windows I use microsoft keyboard and mouse) and the problem seems to have disappeared.
  Looks like the keyboard and mouse drivers were win8 go nuts I think that maybe it's true, it seems that every time I hit the tab key to move between fields.
  I hope that this is the solution.
  Thank you

• Very bad performance for Qosmio X 70-B-10 t battery

  I recently bought a laptop 10 X 70 - b - t. The battery is terrible, I'm just using the computer laptop just surfing the web and using microsoft word etc not play games or watch movies and I get a little less than 2 hours. I then opened and took a quick glance at the battery, it's a PA5109-1BRS but when I saw the MAH is 4200MAH where as the PA5109-1BRS battery are normally 4400MAH according to the toshiba site and "Topaz". This could be the battery problem? laptops not even 20 days.

  It's just for a Qosmio. 200mAh will make very little difference, a few minutes on it. 2 h is actually pretty good, my X 870 scrapes barely 90 minutes.

• Qosmio X 70-A-11 - very bad performance

  Hello

  I am owner of a Qosmio X 70-A-11 for about a year and a half, and recently it started to run poorly, especially when the games.

  I OC the GPU, programs not essential disabled, clean the DDI and the hard disks, use a pc tune up... but no significant results.

  Here are some examles:
  -Battlefield 3: low settings @ 720 p-online 45-50 fps
  -Battle field 4: low setting @ 720 p-35-40 fps online
  -L' crew: low settings @ 1080 p-online fps 17-19
  -Car project: ultra settings @ 1080 p-online 9-10 fps
  -Project cars: low settings @ 1080 p-19 and 20 fps online
  -Light to die: low settings @ 1080 p-online 20-25 fps
  and so on...

  Yes, I do run with the GPU performance (Gtx 770 m), use plan of good power and hava the GPU driver up-to-date.

  Any suggestions?
  Thank you.

  BIOS settings reset default please. your system is running low power mode. I have in the field of battle 4: live @ 1080 p-online 65-70 fps
  Switch to normal cpu dynamik

• Set of workflow Performance Virtual Machine

  Hi all

  I'm trying to fix limits of network performance using the workflow to "Define the performance of the virtual machine" in the library of Orchestrator (library - & gt; vCenter - & gt; Virtual Machine Management - & gt; Basic). My network is a dvPortGroup - and I want to put things like average bandwidth, bandwidth, burst size.

  I know that this stream supports and when I run the stream with the values defined by hand, newspapers said the executed correctly flow. However, when I look at the properties of the dvPortGroup I see no change.

  I'm confused on how this should work. My virtual machine will have two NICs (connected to two different dvPortGroups) and I want to apply the formatting only to a network card. Where is this done shaping? In the virtual machine or the dvPortGroup? Where I should be able to see the applied values?

  Thank you!

  VirtualMachineConfigInfo.networkShaper property can be changed only in ESX 2.5 and earlier versions. For ESX 3.0 and later versions, this property is not populated. In order to change the newtork formatting, you must configure the network traffic by changing the HostNetworkTrafficShapingPolicy which is available on the HostVirtualSwitch or HostPortGroup.

  As a result, the workflow would not change everything well for network shape for ESX 4.0 and this is why you do not see the changes.

  Hope it will be useful.

• Cannot depend on the virtual machine after reboot

  Hello world

  Thanks for reading this post.

  My version VM ESXI4.1.0, 348481.

  Here's my problem:

  One day my VM was hanged, and I tried to restart.

  After the reboot, it can be turn on. And I tried to remove the inventory and to recreate the virtual machine.

  However, I can't find the. VMX file in the data store.

  So I tried to create a new virtual machine and put the files in this folder, but it can also work.

  Please see the capture screen, (http://clip2net.com/s/2aNUG)

  Could someone give me guidance and advice how to solve?

  Thank you very much.

  According to the latest vmware.log file, you have tried to access one of the files base/flat with a wrong header file (40 GB)! This hopefully not bad data on the virtual disk. In any case, I've recreated the missing files (see attachment), so you should be able to redo the work of the virtual machine. Because the vmware.log file did not help, I've recreated the instant chain according to the timestamps of the delta file.

  Please follow it below mentioned steps exactly to avoid problems in case something does not work as expacted:

  • Make sure the flat files as well as files 2 6 delta are still in place (not renamed,...)
  • remove the virtual machine from the inventory of the host (right-click-> "Remove from inventory") ESXi
  • delete the temporary/second .vmx unless there is a special need for it
  • Download the files from the .zip file in the folder of the virtual machine on the data store
  • Click right newly downloaded virtual machine's .vmx file and choose "add to the inventory.
  • create a new snapshot (this is important!)
  • Turn on the virtual machine

  If something is not or you have questions, please send back before moving on to the other steps.

  If everything works as expected you should consider the 'Delete All' snapshots in snapshot Manager. This will release not only almost 200 GB of disk space, but can also improve the performance of the virtual machine. Because the size of the snapshot, remove snapshots may take some time, probably in hours!)

  André

• Very slow performance on EVA4400

  Hello guys,.

  I need help about:

  So I have a configuration consisting of a few servers HP ProLiant DL380 G5 (2 x Quad-Core Xeon X 5470 + 24 GB of RAM) with dual-port HP FC2242SR 4 GB/s HBA (enforcement of a vSphere 4.1) connected to a HP EVA4400 consisting of 24 300 GB 15K FC disks.

  The servers each have 2 network cards connected to a Cisco 2960 G switch and connected to a dvSwitch in vSphere.

  The disks are grouped into a single DiskGroup on the EVA, and I a Vdisk 1040 of GB (LUNS) created with RAID level 5.

  Well, now I have a virtual machine with Windows Server 2008 R2, acting as a file server. Is the virtual material consists of 4 GB of ram and 4 cores of the processor and the LUN I mentioned above is mounted as a data store VMFS to the host and divided up like a vmdk of 40 GB for the OS (with LSI SAS controller) and an another vmdk of ~ 1 TB on a controller separated from Paravirtual... .intended for real storage of files. The two drives are configured as Thin Provisioning

  And also it has a single VMXNET3 NIC.

  The problem is that the file server is running very bad, it is about 10-12 Mbps for writing and 20 to 30 Mbps per bed.

  I tested it by transferring ~ 2, 3 ~ and ~ 4 GB .iso files to and from the server.

  I have tried to my workstation, other virtual machines & physical servers... the performance is the same.

  I also vMotioned the file to another server hosts... still the same thing.

  I thought that maybe it is something wrong with the virtual machine itself... so since I was an another windows 2008 R2 VM residing on a smaller (150 GB) LUN I did the same test by copying .iso to that... and the performance was the same. And I also did a Linux VM... his thing.

  Any ideea what the problem is there... because it just seems very bad performance... it's as much slower even than a normal PC

  Basically there is virtually no traffic on the fibre channel network most of the time... it's not congestion or anything like that.

  The EVA and the FC switches are running the latest firmware.

  Help, please.

  Thank you

  fight the block cache on EVA are fully charged? or is there any failuer disk raid reconstruction is done in the background.

• 0 MB of CACHE physical memory, bad Performance

  My computer has a very bad performance. It shows in taskbar (PERFORMANCE tab) I don't have enough physical memory, physical memory is about 90 MB, total is Mo 4043, also available is Mo 1684, Cached's Mo 1628, sometimes free memory is 0. Is it possible to reduce the size of the physical cache or increase the physical memory (without RAM upgrade)?

  Here is the screenshot of my Task Manager(performance Tab).

  Hello

  Did you make any changes before the show?

  Refer to the article below:

  Optimize Windows 7 for better performance

  http://Windows.Microsoft.com/en-in/Windows7/optimize-Windows-7-for-better-performance

  Note: The data files that are infected must be cleaned only by removing the file completely, which means that there is a risk of data loss.

  Change the size of virtual memory

  http://Windows.Microsoft.com/en-in/Windows7/change-the-size-of-virtual-memory

  For reference:

  Usable memory may be less than the memory installed on Windows 7 computers

  http://support.Microsoft.com/kb/978610

  Note: Changing the BIOS / oxide of additional metal (CMOS) semiconductor parameters incorrectly can causeserious problems that may prevent your computer from starting properly. Microsoft cannot guarantee that problems resulting from the configuration of the BIOS/CMOS settings can be solved. Changes to settings are at your own risk.

  Let us know if that helps.