No announcement yet.

citrix xenapp and revit 2012

  • Filter
  • Time
  • Show
Clear All
new posts

    citrix xenapp and revit 2012

    we have a computer setup to run citrix and revit. Computer specs are

    Dual xeon X5670 2.93ghz
    48gb ram
    1gb nics
    2 500gb SAS – raid
    1 360gb OCZ ibis solid state drive with HSDL link
    nvidia quadro 4000

    xenserver 6.5 free install with 2 virtual servers running 2008r2(one has xenapp 6.5 and the other citrix lic manager.
    We have installed office 2007, sketchup 8, speclink+, bluebeam, costworks, abobe CS5 and they all work fine. Got autodesk citrix network license. Install network license manager on the xenapp server along with revit 2012. Open revit on the server and it opens fine and gets a license. Login with citrix receiver and the user doesn't get a license when revit is open and the program will not open. Login citrix receiver as administrator(same user who everything was install under) and it will get a license and work fine. Anyone have any idea why other users can't get a license? Also any ideas on how to get hardware acceleration enabled in revit and sketchup. Can't get them to recognize the video card

    I think you are much more likely to get some useful info here compared to the General forum (so I deleted the post there, in case anyone is wondering ).
    FWIW I know someone who is borderline raving about XenApp, so hopefully that means some good news will be forthcoming.

    Pragmatic Praxis


      Since there wasn't much traffic, I pinged my friend Ken about this. He isn't a member (WTH? I know. ) so I am passing this along.

      From the sounds of your process, you are logging in as an Administrator on the server, not a user, and getting the license. Per Ken, "I would have the user that can't get a license via Xenapp log directly onto the server as themselves and see if they can open revit and pull a license that way. If they can't then it has nothing to do with Xenapp."

      Now if this is the case, I have no idea what to do from there. The thought of making all users admins makes me cringe, but it is common practice in a Windows environment, especially with Autodesk.

      Pragmatic Praxis


        logged on to xenapp server as domain administrator, installed revit and can pull a license with doamin admin on the server and when logging in with citrix reciever. Log in with another user on the citrix reciever and that user fails to pull a license and gets error 15something which means NLM server failed to contact or something. Have not tried the test of user not being able to pull license, logging onto server and seeing if they can pull a license that way. Thanks!


          Just in case anyone else runs into this issue with citrix or even just NLM for autodesk products on a 2008r2 server. This issue ended up being the version of NLM we had installed. Our office is a IPv4 network with 2003 servers and win 7 machines. The citrix server is a 2008r2 server and this is where the NLM for our citrix products is. We had installed the network license manager for IPv4 and disable IPv6 on the 2008r2 server since we are a IPv4 network. Taking the 4 EXE files from the NLM IPv6 version and overwriting the IPv4 EXE files then NLM started handing out Licensees to everyone. Autodesk said even if you do not use IPv6 for some reason NLM still must use some protocol when installed on a 2008r2 server.
          All is good now, we have hardware acceleration enabled on revit and autocad architecture 2012 versions and they are performing VERY well over the WAN. xenserver 6.0(allows GPU passthrough to VM server with xenapp 6.5) with xenapp 6.5. We are still trying to decide about publishing the apps from xenapp to xendesktop or just leave it how it is setup now. So far we are very pleased with performance compared to hyper-v, remote desktop RDP or even riverbed steelhead that we have tested over the WAN!


            How did you "virtualize" your graphics cards/GPUs? We are planning a Virtualization test, and are having problems with the graphics card part.
            CPUs and RAM, OS and Apps seem pretty straightforward, just the GPU/graphics we are struggling with. We are told that Citrix HDX and Remote FX
            are the way to achive this. Any experience/advice?
            Cliff B. Collins
            Registered Architect
            The Lamar Johnson Collaborative Architects, St. Louis, MO
            Autodesk Expert Elite


              We tried it with xenserver 5.5, but was unsuccessful. Then citrix released xenserver 6.0 (just in the past 2 weeks) and it work. From what I have read it only works nvidia quadro and quadro fx cards, plus you need a bios/processor that supports VT-d and xenserver Enterprise editions and higher.
              Install xenserver 6.0 and it should recognize the GPU. I believe it is a 1:1 ratio so we decided to assign the GPU to server with xenapp on it instead of a virtual desktop. Assigning it to xenapp allows all apps and people access to the GPU, but assigning it to a virtual desktop would have meant buying 1 GPU for each virtual desktop. We have 2 virtual servers running 2008r2 server sitting on top of xenserver. Assign the GPU to the virtual server that has xenapp on it, below are the instructions on how to do that. Then login to that server using remote desktop, go under devices and you should see 2 GPU's and one is the nvidia gpu. Now just install the driver for it and you are done. The instructions say that it is only supported for xendesktop, but like I said it works in xenapp as we have done. You can still use xendesktop and publish the apps from xenapp to those desktops with the GPU still recognized.

              Assigning a GPU to a Windows VM (for Use with XenDesktop)
              XenServer allows you to assign a physical GPU in a XenServer host machine to a Windows VM running on the
              same host. This GPU Pass-Through feature is intended for graphics power users, such as CAD designers, who
              require high performance graphics capabilities. It is supported only for use with XenDesktop.
              GPU Pass-Through is available to Citrix XenServer Enterprise editions and higher.
              While XenServer supports only one GPU for each VM, it automatically detects and groups together identical
              physical GPUs across hosts in the same pool. Once assigned to a group of GPUs, a VM may be started on any host
              in the pool that has an available GPU in the group. Once attached to a GPU, a VM has certain features that are
              no longer available, including XenMotion live migration, VM snapshots with memory, and suspend/resume.
              Assigning a GPU to a VM in a pool does not interfere with the operation of other VMs in the pool. However, VMs
              with GPUs attached are considered non-agile. If VMs with GPUs attached are members of a pool with HA or WLB
              enabled, those VMs are overlooked by both features and cannot be migrated automatically.
              GPU Pass-Through is available to Windows VMs only. It can be enabled using XenCenter or the xe CLI.

              GPU Pass-Through is supported for specific machines and GPUs. In all cases, the IOMMU chipset feature (known
              as VT-d for Intel models) must be available and enabled on the XenServer host. Before enabling the GPU
              Pass-Through feature, visit Citrix XenServer Hardware Compatibility List » Citrix Ready to check the hardware compatibility list. For any further
              questions regarding the hardware compatibility list, e-mail [email protected].
              Before Assigning a GPU to a VM
              Before you assign a GPU to a VM, you need to put the appropriate physical GPU(s) in your XenServer host and
              then restart the machine. Upon restart, XenServer automatically detects any physical GPU(s). To view all physical
              GPU(s) across hosts in the pool, use the xe pgpu-list command.
              Ensure that the IOMMU chipset feature is enabled on the host. To do so, enter the following:
              xe host-param-get uuid=<uuid_of_host> param-name=chipset-info param-key=iommu
              If the value printed is false, IOMMU is not enabled, and GPU Pass-Through is not available using the specified
              XenServer host.
              To assign a GPU to a Windows VM using XenCenter:
              1. Shut down the VM that you wish to assign a GPU.
              2. Open the VM properties: right-click the VM and select Properties.
              3. Assign a GPU to the VM: Select GPU from the list of VM properties, and then select a GPU type. Click OK.
              4. Start the VM.
              To assign a GPU to a Windows VM using xe CLI:
              1. Shut down the VM that you wish to assign a GPU group by using the xe vm-shutdown command.
              2. Find the UUID of the GPU group by entering the following:
              xe gpu-group-list
              This command prints all GPU groups in the pool. Note the UUID of the appropriate GPU group.
              3. Attach the VM to a GPU group by entering the following:
              xe vpgu-create gpu-group-uuid=<uuid_of_gpu_group> vm-uuid=<uuid_of_vm>
              To ensure that the GPU group has been attached, run the xe vgpu-list command.
              4. Start the VM by using the xe vm-start command.
              5. Once the VM starts, install the graphics card drivers on the VM.
              Installing the drivers is essential, as the VM has direct access to the hardware on the host. Drivers are
              provided by your hardware vendor.
              If you try to start a VM with GPU Pass-Through on a XenServer host without an available GPU
              in the appropriate GPU group, XenServer prints an error message.
              To detach a Windows VM from a GPU using XenCenter:
              1. Shut down the VM.
              2. Open the VM properties: right-click the VM and select Properties.
              3. Detach the GPU from the VM: Select GPU from the list of VM properties, and then select None as the GPU
              type. Click OK.
              4. Start the VM.
              To detach a Windows VM from a GPU using the xe CLI:
              1. Shut down the VM by using the xe vm-shutdown command.
              2. Find the UUID of the vGPU attached to the VM by entering the following:
              xe vgpu-list vm-uuid=<uuid_of_vm>
              3. Detach the GPU from the VM by entering the following:
              xe vgpu-destroy uuid=<uuid_of_vgpu>
              4. Start the VM by using the xe vm-start command.


                We are just starting to look at Citrix and Revit seriously. A critcal project on the horizon is a large healthcare facility with most of the design team in Australia but some in the states. Initial testing shows that our latency is 300-400ms and are wondering if this is a deal breaker. Is Citrix still constrained to the <200ms recommendation for the user experience to be acceptable?


                  could you please tell me the way to enable hardware acceleration in revit.?
                  I am using revit architecture 2014+XenApp embeded in xendesktop 7 +server2008R2(baremetal system).

                  When i launch revit achitecture from citrix receiver session, the gpu is not detected.The application detects the GPU when launched directly in the system.


                    Originally posted by jasonbailly View Post
                    We are just starting to look at Citrix.. latency is 300-400ms and are wondering if this is a deal breaker.
                    Jason, we have Citrix set up, and acceptable user experience is subjective. One of our designers in Kuwait was dissatisfied with the performance from a London Citrix farm, yet our Pune, India team is satisfied using the Citrix farms in Seattle, NYC and London. I suspect they prefer Citrix to the 20-30 minute open/synch time for Revit Server projects.

                    I suspect that spanning one continent is acceptable. Multiple continents is dubious. But, then again, routing and bandwidth may be better indicators than latency alone? HTH.
                    Last edited by Markus; March 21, 2014, 07:06 PM.


                    Related Topics