Jump to content

Calvin Chan

Legacy Group
  • Posts

    20
  • Joined

  • Last visited

Posts posted by Calvin Chan

  1. We've been having a lot of trouble getting Sophos Endpoint Protection to work in our Win10 Xendesktop 1912 LTSR (PVS) with App Layering. Before Sophos, we used Symantec Endpoint Protection. Symantec was also a bit difficult to get working, but it was less problematic in our experience.

     

    What do you guys use for your environments? 

     

    We are considering going back to Symantec(Broadcom) again because they worked in the past. However I noticed on the Citrix page - (https://docs.citrix.com/en-us/citrix-app-layering/4/layer/layer-antivirus-apps.html#symantec-endpoint-protection) that the 'latest' version support is still version 14.2 and according to Broadcom, 14.2 is over 10 version releases old.  https://knowledge.broadcom.com/external/article/154575/versions-system-requirements-release-dat.html 

     

    The Citrix doc also seemed a bit outdated.  Does anyone have a working solution and recipe for their environment? 

     

    Windows 10 v1909

    Citrix Virtual Desktops LTSR 1912 CU2

    Provisioning Service LTSR 1912 CU2

    App Layering (ELM latest) - Full User Layer

    No elastic app layers

  2. Quote

    I was able to configure DHCP in a way that a client with a specific hostname  would always get a single IP. I used the below configuration.

    1. In a /24 network, I created a range with only a single IP address and another range as big as required.
    2. Created an option filter under 'IPv4 Filters' with the rule as Option 12 equals my hostname.
    3. Added a class filter to the range with a single IP address with the option filter that I created and configured the action as 'Grant Lease'
    4. In the bigger range ,I added the option filter as a class filter with the action ' Deny Lease'

    I tested this with a DHCP relay agent simulator and found that when the hostname in option 12 is the hostname that I specified in the filter, the lease is given with the IP in the first range.

     

    Below are the relevant parts from the DHCP conf file.

     

    
        # Option filter "teswt1"
        class "teswt1" {
        match if (option host-name="jarvis");
        }
    
        pool {
            infoblox-range 10.192.16.2 10.192.16.2;
            range 10.192.16.2 10.192.16.2;
                    allow members of "teswt1";
            ddns-domainname = pick ( option fqdn.domainname, config-option domain-name );
        }
        pool {
            infoblox-range 10.192.16.4 10.192.16.254;
            range 10.192.16.4 10.192.16.254;
                    deny members of "teswt1";
            ddns-domainname = pick ( option fqdn.domainname, config-option domain-name );
        }

     

     

     

    https://community.infoblox.com/t5/DNS-DHCP-IPAM/Is-it-possibble-to-do-DHCP-Reservation-by-hostname-not-by-mac/m-p/11353#M2344  

    I found it here. Good luck!

     

     

  3. VDA

    Windows v1909,

    Citrix Xendesktop LTSR 1912 VDA

    Citrix PVS 7.15 LTSR

     

    Client PC - Windows 10 v1909 Workspace App 2002 

     

    The VM appears to be mapping the local audio devices, however they all appear as 'Citrix HDX Audio' where as before, it reflects the actual local audio device names.

    Also, in addition to the names not appearing, there is also no audio coming through. Have tried restarting the VM, service, client ICA, client PC, and reinstalling the ICA workspace client.
    Attached screenshot for a little more information.

    Is there any I can investigate as to the real status of the audio devices? or advice to troubleshoot this?

     

     

    Notif_area.png

    devmgr.png

  4. 1 minute ago, Rob Zylowski1709158051 said:

    No that was for Unidesk 2.x only.  In App Layering  the layers are stored thin provisioned so you dont have the same issue.  

     

    Are you publishing to PVS or MCS?  If MCS are you using the compositing engine that will make vSphere vmdk's thin provisioned.  If its PVS is that the size you see in the console or are you looking at the size of the vDIsk?

     

    Hi Rob,

     

    Thank you for that quick response!!! I am publishing to PVS, I was looking at the size on the last page of Publishing Image in the ELM. And interestingly, that lead me to check the actual size in the file storage server where the image is kept... only 50GB. 

  5. I've just created a new OS layer for the Win10 1909 upgrade, and along with that, LTSR 1912 VDA on the platform layer. After publishing the new image, I noticed the final size is getting larger and larger ~ 80GB. My question is, is this tool still valid? https://support.citrix.com/article/CTX226105 Citrix App Layering: Compress Layer Utility. If this isn't what I actually need, what can I do to help shrink down the published image size? Does it even matter? 


    For application layers, I always install/update from network share installers and download them using one of the non-citrix VM.

    After I've completed my work, I always run the JGSpiers script to finalize. 

     

    Thank you everyone in advance.

    ESXi, vSphere 6.5

    Win10 1909

    Farm is 7.15 LTSR, pending 1912 LTSR upgrade in matter of days

    ELM 20.3.0.12

     

  6. 13 hours ago, Thomas Rodgers said:

     

    OK we had a similar issue with this when the Hypervisor tool was installed into the platform layer, we ended up having to recreate it. Installing the Hypervisor tools into a new platform layer > Finalizing the layer > Add New version > Then continue on with the Nvidia drivers, VDA Agent etc.

     

    In our experience, the platform layer has issues if it isn't finalized straight after the Hypervisor tools are installed

     

    Doing it this way also negated the need to add the Unifilter Registry keys for the Nvidia Installation folders

     

    Oof, that's quite troublesome... and it seems like no matter which base version i go to, the NVidia comes out missing on the new version. Wonder if the app layering team has something like this in their KB... probably should start a ticket at this point.

  7. Latest version of ELM (as of today)

    7.15 LTSR CU3, PVS

    Windows v 1709

    Full Elastic enabled for User Layer, but no elastic layering are used for apps. So the VM's are running image only + user layer.

     

    Trying to update something on the platform layer, however the new versions coming out are missing the Nvidia /GRID components, the folders are still there, but emptied or only containing a blank shortcut. Got a screenshot here.. http://prntscr.com/q7qhed

     

    The NVidia program files folder are already added to the registry HKLM\System\CurrentControlSet\Services\unifltr

     

    nvidia-error.png

  8. 1 hour ago, Rob Zylowski1709158051 said:

    I think the files are created when you finalize a layer.  The ones you have you probably created after this feature was introduced so you can probably add a version to the pother layers you dont have and when you finalize them they should be created as well.

     

    That  is just an educated guess.

    Hmm that would make sense, but I finalized a layer 3 days ago (Platform), and I don't see a repair json for it. We've been on the latest ELM since week 1 of its release. It's not a huge problem, we can always contact an engineer to help us generate it manually, but It would be more convenient if it was something we had some control over.

     

  9. 48 minutes ago, Rob Zylowski1709158051 said:

    Search for "Repair a User Layer" in here https://docs.citrix.com/en-us/citrix-app-layering/4/layer/enable-user-layers.html

     

    That should help.

     

    Hi Rob, that worked -  but i noticed there are only about 4-5 of these json files available, yet we have 15+ application layers, and 3 platform layers.

     

    According to the doc, we have to contact an engineer to manually generate those repair files if they aren't generated, is this still valid? 

    Is there a json repair for Platform layers?

     

    Thank you

  10. Hello everyone,

     

    We're currently on 7.15 LTSR CU3, PVS

    Windows v 1709

    Full Elastic enabled for User Layer, but no elastic layering are used for apps. So the VM's are running image only + user layer.

     

    Is it possible to undo changes made on the user layer ex: uninstalled/updated certain programs ? so that the user is realigned to the image?

    We've been having some difficulty keeping all the auto-updaters in check, and sometimes an admin would slip and allowed the process to go through. 

     

    Thank you

  11. I tested by trying to publish another image via a different template, and that didn't work either, which rules out the work I did on the platform layer revision. A wild google search revealed one other person has experienced this issue and it was caused by the certificate.  We re-install a fresh certificate on the ELM and problem has gone away. Thank you Rob for the assistance!

  12. 10 minutes ago, Rob Zylowski1709158051 said:

    Is this something that worked before or is it your first try?  

       This is something that worked previously, I added a new version to the platform layer for troubleshooting purposes on another support ticket.

     

    10 minutes ago, Rob Zylowski said:

    Do you have a template defined in your connector config?  Maybe try without the template as a test if you do?

      hmm, this connector for the output is on Citrix Provisioning Services, no template option on that page

     

    Do you see any errors in your vcenter?  Or the host you chose for publishing?

      no error in vcenter, was able to create target device/packaging machine just fine, finalizing as well.

     

    if you open a ticket support can look at the detailed logs. 

     

    Thank you Rob for your prompt reply in this matter! that was quick.

  13. Trying to publish a new image today and on the last step, we ran into this error:  "The file copy was aborted after 5 transient errors."

    Attached is a log from the ELM console. The credentials tests, connection tests all pass. The service account on the ELM has proper access to filestore... 

    Doesn't seem like anyone else have had this problem.

     

    Our environment is ESXi vSphere 6.5 

    7.15 LTSR CU3
     

    pvs_elm_error.png

×
×
  • Create New...