This article introduces common error scenarios when provisioning Linux VMs by Machine Creation Service.

---

Instructions

In Linux VDA 7.18, users can provision Linux VMs using Machine Creation Service (MCS), you may refer to document for detailed steps on how to achieve this.
Please make sure following requirements are met in the environment before go further on troubleshooting the issue.

• supported hypervisors
  • XenServer
  • Microsoft Azure
• supported AD integration methods
  • winbind
  • sssd

Common scenario 1:

Run /opt/Citrix/VDA/sbin/deploymcs.sh fails with error:

Loaded plugins: langpacks, product-id, search-disabled-repos, subscription-manager
No package ntfs-3g available.
Loaded plugins: langpacks, product-id, search-disabled-repos, subscription-manager
No package tdb-tools available.
Solution: Please follow the document to enable the repository. Please install the epel to install the dependencies that MCS needs.

Common scenario 2:

Create Machine Catalog fails with error during copying the master image:

Solution: Please go to Citrix Studio->Configuration->Hosting and verify the configuration for the resource you are connecting to.Please correct the hosting configuration from Edit Connection or Add Connection and Resources accordingly.

Common scenario 3:

Linux VMs are successfully created by MCS, and are added to a Delivery Group. Citrix Studio shows the Linux VMs as registered. When launching those published Linux VMs, "Invalid Login" is returned and user can’t login to the published Desktop by any means.

Solution:

This is possibility caused by WORKGROUP not configured in /var/xdl/mcs/mcs.conf.
In case of the NetBIOS name of the domain is different from the domain name, WORKGROUP must be configured same as the domain name in mcs.conf.

To check NetBIOS name the domain and domain name, from a Command Prompt window, type set.

• The variable USERDNSDOMAIN displays the DNS domain name.
• The variable USERDOMAIN displays the NetBIOS domain name.

If NetBIOS name is different from the domain name, you would need to configure WORKGROUP in mcs.conf file.
In this case, please following steps to resolve the issue.

1. Go to the Master Image (the template VM you created Linux VMs from) ,go to /var/xdl/mcs/mcs.conf, set WORKGROUP the same as the domain name.
2. Run /opt/Citrix/VDA/sbin/deploymcs.sh.
3. Shutdown the Master Image, take a new snapshot.
4. Go to Delivery Controller Server, Studio, select the machine catalog, click Update Machines on Action panel.
5. Select the snapshot created in step3.
6. Wait till the update completes and relaunch ICA session.

Common scenario 4:

Linux VMs are created by MCS with no error, and are added to Delivery Groups. Citrix Studio shows the Linux VMs as Unregistered.
After a few minutes, the VMs are powered off by Citrix Broker Service with Event ID 3015 outputted into application log on Delivery Controller Service.
Event ID 3015
Event Source: Citrix Broker Service
Message:The Citrix Broker Service marked a power action as failed because virtual machine 'domian\machine' failed to register with the site after being started, restarted, or resumed.
 
Solution:
This is possibility caused by WORKGROUP incorrectly configured in /var/xdl/mcs/mcs.conf.
 
In this case, please following steps to resolve the issue.

1. Go to the Master Image (the template VM you created Linux VMs from) ,go to /var/xdl/mcs/mcs.conf, set WORKGROUP the same as the domain name.
2. Run /opt/Citrix/VDA/sbin/deploymcs.sh.
3. Shutdown the Master Image, take a new snapshot.
4. Go to Delivery Controller Server, Studio, select the machine catalog, click Update Machines on Action panel.
5. Select the snapshot created in step3.
6. Wait till the update completes and relaunch ICA session.​

Loggings that help to troubleshoot MCS related issues on Linux VDAs:

• /var/log/ad_join.log
  • Used for troubleshooting domain joining issues on Linux VDAs created by MCS.
  • Debug level is enabled by fault.

• /var/log/deploymcs.log
  • Used for troubleshooting deploymcs.sh running issues on Master Image.
  • Debug level is enabled by fault.