Install and configure software required for jobs
After you set up the Deadline Cloud worker agent, you can prepare the worker host with any software that is required to run jobs.
When you submit a job to a queue with an associated jobRunAsUser, the job runs
as that user. When a job is submitted with commands that are not an absolute path, that command
must be found in the PATH of that user.
On Linux, you might specify the PATH for a user in one of the following:
-
their
~/.bashrcor~/.bash_profile -
system configuration files such as
/etc/profile.d/*and/etc/profile -
shell startup scripts:
/etc/bashrc.
On Windows, you might specify the PATH for a user in one of the following:
-
their user-specific environment variables
-
the system-wide environment variables
Install digital content creation tool adaptors
Deadline Cloud provides Open Job Description (OpenJD) adaptors for using popular digital content creation (DCC) applications. An adaptor is a command-line program that runs the DCC application on the worker host, keeps the application loaded between tasks, reports render progress to the job's logs, and applies path mapping. For more information, see Adaptor packages.
To use these adaptors in a customer-managed fleet, you must install the DCC software
and the application adaptors. Then, ensure the software's executable programs are available on
the system search path (for example, in the PATH environment variable).
To install a digital content creation (DCC) adaptor on a customer-managed fleet
-
Open a terminal:
-
On Linux, open a terminal as the
rootuser (or usesudoorsu). -
On Windows, open an administrator command prompt or PowerShell terminal.
-
-
Install the adaptor package for your DCC application. The following example installs the Maya adaptor:
pip install deadline-cloud-for-mayaThe following table lists the adaptor package name and documentation link for each supported DCC application. All adaptor packages are available on the GitHub website:
Installing multiple adaptors on the same worker
If you install more than one DCC adaptor on the same worker, install each adaptor
into its own Python virtual environment. Adaptors can pin different version ranges
for shared Python packages (for example, deadline or
openjd-adaptor-runtime). An environment with multiple adaptors might fail
with a ResolutionImpossible error. It might also silently downgrade a shared
package and break the adaptors that need the newer version.