I couldn’t find anything about how to use this in Houdini?

Deadline Cloud [press.aboutamazon.com]

Didn't test it yet, but it looks like you need a login to AWS account to get access to the download of Deadline Cloud submitter :
https://docs.aws.amazon.com/deadline-cloud/latest/userguide/submitter.html [docs.aws.amazon.com]

The doc says that after installation of the submitter you get access to the submitter inside ROP network, like the classic submitter. As I didn't install it yet I can't tell if it's the same ROP node with additional parms, or if it's a separate ROP node.

Okay thanks, I’m gonna try this out and see if it really works.

Curious to Clone some things with Karma in the Cloud and what price point that comes out too per-frame.

Hey did you already get it to work? we're only getting lots of error messages so far.

+1, we're a bit stuck with setting up AWS Deadline Cloud on and trying to integrate Redshift. Any advice or examples, anyone?

Following the documentation I was able to get Deadline Cloud working yesterday. The installer for the Houdini submitter doesn't drop in the needed env settings though so you need to grab this .json and modify it and add it to your packages folder: https://github.com/aws-deadline/deadline-cloud-for-houdini/blob/mainline/packages/deadline_submitter_for_houdini.json [github.com]

The installed HDA creates a separate deadline_cloud node type in /out context so you can keep your existing local deadline setup as well:
Make sure to set the AWS settings in "settings" on your deadline_cloud node the first time you use it.
I found that submission of a test job using vanilla Houdini and retrieval of rendered frames pretty efficient and painless.

As for Redshift or custom HDAs you'll need to make your own EC2 AMI with the cloud worker agent installed to get it to work with deadline_cloud. Install instructions are here: https://github.com/aws-deadline/deadline-cloud-worker-agent [github.com]

Personally I try to stick with Karma for rendering as much as possible and I've been moving further and further away from Maxon and Redshift for anything that I can't handle rendering locally.

Perfect timing for me to get the help I needed to set it up, just the day after your post.

etwiss
Following the documentation I was able to get Deadline Cloud working yesterday. The installer for the Houdini submitter doesn't drop in the needed env settings though so you need to grab this .json and modify it and add it to your packages folder: https://github.com/aws-deadline/deadline-cloud-for-houdini/blob/mainline/packages/deadline_submitter_for_houdini.json [github.com]

The installed HDA creates a separate deadline_cloud node type in /out context so you can keep your existing local deadline setup as well:

Image Not Found

Make sure to set the AWS settings in "settings" on your deadline_cloud node the first time you use it.
I found that submission of a test job using vanilla Houdini and retrieval of rendered frames pretty efficient and painless.

As for Redshift or custom HDAs you'll need to make your own EC2 AMI with the cloud worker agent installed to get it to work with deadline_cloud. Install instructions are here: https://github.com/aws-deadline/deadline-cloud-worker-agent [github.com]

Personally I try to stick with Karma for rendering as much as possible and I've been moving further and further away from Maxon and
Redshift for anything that I can't handle rendering locally.

edward_twiss
As for Redshift or custom HDAs you'll need to make your own EC2 AMI with the cloud worker agent installed to get it to work with deadline_cloud. Install instructions are here: https://github.com/aws-deadline/deadline-cloud-worker-agent [github.com]

Could you please clarify - we need to install a dedicated worker agent because Redshift is not supported natively (btw - where is the list of supported rendering engines or is it only those that are internal to DCC?) and so we need to install Redshift with a license on some EC2 and make it work? Am I understanding this correctly?

Many thanks for the post above, it wasn't obvious why they didn't add their JSON to the packages.