Skip to content
This repository has been archived by the owner on Sep 20, 2024. It is now read-only.

Deadline: reworked pools assignment #3051

Merged
merged 13 commits into from
Apr 13, 2022

Conversation

kalisp
Copy link
Member

@kalisp kalisp commented Apr 12, 2022

Brief description

Previous solution was nonconsistent, expected values for pools in multiple places in Setting.

Description

New collector and validator was introduced. Collector adds pool names from Setting (merged and removed separate pool settings in hosts sections) if not already present on a instance (as Maya has it, potentially others with New Publisher).

Validator checks if name of pools on the instances matches to current values on DL (does a query of DL).

Additional info

Refactored _requests_get and requests_post wrappers into openpype.lib.connections - name up for commenting/changing.

Testing notes:

  1. configure project_settings/deadline/publish/CollectDeadlinePools
    1.a - insert non existent values
    1.b - insert values matching situation on DL
  2. publish on DL and check pools

@kalisp kalisp added the module: Deadline AWS Deadline related features label Apr 12, 2022
@kalisp kalisp self-assigned this Apr 12, 2022
@ynbot
Copy link
Contributor

ynbot commented Apr 12, 2022

Task linked: OP-2005 Default Pool configurable

@m-u-r-p-h-y
Copy link
Member

m-u-r-p-h-y commented Apr 12, 2022

  • Maya render instance (and all render instances in other hosts) should check the settings and prefill the pools and groups from the settings
  • Pools and Groups should not be a drop-down, but text fields as we can combine multiple groups to fine-tune specific scene requirements. Ex. I need to submit my heavy scene to group Maya, Ram256 and GPU to render on machines with Maya installed, with more than 256 GB RAM and GPU card)
    image

I would suggest to add pools and group to Nuke write instance (the same as we have in Maya)

image

@m-u-r-p-h-y
Copy link
Member

there is also an issue in collect render layers (I'm not sure if it is connected to this PR though)

// pyblish.CollectMayaRender : collecting layer: rs_Main // 
// Error: pyblish.plugin : Traceback (most recent call last):
  File "D:\REPO\OpenPype\.venv\lib\site-packages\pyblish\plugin.py", line 522, in __explicit_process
    runner(*args)
  File "D:\REPO\OpenPype\openpype\hosts\maya\plugins\publish\collect_render.py", line 391, in process
KeyError: 'SecondaryPool'
Traceback (most recent call last):
  File "D:\REPO\OpenPype\.venv\lib\site-packages\pyblish\plugin.py", line 522, in __explicit_process
    runner(*args)
  File "<string>", line 391, in process
KeyError: 'SecondaryPool' // 
b'Processing Context Label'

@64qam 64qam requested a review from m-u-r-p-h-y April 12, 2022 15:45
@kalisp
Copy link
Member Author

kalisp commented Apr 12, 2022

collect_render issue was resolved.

Copy link
Member

@m-u-r-p-h-y m-u-r-p-h-y left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

project pool is correctly propagated to render instance

image

@BigRoy
Copy link
Collaborator

BigRoy commented Apr 13, 2022

I'd expect users might want to set the Pool per project and per application, so that e.g. Maya runs in a different pool than Nuke or Harmony. Thoughts?

@m-u-r-p-h-y
Copy link
Member

m-u-r-p-h-y commented Apr 13, 2022

You are correct. I've already created issue ynput/ayon-deadline#5 to do it outside of this PR as it is influencing a lot of hosts

@kalisp kalisp merged commit 3014e11 into develop Apr 13, 2022
@kalisp kalisp deleted the feature/OP-2005_Default-Pool-configurable branch April 13, 2022 08:54
@kalisp
Copy link
Member Author

kalisp commented Apr 13, 2022

@BigRoy I thought that pools are per project and groups would be per host.
But maybe we should add there profile for host, task name?

Sign up for free to subscribe to this conversation on GitHub. Already have an account? Sign in.
Labels
module: Deadline AWS Deadline related features
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants