New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Manage Pods directly instead of using Job controllers #325

Closed
jlewi opened this Issue Jan 18, 2018 · 7 comments

Comments

Projects
None yet
6 participants
@jlewi
Copy link
Collaborator

jlewi commented Jan 18, 2018

Consensus in #45 was that we should manage the pods directly. Opening this issue to track that work.

@ScorpioCPH

This comment has been minimized.

Copy link
Member

ScorpioCPH commented Jan 19, 2018

Hi, folks, I will pick up this if nobody working on this :)

@jlewi

This comment has been minimized.

Copy link
Collaborator Author

jlewi commented Jan 19, 2018

that would be great!

@gaocegege

This comment has been minimized.

Copy link
Member

gaocegege commented Jan 19, 2018

@ScorpioCPH

Awesome! And I think it is hard to decouple #314 and this issue 🤔 Would you file one PR for these two issues together?

@vishh

This comment has been minimized.

Copy link

vishh commented Feb 15, 2018

Just FYI: @kow3ns and I have been noodling on the idea of a stateful Job that would benefit TF Training Operator. Managing pods outside of existing curated controllers is a risky proposition IMHO, but might be a good stop gap solution.
We will post some updates once we get around to having some concrete designs on that.

@gaocegege

This comment has been minimized.

Copy link
Member

gaocegege commented Feb 16, 2018

@vishh We can discuss it in the issue #45

jlewi added a commit that referenced this issue Mar 5, 2018

Create pod instead of job (#344)
This PR is a part of #325:

rename jobName() to genName()
create Pod instead of Job

TODOs (in another PR):

use controller.PodControlInterface and CreatePodsWithControllerRef to create Pod
Listen Pod CRUD and update TFJob status which descried in #314

Jimexist pushed a commit to Jimexist/tf-operator that referenced this issue Mar 7, 2018

Create pod instead of job (kubeflow#344)
This PR is a part of kubeflow#325:

rename jobName() to genName()
create Pod instead of Job

TODOs (in another PR):

use controller.PodControlInterface and CreatePodsWithControllerRef to create Pod
Listen Pod CRUD and update TFJob status which descried in kubeflow#314

@jlewi jlewi added the priority/p1 label Mar 7, 2018

@jlewi

This comment has been minimized.

Copy link
Collaborator Author

jlewi commented Mar 7, 2018

@ScorpioCPH #344 is merged. Can we close this? Or is there more work to be done?

@ScorpioCPH

This comment has been minimized.

Copy link
Member

ScorpioCPH commented Mar 7, 2018

/close

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment