Skip to content
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

[31/N][VirtualCluster] add virtual cluster documents #470

Merged
merged 1 commit into from
Jan 20, 2025
Merged

Conversation

wumuzi520
Copy link
Collaborator

Why are these changes needed?

This PR is 31/N implementation of the VirtualCluster (see issue #409), Add virtual cluster documents.

Related issue number

Checks

  • I've signed off every commit(by using the -s flag, i.e., git commit -s) in this PR.
  • I've run scripts/format.sh to lint the changes in this PR.
  • I've included any doc changes needed for https://docs.ray.io/en/master/.
    • I've added any new APIs to the API Reference. For example, if I added a
      method in Tune, I've added it in doc/source/tune/api/ under the
      corresponding .rst file.
  • I've made sure the tests are passing. Note that there might be a few flaky tests, see the recent failures at https://flakey-tests.ray.io/
  • Testing Strategy
    • Unit tests
    • Release tests
    • This PR is not tested :(

Copy link
Collaborator

@xsuler xsuler left a comment

Choose a reason for hiding this comment

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

LGTM

**Raylet Failover**
When Rayle dies or restarts, the GCS Server will detect that the node is dead and then update the status of the corresponding node in the virtual cluster, and the virtual cluster manager will periodically check for dead nodes and replace them with new ones.

There are 4 scenario that the virtual cluster manager will replenish new nodes for the virtual cluster, and the state of the virtual cluster need flush or publish:
Copy link
Collaborator

Choose a reason for hiding this comment

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

There are four scenarios in which the virtual cluster manager will replenish new nodes for the virtual cluster, and the state of the virtual cluster needs to be flushed or published:

1. Select new alive undivided node with the same node type as the dead one from :math:`V_{primary}` to replace the dead one.
2. If there are still dead nodes inside :math:`V_{job}`, then continue to new alive undivided node with the same node type as the dead one from it's parent virtual cluster nodes :math:`V_{divisible}` to replace the dead one.
3. If there are any dead node inside :math:`V_{job}` is replenished, then flush the :math:`V_{job}` to the storage and publish it to all the worker nodes.
4. If there are any dead node inside :math:`V_{divisible}` (parent virtual cluster of :math:`V_{job}`) is replenished, then just flush the :math:`V_{divisible}`, without publish.
Copy link
Collaborator

Choose a reason for hiding this comment

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

without publishing

@wumuzi520 wumuzi520 force-pushed the dev-hc branch 4 times, most recently from 934daa2 to 5a67f24 Compare January 20, 2025 10:12

**Output:**

`Virtual cluster 'myCluster' created successfully`
Copy link
Collaborator

Choose a reason for hiding this comment

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

logical1

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

2 participants