-
Notifications
You must be signed in to change notification settings - Fork 25k
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
[ML] Prefer cluster state config to index documents #36014
[ML] Prefer cluster state config to index documents #36014
Conversation
This reverts commit 417624f.
Pinging @elastic/ml-core |
No issues with the PR really, but what is the reasoning for the reverting? Why should we prefer the cluster state over the indexed configs? |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Well @davidkyle , there you stand, with your code in your hands:
https://www.youtube.com/watch?v=c0SvsinEUfY
:D
Actually doing this uncovered a bug where on closing a job the check that the job's datafeed must be stopped was only checking the index configs and not the clusterstate jobs. This explains an intermittent failure I'd seen in the rolling upgrade tests. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
👍 👍 👍 👍 👍 👍 👍
A partial revert of #35940
Duplicate configs in both the clusterstate and index are still possible but now clusterstate config is preferred.
After discussing the changes in #35940 it was decided that during a move operation the original should always be preferred to the new until the operation has completed. This revert also makes the code simpler to read.
Alas the wasted time
@benwtrent you reviewed #35940 would you take a look at this please