-
-
Notifications
You must be signed in to change notification settings - Fork 1.8k
Never use an asynchronous Client when calling top level compute function #11790
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
Conversation
0d6b643 to
6501572
Compare
Unit Test ResultsSee test report for an extended history of previous test failures. This is useful for diagnosing flaky tests. 9 files ± 0 9 suites ±0 3h 30m 49s ⏱️ + 2m 57s For more details on these errors, see this check. Results for commit d6d379c. ± Comparison against base commit 5f61e42. This pull request removes 18 and adds 21 tests. Note that renamed tests count towards both.♻️ This comment has been updated with latest results. |
|
There's probably some fallout in the distributed CI. Particularly |
dask/dask-schema.yaml
Outdated
| type: [string, 'null'] | ||
| description: | | ||
| If False, raise an exception if dask.compute is attempting to select | ||
| an asynchronous Client. If True, fall back to a sync scheduler. |
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.
It looks like this is supposed to be boolean, not a string? Also, what happens on null?
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.
That description is off
|
Test failure is unrelated. Nanny couldn't come up due to some windows foo and the test timed out |
|
thanks |
When using an asynchronous version of the distributed Client, the function
dask.computeandCollections.computemethods are not properly working because theClient.getis returning a coroutine object instead of the actual result. This typically crashes with the repacking logic indask.compute.Making this just work is not possible from my understanding of asyncio and I see two options. For ordinary users it is best to raise in this situation. For our testing infrastructure it is still nice if there was a well defined behavior. For example, read_parquet is triggering a quantile computation this way regardless of the client used.
xref #11736