Record Ray actor and placement group ids as strings - #181
Merged
Conversation
Ray reports actor_id and placement_group_id as hex strings, but both the polars schema and the perspective schema have declared them numeric since the initial commit. Tracking any actor task therefore raised TypeError from get_df(), and the dashboard rendered every id as 0.0 — silently wrong rather than visibly broken. No test had ever tracked an actor task, only plain remote functions, which is how this survived. Add one that asserts the ids round-trip as hex strings. Also from the same review pass: Sync whatever landed before a batch raised. `apply` mutated tables and then refreshed state at the end, so a bad row left the earlier tables applied but unsynced, and the next batch's replayed schemas reported no change — leaving them invisible until an unrelated update arrived. Check the dashboard mode mismatch for every mode, not just "local". An existing actor attached with the wrong mode was only reported in one of the three cases. Wait on the custom table's row count rather than its name, which passed before the update was applied and so would not have caught a dropped update. Signed-off-by: Tim Paine <3105306+timkpaine@users.noreply.github.com>
Contributor
timkpaine
marked this pull request as ready for review
August 21, 2026 23:16
This was referenced Aug 21, 2026
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Found by an adversarial review pass and confirmed empirically.
The bug
Ray reports
actor_idandplacement_group_idas hex strings, but both the polars schema and the perspective schema have declared them numeric since the initial commit (Jan 2024). Tracking any actor task:get_df()raisesRayTaskError(TypeError)0.0— silently wrong rather than visibly brokenVerified before the fix:
and after:
No test had ever tracked an actor task — only plain remote functions — which is how this survived. This adds one.
Also in this pass
Dashboard.applypartial batches. Tables were mutated and state refreshed only at the end, so a bad row left earlier tables applied but unsynced; the next batch's replayed schemas reported no change, leaving them invisible. Now refreshed infinally, with a regression test.dashboard="local", so two of the three cases attached to a wrongly-configured actor silently.