storcon: use separate scheduling context for each tenant in fill/drain

VladLazar commented 3 months ago

Problem

Service::drain_node and Service::fill_node both use a ScheduleContext in an attempt to spread out shards belonging to the same tenant. However, they're not resetting the schedule context after processing each tenant.

Summary of changes

Reset the scheduling context after each tenant. We are using the fact that the shards inServiceState::tenants are already ordered by tenant id.

Checklist before requesting a review

[ ] I have performed a self-review of my code.
[ ] If it is a core feature, I have added thorough tests.
[ ] Do we need to implement analytics? if so did you add the relevant metrics to the dashboard?
[ ] If this PR requires public announcement, mark it with /release-notes label and add several sentences in this section.

Checklist before merging

[ ] Do not forget to reformat commit message to not include the above checklist

github-actions[bot] commented 3 months ago

3228 tests run: 3111 passed, 0 failed, 117 skipped (full report)

Flaky tests (1)

#### Postgres 14 - `test_subscriber_restart`: [release](https://neon-github-public-dev.s3.amazonaws.com/reports/pr-8116/9594497952/index.html#suites/8be0c222d5601535470e7e5978bbfb03/1171c9bff3b14c4/retries)

Code coverage* (full report)

functions: 32.3% (6844 of 21166 functions)
lines: 49.7% (53396 of 107369 lines)

* collected from Rust tests only

_{The comment gets automatically updated with the latest test results
83fdc07b20afb411eb00587331b88259a1ff6742 at 2024-06-20T09:41:57.782Z :recycle:}

VladLazar commented 3 months ago

Replaced by https://github.com/neondatabase/neon/pull/8119

neondatabase / neon