capistrant opened a new pull request, #20155: URL: https://github.com/apache/druid/pull/20155
### Description The way we have been setting up destinations, we have been setting ourselves to hotspot loads onto a small subset of historicals in a tier, especially if the number of replicas to load on the tier is low. This leads to quickly skewing storage utilization if a lot of new load requests show up in a single coordinator cycle (like a large batch ingest, a change in partial load rules that covers many segments and needs reconciliation, etc.). My goal here is to tap into the existing round robin cursor OR the segment assignment policy (depending on cluster config) to better spread load requests across available servers. One fairly large change I ended up making is folding the 2nd priority for partial loads (stale canceled load targets) into the 1st priority since they are now fresh load eligible right away and are candidates to take the load in the initial set of servers, so keeping a separate step for them would be duplicative. #### Release note can be listed under improvements in vsf / partial load support if such a section exists. Mentioning improved loading algorithm to disperse requests across historical tiers or something in that direction <hr> ##### Key changed/added classes in this PR * `StrategicSegmentAssigner` <hr> <!-- Check the items by putting "x" in the brackets for the done things. Not all of these items apply to every PR. Remove the items which are not done or not relevant to the PR. None of the items from the checklist below are strictly necessary, but it would be very helpful if you at least self-review the PR. --> This PR has: - [ ] been self-reviewed. - [ ] added documentation for new or modified features or behaviors. - [ ] a release note entry in the PR description. - [ ] added Javadocs for most classes and all non-trivial methods. Linked related entities via Javadoc links. - [ ] added or updated version, license, or notice information in [licenses.yaml](https://github.com/apache/druid/blob/master/dev/license.md) - [ ] added comments explaining the "why" and the intent of the code wherever would not be obvious for an unfamiliar reader. - [ ] added unit tests or modified existing tests to cover new code paths, ensuring the threshold for [code coverage](https://github.com/apache/druid/blob/master/dev/code-review/code-coverage.md) is met. - [ ] added integration tests. - [ ] been tested in a test Druid cluster. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
