Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-94057

Don't look for chunks to balance on draining shards more than once

    • Type: Icon: Improvement Improvement
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • 8.1.0-rc0
    • Affects Version/s: None
    • Component/s: None
    • None
    • Catalog and Routing
    • Fully Compatible
    • CAR Team 2024-09-16

      When the balancer looks for chunks to be moved, it first selects a sharded collection (picked randomly) and then we pick the chunks to be moved for that collection following the next prioritization:

      • A. First, we look for chunks that must be moved from draining shards.
      • B. Then, we look for chunks that are violating a zone range.

      • C. Finally, we select chunks to distribute the collection across the cluster based on data size.

      When there are draining shards but there isn't any suitable chunk to move from those shards, we'll reevaluate those draining shards in steps B and C, which is redundant.

      The goal of this ticket is to skip the draining shards on steps B and C as the donor shards.

            Assignee:
            silvia.surroca@mongodb.com Silvia Surroca
            Reporter:
            silvia.surroca@mongodb.com Silvia Surroca
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: