Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-71609

`splitOrMarkJumbo` must not over-split chunks

    • Type: Icon: Bug Bug
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • 6.0.4, 6.3.0-rc0
    • Affects Version/s: None
    • Component/s: None
    • None
    • Fully Compatible
    • ALL
    • v6.2, v6.0
    • Sharding EMEA 2022-12-12, Sharding EMEA 2022-12-26

      Now that balancing is based on data size, chunks are by default of heterogeneous sizes and it is not needed anymore to over-split them (better to keep the routing table as small as possible).

      When a migration fails because of a frequent shard key, the splitOrMarkJumbo method is called and currently takes care of dividing the chunk based on the list of split points [potentially multiple].

      This method should be simplified in order to split only on the first available point because it is only needed to mark as jumbo the chunk containing the frequent shard key. This can be achieved by passing a vector only containing the first split point as argument to the splitChunkAtMultiplePoints function.

            Assignee:
            pierlauro.sciarelli@mongodb.com Pierlauro Sciarelli
            Reporter:
            pierlauro.sciarelli@mongodb.com Pierlauro Sciarelli
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: