Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-88100

Encode UUID of a collection and nss differently when computing QueryShapeHash for "find" command

    • Type: Icon: Task Task
    • Resolution: Fixed
    • Priority: Icon: Major - P3 Major - P3
    • 8.0.0-rc0
    • Affects Version/s: None
    • Component/s: None
    • Query Execution
    • Fully Compatible
    • QE 2024-04-01

      Currently UUID and nss information is encoded without any discriminator, meaning we do not differentiate between UUID and nss when preparing the buffer for QueryShapeHash computation.

      This may result in collisions when collection UUID and nss result in the same byte sequence.

      To solve this issue we can append additional bit 0/1 for UUID or Nss encoding, to avoid having any collisions

            Assignee:
            mindaugas.malinauskas@mongodb.com Mindaugas Malinauskas
            Reporter:
            denis.grebennicov@mongodb.com Denis Grebennicov
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

              Created:
              Updated:
              Resolved: