Uploaded image for project: 'Core Server'
  1. Core Server
  2. SERVER-13126

Honor read preference after replica fail-over in 2.4

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major - P3
    • Resolution: Gone away
    • Affects Version/s: 2.4.9
    • Fix Version/s: None
    • Component/s: Sharding
    • Labels:
      None
    • Operating System:
      ALL

      Description

      If you read from a secondary through a mongos, and the underlying replica set has a fail-over, such that the secondary becomes a primary, mongos will still keep reading from that node, even though it is now in violation of the read preference. (Connections are pinned.)

      This behavior is fixed in 2.6, but not in the 2.4.x branch. It is suggested to backport the fix, or create a different fix in 2.4 such that read preferences are honored even in the face of replica fail-overs.

        Attachments

          Activity

            People

            Assignee:
            backlog-server-sharding Backlog - Sharding Team
            Reporter:
            andre.spiegel Andre Spiegel
            Participants:
            Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

              Dates

              Created:
              Updated:
              Resolved: