Uploaded image for project: 'Kudu'
  1. Kudu
  2. KUDU-520

When a single replica gets rebooted it never rejoins the quorum

    XMLWordPrintableJSON

Details

    • Bug
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • M4
    • None
    • consensus, tserver
    • None

    Description

      Replicas expect a change config message to get out CONFIGURING state. When the whole cluster goes down leaders will push one at the start. However when a single replica goes down and the remainder of the quorum kept going it's unlikely that the first message will be a config change, thus the replica never leaves the CONFIGURING state.

      On bootstrap, when receiving the first message from the leader, replicas should check that the log matching property is uphelp and that the term is the same and if so they should just accept the message and move on.

      Attachments

        Issue Links

          Activity

            People

              dralves David Alves
              dralves David Alves
              Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: