Uploaded image for project: 'Couchbase Kubernetes'
  1. Couchbase Kubernetes
  2. K8S-3463

Delta-Recovery Upgrade - Orchestrator swap can cause rebalance to exit

    XMLWordPrintable

Details

    • Bug
    • Resolution: Done
    • Major
    • 2.6.4
    • None
    • None
    • None
    • 9 - Cataclysm
    • Critical
    • 2

    Description

      When we are performing an upgrade, the first node to upgrade will be ceded the orchestrator. If there is a rebalance occurring during this transfer, it will stop the rebalance, causing delta-recovery balances fail.

      We should use the /cluster-info/ to query which node is the orchestrator, we should verify that this node IS on the upgrade version prior to beginning the first delta-recovery rebalance. Otherwise this will always fail.

      We should add a 1 minute timeout, if we do not see the orchestrator on the latest version, we should begin the rebalance.

      Attachments

        Issue Links

          relates to

          Page - A new page either standalone or within a series of pages must be added K8S-3446 Cloud Native 2.6.4 - Release Notes

          • Major - Major loss of function.
          • Closed
          For Gerrit Dashboard: K8S-3463
          # Subject Branch Project Status CR V
          209640,2 K8S-3463 Added check for orchestrator version master couchbase-operator Status: ABANDONED 0 +1
          209671,3 K8S-3463 Added check for orchestrator version 2.6.x couchbase-operator Status: MERGED +2 +1

          Activity

            People

              yusuf.ramzan Yusuf Ramzan
              justin.ashworth Justin Ashworth
              Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved:

                Gerrit Reviews

                  There are no open Gerrit changes

                  PagerDuty