Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-5050

autofailover should not auto failover anybody when rebalance is marked as in-progress. Or at least be more careful there

    XMLWordPrintable

Details

    • Bug
    • Resolution: Fixed
    • Major
    • 1.8.1
    • 1.8.0
    • ns_server
    • Security Level: Public

    Description

      In one of customers:

      • rebalance was running
      • node orchestrating rebalance failed
      • node orchestrating rebalance was automatically failed over by new master
      • node orchestrating rebalance re-joined (still thinking it was rebalancing see MB-5049)
      • we had config conflict and bad things as a result

      Looks like during rebalance we should just disable autofailover because during rebalance humans can take care of correct actions

      Attachments

        For Gerrit Dashboard: MB-5050
        # Subject Branch Project Status CR V
        14789,2 disallow automatic failover during rebalance.MB-5050 branch-181 ns_server Status: MERGED +2 +1
        14833,1 Merge remote-tracking branch 'origin/branch-181' into branch-18 branch-18 ns_server Status: MERGED +2 +1
        15066,1 Merge remote-tracking branch 'couchbase/branch-18' master ns_server Status: MERGED +2 +1

        Activity

          People

            alkondratenko Aleksey Kondratenko (Inactive)
            alkondratenko Aleksey Kondratenko (Inactive)
            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