Uploaded image for project: 'Couchbase Server'
  1. Couchbase Server
  2. MB-6209

auto-failover should not fail over a node in the cluster if there are not enough number of replicas in the cluster

    Details

    • Type: Bug
    • Status: Closed
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 2.0-beta-2
    • Component/s: ns_server
    • Security Level: Public
    • Labels:

      Description

      to repro :

      start from a imbalanced cluster wehre there are 1024 active vbuckets and less than 1024 vbuckets , then shut down the node and then you can see that autofailover fails over the node

      # Subject Project Status CR V
      For Gerrit Dashboard: &For+MB-6209=message:MB-6209

        Activity

        Hide
        farshid Farshid Ghods (Inactive) added a comment -

        reproduced this issue once on cluster_run environment with 64 vbuckets , will try on a cluster and attach diags.

        Show
        farshid Farshid Ghods (Inactive) added a comment - reproduced this issue once on cluster_run environment with 64 vbuckets , will try on a cluster and attach diags.
        Hide
        farshid Farshid Ghods (Inactive) added a comment -

        Hi Aliaksey,

        can you please merge this fix to 1.8.1 branch ?

        Show
        farshid Farshid Ghods (Inactive) added a comment - Hi Aliaksey, can you please merge this fix to 1.8.1 branch ?
        Hide
        alkondratenko Aleksey Kondratenko (Inactive) added a comment -

        Fix is in gerrit

        Show
        alkondratenko Aleksey Kondratenko (Inactive) added a comment - Fix is in gerrit
        Hide
        kzeller kzeller added a comment -

        in RN: When you try to automatically failover a node, Couchbase Server
        will not perform the failover if there are not enough
        replica nodes in the cluster to support the failed node.

        Show
        kzeller kzeller added a comment - in RN: When you try to automatically failover a node, Couchbase Server will not perform the failover if there are not enough replica nodes in the cluster to support the failed node.

          People

          • Assignee:
            alkondratenko Aleksey Kondratenko (Inactive)
            Reporter:
            farshid Farshid Ghods (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