Uploaded image for project: 'CDH (READ-ONLY)'
  1. CDH (READ-ONLY)
  2. DISTRO-685

Duplicate hostnames with differing subdomains cause problems with multiple Masters

    Details

    • Type: Bug
    • Status: Open
    • Priority: Major
    • Resolution: Unresolved
    • Affects Version/s: CDH 5.3.0
    • Fix Version/s: None
    • Component/s: HBase
    • Labels:
      None
    • Environment:
      Ubuntu 12.04 LTS

      Description

      I set up a cluster with 3 Master roles on 3 hosts. They all have the same hostname, but different subdomains.

      Example:
      hbase-1.zone1.dc1.domain.com
      hbase-1.zone2.dc1.domain.com
      hbase-1.zone3.dc1.domain.com

      When the cluster comes up, there is a yellow health alert because a standby Master could not be found. The output of this alert shows that the active Master is (Availability: Active, Health: Good), but the other two Masters are (Availability: Unknown, Health: Good). If I shut down the active Master, neither of the other two Masters take over and the cluster alerts because it has no Master.

      I was able to fix this issue by adding Master roles to:
      hbase-2.zone1.dc1.domain.com
      hbase-3.zone2.dc1.domain.com
      These both came up fine and are declared as 'Backup' in CDH. They also are (Availability: Standby, Health: Good) while the other two bad Masters are still showing unknown availability.

      This leads me to believe that there is an issue using the same hostnames even though the FQDN is different. If this is a known limitation, it would be nice if the Cloudera Manager setup either made users aware of this, or prevented adding nodes with the same hostname.

      • Cloudera Manager was used to set up the cluster with parcels and mostly default settings.

        Attachments

          Activity

            People

            • Assignee:
              Unassigned
              Reporter:
              rohan25 Rohan Soni
            • Votes:
              0 Vote for this issue
              Watchers:
              1 Start watching this issue

              Dates

              • Created:
                Updated: