Hadoop HDFS
  1. Hadoop HDFS
  2. HDFS-3714

Disallow manual failover to already active NN when auto failover is enabled

    Details

    • Type: Bug Bug
    • Status: Open
    • Priority: Minor Minor
    • Resolution: Unresolved
    • Affects Version/s: 2.0.0-alpha
    • Fix Version/s: None
    • Component/s: auto-failover
    • Labels:
      None

      Description

      If nn1 is active and nn2 is standby "hdfs haadmin -failover nn2 nn1" says that failover to nn1 is successful even though nn1 is already active. When only manual failover is active, there is a check that tells users that the user can't failover to an already active service. We should have the same check for auto failover.

      bash-4.1$ hdfs haadmin -failover nn2 nn1
      Failover to NameNode at brut01.sf.cloudera.com/172.22.35.149:17020 successful
      bash-4.1$ hdfs haadmin -failover nn2 nn1
      Failover to NameNode at brut01.sf.cloudera.com/172.22.35.149:17020 successful
      bash-4.1$ hdfs haadmin -failover nn2 nn1
      Failover to NameNode at brut01.sf.cloudera.com/172.22.35.149:17020 successful
      

        Activity

        Hide
        Eli Collins added a comment -

        Thanks to Stephen Chu for identifying this bug.

        Show
        Eli Collins added a comment - Thanks to Stephen Chu for identifying this bug.

          People

          • Assignee:
            Unassigned
            Reporter:
            Eli Collins
          • Votes:
            0 Vote for this issue
            Watchers:
            6 Start watching this issue

            Dates

            • Created:
              Updated:

              Development