HBase
  1. HBase
  2. HBASE-46

Robust start and stop for cluster, regionserver

    Details

    • Type: New Feature New Feature
    • Status: Closed
    • Priority: Major Major
    • Resolution: Fixed
    • Affects Version/s: None
    • Fix Version/s: 0.1.2, 0.2.0
    • Component/s: master, regionserver, scripts
    • Labels:
      None

      Description

      When a new region server is started, it should have a period during which it will retry to connect to the master, rather than just immediately shutting down when it can't contact the master.

      Also, when HBase and HDFS are started at around the same time, HBase shouldn't immediately exit. Instead, wait a little while for everything to come online.

        Issue Links

          Activity

          Hide
          Jim Kellerman added a comment -

          Right now in 0.2, the master waits for hdfs to exit safe mode and the region server waits forever or until it is killed trying to contact the master.

          Is there more we want to do for this issue (aside from fix the issues linked to this one)?

          Show
          Jim Kellerman added a comment - Right now in 0.2, the master waits for hdfs to exit safe mode and the region server waits forever or until it is killed trying to contact the master. Is there more we want to do for this issue (aside from fix the issues linked to this one)?
          Hide
          stack added a comment -

          +1 on closing this issue as no longer relevant (IIRC, way back then, the regionserver was going down because HDFS was sick but this issue is missing that description so should be closed).

          Show
          stack added a comment - +1 on closing this issue as no longer relevant (IIRC, way back then, the regionserver was going down because HDFS was sick but this issue is missing that description so should be closed).
          Hide
          Jim Kellerman added a comment -

          I should also mention that the behavior of the Master and Region Server is the same on the 0.1 branch as it is on 0.2.

          If we find more startup issues we can open a new Jira for them.

          Show
          Jim Kellerman added a comment - I should also mention that the behavior of the Master and Region Server is the same on the 0.1 branch as it is on 0.2. If we find more startup issues we can open a new Jira for them.

            People

            • Assignee:
              Jim Kellerman
              Reporter:
              Bryan Duxbury
            • Votes:
              0 Vote for this issue
              Watchers:
              0 Start watching this issue

              Dates

              • Created:
                Updated:
                Resolved:

                Development