HBase
  1. HBase
  2. HBASE-10300

Insure a throw of DoNotRetryIOException when a regionserver aborts

    Details

    • Type: Bug Bug
    • Status: Resolved
    • Priority: Major Major
    • Resolution: Invalid
    • Affects Version/s: 0.98.0, 0.99.0
    • Fix Version/s: None
    • Component/s: None
    • Labels:
      None

      Description

      As discussed on HBASE-10292, we may not be throwing DoNotRetryIOExceptions back to the client when aborting the server, especially when handling fatal coprocessor exceptions.

        Activity

        Hide
        Devaraj Das added a comment -

        My understanding was that DoNotRetryIOException is thrown for cases where there is no point retrying at a global level (for e.g., a table is disabled and the client shouldn't try to get data from its regions). When a regionserver aborts, and a client was connected to it before should retry it's operations to the failed over regionserver(s), no? Why should it be a DoNotRetryIOException..

        Show
        Devaraj Das added a comment - My understanding was that DoNotRetryIOException is thrown for cases where there is no point retrying at a global level (for e.g., a table is disabled and the client shouldn't try to get data from its regions). When a regionserver aborts, and a client was connected to it before should retry it's operations to the failed over regionserver(s), no? Why should it be a DoNotRetryIOException..
        Hide
        Andrew Purtell added a comment -
        Show
        Andrew Purtell added a comment - Ping Anoop Sam John
        Hide
        Anoop Sam John added a comment -

        When I made comment in HBASE-10292 I was having this also in mind. That is why I was saying might be it is correct. Yes CP impl thrown an unexpected error (Not IOE) and it might be a temp issue with CP in that RS. So in another RS, the CP might work well. We can close this issue as of now I guess.

        Show
        Anoop Sam John added a comment - When I made comment in HBASE-10292 I was having this also in mind. That is why I was saying might be it is correct. Yes CP impl thrown an unexpected error (Not IOE) and it might be a temp issue with CP in that RS. So in another RS, the CP might work well. We can close this issue as of now I guess.
        Hide
        Andrew Purtell added a comment -

        Great

        Show
        Andrew Purtell added a comment - Great

          People

          • Assignee:
            Unassigned
            Reporter:
            Andrew Purtell
          • Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

            Dates

            • Created:
              Updated:
              Resolved:

              Development