Uploaded image for project: 'MariaDB ColumnStore'
  1. MariaDB ColumnStore
  2. MCOL-4292

columnstore-docker-cluster: ColumnStore engine not accessible during restart of a cluster member in HA mode

    XMLWordPrintable

Details

    • Bug
    • Status: Closed (View Workflow)
    • Major
    • Resolution: Not a Bug
    • None
    • N/A
    • N/A
    • mariadb/skysql-enterprise-columnstore-handoff sky-1.5.4-1 95e34753d40f

    Description

      Don't know if this is a bug or expected behaviour.

      Scenario: A multi node ColumnStore cluster with 3 nodes and shared metadata storage (activated HA feature)

      In case one cluster member is restarted read and write queries to the ColumnStore engine aren't possible for less than a minute. Once the restarted node is back up again reads and writes are possible again.

      How to reproduce:

      jens@ubuntu18:~/columnstore-docker-cluster$ docker-compose up -d
      Creating network "columnstore-docker-cluster_default" with the default driver
      Creating volume "columnstore-docker-cluster_data1" with default driver
      Creating volume "columnstore-docker-cluster_data2" with default driver
      Creating volume "columnstore-docker-cluster_data3" with default driver
      Creating volume "columnstore-docker-cluster_storagemanager" with default driver
      Creating mcs3 ... done
      Creating mcs2 ... done
      Creating mcs1 ... done
      Creating mx1  ... done
      

      jens@ubuntu18:~/columnstore-docker-cluster$ docker exec -it mcs1 demo
      Waiting for PM1 to be initialized ............... done
      Adding PM1 to cluster ... done
      Adding PM2 to cluster ... done
      Adding PM3 to cluster ... done
      Validating ... done
      Adding PM1 to MaxScale ... done
      Adding PM2 to MaxScale ... done
      Adding PM3 to MaxScale ... done
      Adding service ... done
      Adding listener ... done
      Adding monitor ... done
      

      jens@ubuntu18:~/columnstore-docker-cluster$ docker exec -it mcs1 mariadb -e "CREATE DATABASE test; CREATE TABLE test.tmp1 (i int) engine=columnstore; insert into test.tmp1 values (42); select * FROM test.tmp1"
      +------+
      | i    |
      +------+
      |   42 |
      +------+
      

      jens@ubuntu18:~/columnstore-docker-cluster$ docker restart mcs2; for i in $(seq 1 10); do echo $i; docker exec -it mcs1 mariadb -e "SELECT * FROM test.t
      mp1"; sleep 1; done
      mcs2
      1
      ERROR 1815 (HY000) at line 1: Internal error: DistributedEngineComm::write: Broken Pipe error
      2
      ERROR 1815 (HY000) at line 1: Internal error: DistributedEngineComm::write: Broken Pipe error
      3
      ERROR 1815 (HY000) at line 1: Internal error: IDB-2039: Data file does not exist, please contact your system administrator for more information.
      4
      ERROR 1815 (HY000) at line 1: Internal error: IDB-2039: Data file does not exist, please contact your system administrator for more information.
      5
      ERROR 1815 (HY000) at line 1: Internal error: IDB-2039: Data file does not exist, please contact your system administrator for more information.
      6
      ERROR 1815 (HY000) at line 1: Internal error: IDB-2045: At least one PrimProc closed the connection unexpectedly.
      7
      +------+
      | i    |
      +------+
      |   42 |
      +------+
      8
      +------+
      | i    |
      +------+
      |   42 |
      +------+
      9
      +------+
      | i    |
      +------+
      |   42 |
      +------+
      10
      +------+
      | i    |
      +------+
      |   42 |
      +------+
      

      Attachments

        Activity

          People

            toddstoffel Todd Stoffel (Inactive)
            jens.rowekamp Jens Röwekamp (Inactive)
            Votes:
            0 Vote for this issue
            Watchers:
            2 Start watching this issue

            Dates

              Created:
              Updated:
              Resolved:

              Git Integration

                Error rendering 'com.xiplink.jira.git.jira_git_plugin:git-issue-webpanel'. Please contact your Jira administrators.