Uploaded image for project: 'OpenDJ'
  1. OpenDJ
  2. OPENDJ-7341

Changes made after disk space recovery might not be sync'd to other nodes

    Details

    • Type: Bug
    • Status: Dev backlog
    • Priority: Major
    • Resolution: Unresolved
    • Affects Version/s: 5.5.0
    • Fix Version/s: None
    • Component/s: replication
    • Labels:
    • Support Ticket IDs:

      Description

      To reproduce this issue, set up a two node replication.

      On node 1:

      1. Drop free disk space to below disk-low-threshold but above disk-full-threshold
      2. Increase free disk space to above disk-low-threshold. Node 1 reports error "Could not restart the Replication Server, bind to listen port".
      3. Drop free disk space to below disk-low-threshold but above disk-full-threshold
      4. Drop free disk space to below disk-full-threshold. Node 1 DS would connect to node 2 RS.
        Node 1 reports:
        [07/Jul/2020:15:39:42 +0800] category=SYNC severity=ERROR msgID=73 msg=An unexpected error occurred when searching for generation id for domain "dc=example,dc=com": Unwilling to Perform There are insufficient resources to perform the operation
        [07/Jul/2020:15:39:42 +0800] category=SYNC severity=NOTICE msgID=167 msg=Warning : domain dc=example,dc=com fractional replication configuration is inconsistent with backend data set : need resynchronization or fractional configuration to be changed
        

        Node 2 reports:

        [07/Jul/2020:15:39:42 +0800] category=SYNC severity=NOTICE msgID=131 msg=Directory server DS(11997) for domain "dc=example,dc=com" has changed its status to Bad generation id
        
      1. Increase free disk space to below disk-low-threshold but above disk-full-threshold.
      2. Increase free disk space to above disk-low-threshold. Node 1 RS would report error "Could not restart the Replication Server, bind to listen port". Node 1 DS is still connected to Node 2 RS.
      3. Create a new entry or delete an entry on node 1. The operational is successful. However, node 2 RS reports error
        [07/Jul/2020:15:42:35 +0800] category=SYNC severity=WARNING msgID=135 msg=Replication server RS(18124) ignoring update 000001732839fa0d2edd00000011 for domain "dc=example,dc=com" from directory server DS(11997) at opendj.example.com/192.168.56.2:35432 because its generation ID 8970 is different to the local generation ID 8970
        

      The change is not sync'd to node 2 DS.

        Attachments

          Activity

            People

            • Assignee:
              Unassigned
              Reporter:
              yinyan.cao Yinyan Cao
            • Votes:
              0 Vote for this issue
              Watchers:
              2 Start watching this issue

              Dates

              • Created:
                Updated: