Uploaded image for project: 'Indy Node'
  1. Indy Node
  2. INDY-334

Unable to add a NYM after node count came back up after being below consensus

    XMLWordPrintable

    Details

    • Type: Bug
    • Status: Complete
    • Priority: Highest
    • Resolution: Done
    • Affects Version/s: None
    • Fix Version/s: None
    • Component/s: None
    • Labels:
    • Sprint:
      H3, H4, H5

      Description

      I have a pool of 7 nodes so I can lose 2 nodes without any issues. Once I lose 3 nodes the pool can no longer reach consensus. Once I had 5 active nodes again the pool still could not reach consensus.

      Setup
      Have a 7 node pool

      Steps

      1. Send a few transactions on a 7 node pool to make sure everything is running.
      2. Stop the sovrin-node service on node 7 (6 out of 7 active)
      3. Send a transaction to make sure the pool can still reach consensus (It should work)
      4. Stop the sovrin-node service on node 6 (5 out of 7 active)
      5. Send a transaction to make sure the pool can still reach consensus (It should work)
      6. Stop the sovrin-node service on node 5 (4 out of 7 active)
      7. Send a transaction to make sure the pool can still reach consensus (It should not work)

      At this point the pool should not be able to reach consensus
      Start the sovrin-node service on node 5 (5 out of 7 active)
      Now I have 5 out of 7 active nodes which is enough for consensus as we can see above.
      Send a transaction to make sure the pool can still reach consensus (It should work)

      Issue
      Once I returned to have enough nodes to reach consensus the pool was still unable to reach consensus.
      Send a transaction to make sure the pool can still reach consensus (It should work)

      I then restarted both nodes 6 and 7 so I had all nodes active in the pool again.
      Send a transaction to make sure the pool can still reach consensus (It should work)
      The pool still could not reach consensus.

      I disconnected the CLI and reconnected to the pool.
      Send a transaction to make sure the pool can still reach consensus (It should work)
      The pool still could not reach consensus.

      Workaround
      In the end I had to stop all nodes in the pool and start them up one at a time and then the pool was functional again.

        Attachments

        1. Node1.log
          2.28 MB
        2. node1.txt
          26 kB
        3. Node2.log
          2.28 MB
        4. node2.txt
          157 kB
        5. Node3.log
          2.28 MB
        6. node3.txt
          153 kB
        7. Node4.log
          263 kB
        8. node4.txt
          153 kB
        9. Screenshot.PNG
          Screenshot.PNG
          155 kB
        10. Screenshot.PNG
          Screenshot.PNG
          156 kB

          Issue Links

            Activity

              People

              Assignee:
              danielhardman Daniel Hardman
              Reporter:
              krw910 Kelly Wilson
              Watchers:
              Alexander Shcherbakov, Kelly Wilson, Lovesh Harchandani, Victor Muzychenko, Vladimir Shishkin
              Votes:
              0 Vote for this issue
              Watchers:
              5 Start watching this issue

                Dates

                Created:
                Updated:
                Resolved: