Uploaded image for project: 'controller'
  1. controller
  2. CONTROLLER-1645

shard moved during 1M bgp prefix advertizing (with tell-based=true)

XMLWordPrintable

    • Icon: Bug Bug
    • Resolution: Unresolved
    • None
    • None
    • clustering
    • None
    • Operating System: All
      Platform: All

    • 8318

      https://jenkins.opendaylight.org/sandbox/job/bgpcep-csit-3node-periodic-bgpclustering-only-carbon/8/

      After the bgp peer is connected, 1M prefixes are being advertized. From the log.html we see that updates for 1M prefixes were sent between

      017-04-27 07:36:16,151 INFO BGP-Dummy-1: Common single UPDATE will be generated for both NLRI & WITHDRAWN lists
      2017-04-27 07:36:16,151 INFO BGP-Dummy-1: Let's go ...
      2017-04-27 07:36:16,160 INFO BGP-Dummy-1: Iteration: 0 - total remaining prefixes: 1000000
      ...
      2017-04-27 07:37:18,545 INFO BGP-Dummy-1: Iteration: 999000 - total remaining prefixes: 1000
      2017-04-27 07:37:18,603 INFO BGP-Dummy-1: All update messages generated.
      2017-04-27 07:37:18,603 INFO BGP-Dummy-1: Storing performance results.
      2017-04-27 07:37:18,603 INFO BGP-Dummy-1: ########## Final results ##########
      2017-04-27 07:37:18,603 INFO BGP-Dummy-1: Number of iterations: 1000000

      During this time there was a shard leader leader movement

      https://logs.opendaylight.org/sandbox/jenkins091/bgpcep-csit-3node-periodic-bgpclustering-only-carbon/8/archives/log.html.gz#s1-t7-k2-k3-k7-k3-k1-k6-k1-k1-k1-k1-k1-k2-k1-k1-k2-k2-k2-k1-k6-k2-k2-k1

      Finally only 26k prefixes are present in odl.

      https://logs.opendaylight.org/sandbox/jenkins091/bgpcep-csit-3node-periodic-bgpclustering-only-carbon/8/archives/log.html.gz#s1-t10-k2-k2

            Unassigned Unassigned
            pgubka@cisco.com Peter Gubka
            Votes:
            0 Vote for this issue
            Watchers:
            4 Start watching this issue

              Created:
              Updated: