Uploaded image for project: 'OpenShift Bugs'
  1. OpenShift Bugs
  2. OCPBUGS-42659

Scaling up to 450 UDNs takes ~9hours

XMLWordPrintable

    • None
    • False
    • Hide

      None

      Show
      None

      Description of problem:
      While scaling up to 450 UDNs i.e 450 namespaces each with its own UDN takes a long time. The time is mostly spent to add logical routes and logical switches to each node. To get to 450 UDNs it takes about 9h. For baseline it takes 10mins to scale up to 450 namespaces without any UDNs

      Version-Release number of selected component (if applicable):

      4.18.0-0.nightly-2024-09-30-051028

      How reproducible:

      Always

      Steps to Reproduce:

      1. Run a client-server workload with 10 pods per namespace using kube-burner at 20 QPS

      2. Deploy with 450 namespaces each with its own UDN

      3. Check how much time it takes for all the pods to get primary network ip (UDN)

      Actual results:

      Creation of UDNs one after another takes a long time. It seems there is some wait as the creation of CRDs for all 450 namespace is done at the very start

      Expected results:

      A much quicker deployment to networks

            trozet@redhat.com Tim Rozet
            mohit-sheth Mohit Jitendra Sheth
            Anurag Saxena Anurag Saxena
            Votes:
            0 Vote for this issue
            Watchers:
            3 Start watching this issue

              Created:
              Updated: