Uploaded image for project: 'OpenShift SDN'
  1. OpenShift SDN
  2. SDN-3507

LOGS: Revisit the logging in OVNK code {PostFeatureFreeze: Approved}


    • Icon: Story Story
    • Resolution: Done
    • Icon: Critical Critical
    • openshift-4.14
    • None
    • None
    • None
    • SDN Sprint 239, SDN Sprint 240, SDN Sprint 241, SDN Sprint 242
    • 0
    • 0.0

      1. We need to be strict about what's at level 5 and what's below that, adding too much logs can make it difficult to catch the important logs while adding too less logs can make it hard to debug. There needs to be a balance.
      2. nbctl logs were smaller than the libovsdb transact logs, as a result in real deployments we see log rotation happening very fast and there is a limit to number of log files stored on nodes from a specific pod. -> we need to make an effort to reduce the log size as much as possible.


      One idea that @kyrtapz had was to remove all nil fields in the transact and configure logs, since they anyways don't provide any useful info and they can cut down the size of a single transact line.

      See https://github.com/ovn-org/ovn-kubernetes/issues/3183 

            ffernand@redhat.com Flavio Fernandes (Inactive)
            sseethar Surya Seetharaman
            0 Vote for this issue
            4 Start watching this issue