1. Tom Lehman

Tom Lehman

Forum Replies Created

Viewing 12 posts - 1 through 12 (of 12 total)
  • Author
    Posts
  • in reply to: Infrastructure-metrics queries #7867
    Tom Lehman
    Participant

      I think the main consideration is that HC is based on 64 bit counters, and the non HC is based on 32 bit counters.  So for a high speed interface a 32 bit counter may roll over quite often.  Using a 64 bit counter is more convenient in terms of not having to monitor for counter roll overs.  The data in terms of packet counts should be equivalent.

      in reply to: FABRIC UCSD dataplane link is down #7560
      Tom Lehman
      Participant

        The fiber has been repaired in Los Angeles, and the FABRIC UCSD dataplane link is now back in service.

        This topic will be closed.

        in reply to: FABRIC CERN Dataplane Network Maintenance/Upgrade #7518
        Tom Lehman
        Participant

          The FABRIC CERN dataplane link upgrades are complete. This issues will be closed.

          in reply to: FABRIC MICH dataplane link is down #7466
          Tom Lehman
          Participant

            The FABRIC MICH dataplane is back in service.  This topic will be closed.

            in reply to: FABRIC RUTG Dataplane link is currently down #7457
            Tom Lehman
            Participant

              closing this post, as FABRIC RUTG dataplane is back in service.

              in reply to: FABRIC RUTG Dataplane link is currently down #7447
              Tom Lehman
              Participant

                The FABRIC RUTG dataplane connection is back in service.

                Network dataplane services from FABRIC RUTG to other sites should now work.

                Tom Lehman
                Participant

                  Hi Sunjay,

                  It should be working now, we had to make an adjustment on a BGP peering with an upstream provider.

                  Please check the external connectivity now and let us know if all is ok for your slices.

                  Thanks,

                  Tom

                  in reply to: FABRIC MASS Dataplane is currently link down #7327
                  Tom Lehman
                  Participant

                    Closing this issue since the FABRIC MASS dataplane link is back in service

                    in reply to: FABRIC MASS Dataplane is currently link down #7326
                    Tom Lehman
                    Participant

                      Maintenance work has been completed at MGHPCC and the FABRIC MASS dataplane link is back in service.

                      Tom Lehman
                      Participant

                        Leigh, I do not see VLAN 2591 on Utah, or 3874 on Clemson.  Do you think that FABRIC slice is still up?

                        Tom

                        in reply to: IPv6 on FABRIC: A hop with a low MTU #4168
                        Tom Lehman
                        Participant

                          @yoursunny Thanks for the nice testing table/results.  We do have some links in the network which need to be set higher to allow hosts to use MTU of 9000.   We need to do some more testing for all the links in the network to see if we can find a single value that works everywhere.  We are also adding many new wide area links over the next month.   So we will probably update the MTU settings across the network sometime in June.   We will report back here on the forum with more details once we have made this change.   For now, hopefully you can use MTU 8948 as you show in your table, until we make the change.

                          The MASS site/links should now support a host MTU of 8948.  Please let us know if you still have issues with MASS node.

                          in reply to: FABRIC Dataplane Network Topology Updates #1604
                          Tom Lehman
                          Participant

                            This work to add the new dataplane links to the FABRIC network topology is complete.

                          Viewing 12 posts - 1 through 12 (of 12 total)