Minutes of Working Group Meeting

IfA-Manoa Facility

June 10, 2010




  1. Current Network Status
  2. On March 16 - Another power outage in Waimea took down the MKOCN switch at CFHT and, consequently, our path to the SX southern path from 19:20 to 20:18. According to Kanoa, the UPS at CFHT Waimea hasn't changed. What made the difference is that in the last four years, there had only been one power outage lasting longer than 12 minutes in Waimea, but recently, there have been a rash of long outages. Kanoa is working on getting new batteries to replace existing ones that are several years old.

    On March 24 - Starting from 7:00am, there were problems with one of the core UH Juniper routers causing spotty and intermittent -- or even complete failure of -- connectivity to some remote sites. UH-ITS eventually gave up on fixing the router and switched to a backup unit with a previous version of the OS software around 12:30pm. Recovery was made more complicated because UH was operating with reduced staffing due to HGEA administrative days off. Off-duty engineers were initially working remotely with operators and then have had to come in when the situation did not improve.

    On the Big Island, since research traffic goes through the SX south instead of Manoa, observatories were seeing problems only for commodity traffic to the Internet.

    On April 13 - The SX southern path was down between midnight and 5:00am, it turned out to be a scheduled CENIC maintenance about which UH was not notified. All Big Island research traffic was rerouted through Oahu. UH-ITS has made a request to CENIC to be notified of such events in the future.

    On April 28 - Due to AC failure at the Spencer Beach landing site, the SX Juniper m7i was down between 7:30pm and 9:30pm. All Big Island traffic was rerouted through Oahu.

    On May 14 - Between 7:00am and 9:00am, all Big Island traffic was rerouted through Oahu due to a scheduled maintenance of the SX south.

  3. FY 10/11 Budget
  4. There is little change from previous year. The rate per observatory will be $8,002 vs $7,846 last year. The only increases are in the RCUH salary and the Cisco maintenance contract. We will be adding two additional CAT3750s to our Cisco contract, one for redundancy in Hilo and one at the Summit for additional ports.

  5. Brief Updates of various Projects
  6. On network performance: Pui Hin reported that transfer rates within Hawaii has been around 500Mbps or better. While running some iperf tests, Pui Hin found that the interface counters of the MKOCN cacti graphs were not properly updated at high bandwidth. The problem was subsequently fixed by Gavin Seo. It turned out that the software uses 32bit counters by default, any values over 110 Mbit were truncated. Switching to 64bit counters solved the problem.

    Hilo fiber project: Just a brief mention that aside from CSO, all Hilo facilities have moved to the new fibers. With Simon Chan and Ruisheng Peng working on it, Pui Hin is hopeful that the fiber panel in the handhole will be obsoleted by the next meeting.

    Hilo Redundancy: On June 8, the additional Cat3750, referred to as HURP-Summit3750 was successfully inserted in the MKOCN between the Juniper m7i and the Summit.

    In the next few weeks Miranda and Pui Hin will be scheduling the move of the Cisco 7206VXR NPE-G1 (HURP7206) and the CFHT Waimea connection through the HURP-Summit3750. Only CSO will be affected in the first move. For the move of the CFHT connection, CFHT Waimea will be down for 10 to 15 minutes; traffic for the rest of the observatories will be rerouted through Oahu.

    Finally, VRRP (Virtual Router Redundancy Protocol) needs to be configured to automatically switch the default route of the observatories' traffic to the HURP7206 should the Juniper m7i becomes unavailable.

  7. New participant to the MKOCN - update
  8. Pui Hin is in discussion with UH Hilo on the connection of the Hoku Ke`a, UHH 36" telescope, to the MKOCN.

    Pui Hin also reported continuous interest on the part of the US Naval Observatory to fund fiber connection to the VLBA on Mauna Kea.

  9. Other Business
  10. In response to Simon Chan's question, Pui Hin explained that due to the huge volume of over 10TB per night, it will be impossible for the data to be generated by the PanSTARRS array, when it is built on Mauna Kea, to share the common GE, at least not at its current capacity.

  11. Next meeting
  12. September 9th, Thursday, 10:00am at the SMA Hilo building.

MKOCN Homepage Back to Minutes

Last updated September 13, 2010 by by Miranda Hawarden-Ogata