[update – 12:58, 27 Dec 2017] Only the L47 top-of-rack switchgroup is down. All other groups have been patched and successfully rebooted.
[update – 12:07, 27 Dec 2017] All router maintenance is complete. Top-of-rack switches are still being rebooted.
—
We have scheduled a major maintenance/upgrade for our SDSC core network on Wednesday, December 27th, starting at 10am. During this window we will be completely upgrading our core Juniper routers with new hardware, and updating code on all of our Juniper top-of-rack switches. Depending on your connection to our networking, you will see varying levels of network disturbance during the window.
For the router upgrades:
Beginning at 1000, we will be adjusting all routing of 132.249.0.0/16 and 198.202.64.0/18 to flow through a single device. You are likely to see a brief loss of traffic during this portion of the window, while routing re-converges and sessions re-establish. Once our first chassis is completely upgraded, we will have another adjustment window to move all traffic to the new setup. This is targeted for 1300 on Wednesday, but could move up if everything goes smoothly. Once the other chassis is fully upgraded, we will have one final window of routing adjustment, to return to normal services. This will be targeted for 1600, but could also float within the window.
For the Top-of-Rack (TOR) updates:
Beginning at 1000, we will be rebooting each group of chassis units, one at a time – roughly every 15 minutes, or as the previous chassis stabilizes. We will ensure that no two chassis groups are down at one time. This will result in an outage between 5 and 15 minutes for each group. This outage will effect ALL 1GbE connections, as well as a handful of 10GbE connections that connect to Juniper devices.
Customer Notes:
1- If you are connected to any of our Arista switches, you will not see the TOR outages.
2- If you are a UCSD customer, you will only see the TOR outages.