Category: Operations

Posted on

New and Improved! (and downtime)

Today, the cable folks came by and upgraded all of our networking interface to new and faster devices. Unfortunately, it also meant I had to reconfigure all of the machines involved in the weather station, as they depend on static addresses to find each other. Laptops and cell phones and PCs that take whatever IP address they are given were up and running as soon as they were done (nearly three hours without internet! Can you imagine!) but the weather devices took a few hours: resetting DHCP reservations, a new router configuration to learn, digging up obscure instructions for some small devices, and tweak on settings. We were down, not posting updates to the website, from 1:40 pm to 7:55 pm, but we seem to be up and running now.

6-hour gap visible in the weather reports for 18-August-2026
6-hour gap visible in the weather reports for 18-August-2026
Posted on

Downtime

Several of the weather reporting systems we have running broke down at 5:15 PM yesterday. While setting up fans in the windows after a long hot day, I suspect I bumped the GW1100 gateway device and cause it to malfunction. The gateway gets signals from all the local sensors, and makes that data available on the LAN. From there, a Raspberry Pi reads the data, formats it into several pretty formats, and forwards it on the website where you read it, and also forwards the data on to other online sites.

The gateway was in a quirky state neither on nor off, and the Raspberry Pi locked up waiting for it complete a conversation with gateway.

I didn’t notice the problem until this morning and after a lot of running around getting diagnostic info, turning things off and back on again (still the #1 solution to IT problems!) and jiggling a lot of wires, it appears we’re back up and running.

Posted on

Offline yesterday

Alert weather watchers observed the websites (https://wx.iayft.com, https://contoocook.org and https://contoocook.org/cell.php) were offline Sunday morning, July 27th. My internet hosting provider (Linode/Akamai) had a long-term site power outage and troubled recovery of their HVAC systems resulting in a long, slow, manual restoration. The were good about keeping their users informed. I expect many lessons were learned and I’m looking forward to the follow-up to the incident. No data was lost, and the sites are back up and running now. If anyone notices anomalies, let me know using the “Contact” form linked above.