It’s time for us to arrange to reboot most of our network switches again. We need to do this for a couple of reasons.
The first is that the clocks that they use for management and reporting purposes tick in centiseconds. That means that they wrap at just under 497 days of uptime, which in particular results in error log messages being jumbled.
The second reason is to apply firmware updates. We apply important updates as soon as we can, of course, but this “yearly” reboot allows us to remain using a (reasonably) current version, which can be important when talking to the manufacturer or swapping individual components.
We had hoped that we could take advantage of the forthcoming electrical work in the Forum and Bayes to do these reboots for us, but unfortunately the likely timing of this now means that we can’t wait so we’ll have to schedule reboots sooner. For Appleton Tower we usually reboot each floor’s switches first thing in the morning over a couple of days, avoiding major deadlines.
We also have six core switches, for which we always schedule reboots in advance. These are normally done over a series of lunch-times, as the manufacturer advises that physical intervention may be required in a small number of cases.