*sigh* Well, I've successfully duplicated my nagios monitoring server, which has been hosted at Bytemark for over 16 years, and now have a clone successfully running on my own infrastructure. For the last day or two, notifications and statuses all seem to have been matching between the two so I think I can cross that off the list. (I would prefer to monitor from off-site but this will do for now.)

I have also duplicated my collectd statistics collection server too, but my next job will be to log into 85 separate servers one at a time and reconfigure each one of them to send their statistics to the new server. Then I can copy the rrd archive files over from the old box to the new one and, with any luck, end up with only a reasonably small gap in the graphs. (Then I'll need to log into each of the 85 servers again to stop them sending to the old statistics server!) (Yes, I know there are tools that automate changes like that, but I can't say I trust that kind of automation.)

Then I think the only other things on my Bytemark footprint are (1) several domains still using one of the VMs as a nameserver, and (2) just about everything still using one of the VMs as an NTP server. At least NTP is just a matter of changing DNS and then poking each machine's NTP daemon until it notices the change. The nameserver bit involves changing the notify/transfer topology and adjusting delegations to point to my little DNS CDN, but I think there are only a few dozen domains left to do, so it shouldn't take forever. (Yep, just checked, just 43 domains still secondarying to that server, shouldn't take too long.)

I have just over a week until the smoking remains of Bytemark turn off their VM hosting servers forever. No pressure, then. And they used to be so good.