It's just not our week, I guess. For the past couple of days, gabe.freedesktop.org (www.fd.o, ns1.fd.o, lists.fd.o) has been playing up: it will just mysteriously die with no explanation. It appears to still be alive, but we can't access it at all. Bringing it back up yields nothing suspicious before it dies, so we're suspecting some bad hardware. This wouldn't be too major a problem, bar the loss of the above-named services and dri.fd.o, openclipart.org, and a few other services, if it weren't for other things conspiring against us.

All revision control services (git/CVS/SVN committing and checkout, web access to same), Bugzilla, xorg.fd.o, people.fd.o and planet.fd.o, are on separate machines which are unaffected by this failure. But, unfortunately, we got less than 24 hours notice that our secondary DNS server at MIT was being moved, and would be offline for a while. Two DNS servers gone. Our third DNS server, hosted in the UK, is still servicing requests just fine, but apparently our registrar never noticed the update. So the third DNS server according to the glue records is home.keithp.com, which is also off the air right now for some reason.

So, ns0.bluelinux.co.uk is still serving names, and you can still access some services if you use that, but we're not sure how long until normal DNS is back up, let alone how long it will take us to get gabe fully fixed.

Sorry for the outage, please bear with us. If you think it's frustrating not being able to check www.fd.o, try checking the remote console and watching /proc/meminfo hopelessly spew into the serial console as you are unable to log in ...