If you've noticed slower than normal download speeds to FSMG over the past few weeks, there was a fault on one of the paths used by our upstream provider that affected a small subset of our traffic
This path has been isolated while it is being repaired which has fixed the issue
You may have had issues reaching FSMG in the past hour.
We had some firewall rules unintentionally start dropping some internal traffic which confused our anycast BGP system.
This caused our service to be unavailable in small bursts between 9pm and 10pm
A big thanks to @MetanameNZ for helping us quickly fix the issue by rolling back our DS records to use our original KSK. As well as helping us fix the root problem so this won't occur again.
We got in a state where we had both our registrar's DNSSEC infra as well as our own enabled
A KSK rollover occurred on one set of infra which got published as the DS record for https://t.co/OjyBaanSJV
Since our zone was signed with a different KSK, DNSSEC validation would fail
This afternoon at 2:00pm UTC+12 we will be taking our Wellington node out of service for 1.5 hours to upgrade it to the latest stable version of Debian.
We don't expect any outage as our Hamilton node will take over serving content during this window.
In an hour from now we will be performing maintenance on our Wellington node to upgrade it to 10gig.
Traffic should be automatically rerouted to our Hamilton node so we don't expect an outage.
@ss2342 It sure does, if you are on the REANNZ network you can access the node over their 100gig national backbone, otherwise you will be limited to the speed your provider peers with REANNZ at (normally 10gig at every IX)
This afternoon at 1:00pm UTC+12 we will be taking our Hamilton node out of service for an hour to upgrade it to the latest stable version of Debian.
We don't expect any outage as our Wellington node will take over serving content during this window.