6 Feb 2017
18:06 UTC
Tom Hill
Unfortunately nothing has come to light from any of the debug information post reboot, so we'll need to defer to Cisco which will take some time.
Similarly we're not able to find any documented bugs for this hardware that would meet the criteria, or indeed resolved caveats in the Cisco release notes.
As it happens this router is also due to be moved to another rack in the coming weeks, so we will as a matter of course apply the lastest software release just in case.
Further debugging with the downstream customer in question will take place with our backup router, in the case that this is a repeatable problem, such that it does not cause any outage at all if it goes down.
6 Feb 2017
17:02 UTC
Tom Hill
One of our two gateway routers in York, ar1.dc1.yo26.yrk, spontaneously rebooted at around 16:53 UTC+0. This coincided roughly with a new BGP client flapping their session, and is likely to be an unhandled exception.
The backup router took over all duties as it should, so we seem to have been OK, but this event will have caused a very short pause in connectivity for traffic travelling between York & Manchester.
The root cause of the failure is as yet unidentified, but we will be analysing the data from the router as best we can, and will update when we have more information.