On Tue, Dec 13, 2005 at 06:25:19PM -0500, Ben Stern wrote: > I remember the fire. Care to expand upon the "insane routing updates from > ALGX" issue? I only remember one instance when that was possible, and if I > recall correctly, it was avoided.
There was a period (several months) a while back (2002) when ALGX routing engineers would perform routing updates early on Saturday mornings. On several of those mornings (not sequential; scattered) they goofed. I could tell that they goofed because I had systems inside and outside the affected space, so just using traceroute was sufficient to observe the problems. The goofs took various forms, none of which I ever tried to diagnose in detail (because I wasn't in a position to fix them anyway) but loops and route flapping were among the symptoms. I can dig out the old email messages that I composed and sent at the time if you want more details such as dates/times, observations, etc. But it's pretty much a non-issue now: some form of corrective action has apparently taken place, whether it's more scrutiny of proposed changes before deployment or taking the "enable" password aware from someone. ;-) I just cited it as an example of a situation that didn't actually require a report from me, because the people who needed to be aware of it already were.
