[Outages-discussion] [outages] TWC outage

[Jared Mauch]

> On Aug 27, 2014, at 9:37 AM, Warren Kumari <warren at kumari.net> wrote:
> 
> On Wed, Aug 27, 2014 at 9:24 AM, Mitch Patterson <mitpatterson at gmail.com> wrote:
>> Not sure how much I buy that. Must have been major routine maitiance to have
>> that wide of an affect
> 
> Not necessarily -- screwing things up royally from a single device is
> not impossible / hard.
> 
> For example, that classic example of this is managing to redistribute
> BGP into your IGP. Then there is the "I'll just change the policy on
> this one route reflector and, oh goodness me, I just set local-pref to
> 1 on all prefixes from $some_peer. That can't be good..." game. There
> is also the "This control-plane policer ACL is ugly. I'll refactor it
> to make it prettier.... 'k, I need port 22 for SSH, and 161 for SNMP.
> What's this random port 179? Must be a typo, I'll move it lower in the
> filter..." option.
> There are also all sorts of vendor bugs you can tickle, etc.

I know I’ve caused outages by my human aspect/nature.  If this was a human
caused outage (which I suspect), I would feel horrible reading about my job
on CNN or getting the breaking news alert (USA Today made my phone ding this
morning from this).

- Jared