[Outages-discussion] Unpleasant discovery

George Herbert george.herbert at gmail.com
Sat Oct 27 16:13:32 EDT 2012



On Oct 27, 2012, at 12:44 PM, Alex Rubenstein <alex at corp.nac.net> wrote:

> What lead you to believe that just by being in the proverbial 'cloud', that you would have reliable service that dodges natural disasters?
> 
> When you buy a service that doesn't have transparency, don't expect transparency. 

I spend most of my time as a consultant; *I* certainly know better on the risks side, which is why I spent 36 hours beating through the location obfuscation for $CURRENTCLIENT.

It was statistically possible that everything could be there.  We knew part was.  Finding out that it all really is, is slightly annoying, even having 20+ years experience.  Not shocking, but annoying.

I posted this primarily to poke other people here who may be in clouds or SAAS users to go check.  And if appropriate, run local backups, set up standby servers in clouds physically elsewhere, get alternate DNS control, etc.

Right now would be good; Monday might be too late to get backups out or other prep work completed before the storm hits.  We started already, before we confirmed it was all there.

Likely case is low moderate disruption but the various weather systems coming together like that could lead to unpredictable intensity or duration upsides and serious power and network disruption.


George William Herbert
Sent from my iPhone



More information about the Outages-discussion mailing list