Thank you for your response.This is why, even though I have a legit copy of XP, I'm going to use a hacked DRM free copy of TinyXP and Ubuntu.
I’m sorry to inform you that the Windows Genuine server might be down for few days. I have escalate the issue to our Genuine team, kindly try to validate again on Tuesday 28 Aug 2007.
Thank you for contacting Microsoft Technical Support.
Burhanistan: It takes MS more than a day to restore their own servers. One would think a company with that much liquid green on hand would have every single piece of its server infrastructure mirrored in several places around the country. Bad PR. In other news, Server 2008 is kind of lame.I find it astonishing that this did happen, as well. I used to work at MSN (which likely hosts the Windows Advantage Servers although I don't know for sure), and there were groups, including my own, that knew how to keep a site up. Geolocating your datacenters, 3rd party caching and proxying systems like Akamai (and load-balancing those so you can switch between if an Akamai itself fails), rapid detection of failure and automatic repairing without human intervention. Hell, we could and did provision an entire multi-thousand server datacenter from scratch in one day: catch the 6:30a flight from Seattle to the Bay Area, with a crew rolling in the few hundred pre-made racks and bolting them to the floor and connecting the uplink cables as we entered the datacenter a little before 10am. Plug my laptop into one server to put one copy of the bootstrap + full application suite onto a server and then sit back, watching lights flicker on across the servers as it self-replicates to the other few thousand servers. Grab lunch down the street, then catch the 5pm flight home with all your servers running, crawling the web, and happy.
So... having worked at MSN, 3 is the likely choice; they either didn't anticipate a type of failure, or don't know how to recover quickly. Lord knows that Windows would have paid plenty of money if needed for the infrastructure to ensure this didn't happen; that they didn't means that someone in operations didn't ask for the funding. Some operational groups were not the best of the best, and not everyone was, well... me. :)
- The architecture of the servers has single points of failure somewhere, and is poorly designed
- There's something suspicious going on, resulting in this outage being related to something sinister (the "How can they know it'll be up Tuesday?" question)
- The systems engineers working on that operations group are inept and didn't think of how things could fail, how to design around failure cases, and then to have an emergency plan bringing the system completely online in N minutes/hours, as needed.
« Older The hot new PC game "Bioshock"... | The Official Berkeley Breathed... Newer »
This thread has been archived and is closed to new comments
posted by [expletive deleted] at 4:16 PM on August 25, 2007