Amazon's mammoth compute cloud experienced problems on Friday, disrupting the constellation of sites that depend on it, such as Heroku. The faults in the company's Northern Virginia US-EAST-1 data center cluster were spotted by Amazon at 7:32am Pacific Time, and as of the time of writing they were still experiencing problems. …
All Flickr members were informed over 24 hours ago that there would be a maintenance period this morning - so it can hardly have been an outage at all, unless you think major sites are now planning outages to rectify faults they already know will happen ???
The site was up again at the announced time with no problems.
"All Flickr members were informed over 24 hours ago...." It would seem, with so many jih-nah-lists queuing up to slate Big Cloud, maybe Amazon needs to start a PR exercise to also notify industry journos of downtime so they don't get overly excited in future.
@ Matt Bryant
I guess that explains the comment regarding their root cause analysis and not actually stating what the actual fault was...
Pint coz its Friday night
To be fair, IT Hack, the full answer might involve commercial secrets. AWS does try to be as open as possible about outages and their causes, particularly when the information might benefit the wider community, so that's the only reason I can think of - apart from embarrassment - why they might not want to spill the details.
@ Joe Green
No...I understand that and that is a very salient point. However narrowing down the cause to something that aids the rest of us is not bad thing. I don't expect IP addresses and machine names but a general statement of whether it was power or bad code which has been rectified is not a bad thing.
Bearing in mind that we have such controversy surrounding "cloud" there has to be some level of openness when the shit hits the fan. I mean..if I am going to entrust my data to a third party I certainly want to know what went wrong at some level...not a statement that "we did a root cause" with nothing else. So I reckon it also makes commercial sense to have some transparency. Even if they said that effected companies have been given full details of the root cause analysis would have been fine.
When I worked for a major storage vendor the one thing that was drummed into us as technical support managers was to be as honest and open with our customers in the event of outages. We managed to get a pretty good reputation for providing not only top notch support but honesty (well as far as tech support was concerned) as well.
Splicing in a new NSA feed?
Speaking of Flickr
I see that XKCD took a swing at their crap UI
Office 365 Lync had a little lie down earlier too.
Re: Friday 13th.
But I thought the billions we've invested in anti-virus software not to mention the architectural changes we've made to the systems eliminated that virus years ago!
"We all feel it"
- Stick a 4K in them: Super high-res TVs are DONE
- Review You didn't get the MeMO? Asus Pad 7 Android tab is ... not bad
- BEST BATTERY EVER: All lithium, all the time, plus a dash of carbon nano-stuff
- If it weren't for that GIANT ASTEROID. Sigh. 'Colossal bad luck', old DINOSAUR chap
- Bring back error correction, say Danish 'net boffins