Things should be back to normal now. Trevor just moved www.ycombinator.com to Slicehost, and I just told HN to refer to static stuff there instead of serving it locally. Interesting how easy it is to move your whole web site. Good customer service is important when users can switch so easily.
Thank you ! Procrastination becomes much harder when HN does not provide a continuous flow of interesting distraction.
Cause of YC/HN outage discovered
71–80 of 110 posts
Re: Cause of YC/HN outage discovered
#72Earlier quoted context omitted.
VPSes are still shared hosting and are not immune from the problem.
They won't shut you down for over-using resources. As a matter of fact you can't over-use resources, you can only use the resources you are given to you. I know they are different type of VPS, I am referring to Slicehost/Linode type VPS which has hard limit and acts very much like dedicated single box. More important to me is that I am really really really surprised that Ycombinator was running on a shared account. T…
Re: Cause of YC/HN outage discovered
#73I'm not sure what is more ridiculous: that they disabled the site without warning or that they were too lazy to look at their logs or do a simple Google search to find out that Hacker News is real site with lots of regular traffic.
What I really loved is that they just demonstrated a fair degree of incompetence to a site used pretty much exclusively by people who are very good potential customers! This site is a wet-dream for their marketing department. What they should have done is upgrade the website to a dedicated server for free and let that news hit the front page.
Re: Cause of YC/HN outage discovered
#74Shared hosting is generally sucky for anything remotely successful. I'm amazed you got away with storing your static assets there so long! When I got 100k hits in a day on my first blog, Dreamhost promptly shut it down without warning (in the middle of a slashdotting!)
Dreamhost? Their main advertising point is UNLIMITED TRAFFIC!!! I guess I'll scratch them off the list of potential hosts.
I jumped on a deal at the beginning of this year for ~$9 for 1 year of shared hosting. I now use that account purely for running things like scripts through SSH for doing basic data grabs with wget for further analysis elsewhere.
I've more than got my money's worth and I've yet to see any kind of email message asking me to upgrade.
Re: Cause of YC/HN outage discovered
#75Things should be back to normal now. Trevor just moved www.ycombinator.com to Slicehost, and I just told HN to refer to static stuff there instead of serving it locally. Interesting how easy it is to move your whole web site. Good customer service is important when users can switch so easily.
Any particular reason for Slicehost as opposed to any of the other similar options (Linode, prgmr)? I'm not affiliated with any, though I run my site on VPSes from Linode. I was convinced by a Dec 2009 post by an HNer, uggedal: http://journal.uggedal.com/vps-performance-comparison
Edit: Slicehost help articles & resources, well organised & exhaustive is a plus though.
Re: Cause of YC/HN outage discovered
#76It makes it easy to minify, combine, gzip and push your css, js and image assets into the Amazon Cloudfront CDN with far-future expiration headers. It also automagically detects background images referenced in your css, and puts them in the CDN. It rewrites the css to use the new CDN image urls.
It's all done through a trivial REST API.
Would love some feedback, and to find out if/how it's breaking any of your complex css/js.
Re: Cause of YC/HN outage discovered
#77Re: Cause of YC/HN outage discovered
#78Things should be back to normal now. Trevor just moved www.ycombinator.com to Slicehost, and I just told HN to refer to static stuff there instead of serving it locally. Interesting how easy it is to move your whole web site. Good customer service is important when users can switch so easily.
Re: Cause of YC/HN outage discovered
#79Re: Cause of YC/HN outage discovered
#80Earlier quoted context omitted.
Not this site, www. There was no sudden spike in traffic. If they'd bothered to check the logs they'd have found that the load, whatever it was, was no higher than it had been.
So were they just jerking you around then?