108 comments

[ 0.74 ms ] story [ 215 ms ] thread
Good news, it's up now!
It was pretty horrible. I thought where I’d read about it and couldn’t think of any place that wasn’t down.
As one of the people furiously hitting reload, I apologise.
I went digging into what was wrong with my network. I'm deep into updating my pihole now, for better or for worse. Wish me luck!
Argh. And if you run it in Docker like I do, I have to relearn the lesson that taking down the container, losing internet, then trying to do a ‘docker pull’ is a a bad idea.

Sometimes I feel the attraction of some spyware mesh system for home networking.

That's why you run two pihole instances on independent hardware :)
You’ve worked out how I ended up with 3, virtualised and on different hardware.
I was able to post an angry tweet about my ISP, then I realized Twitter was working.
Very high uptime overall. Credits to PG and his lisp/arc inspired codebase! And to dang (who I assume does a bit of coding on it from time to time from previous comments I have read..sorry if I read wrong)
This is one of a few lightweight, reliable sites where any problem makes me assume that my internet connection must somehow have gone awry.
Today it has been raining heavily, and the only reason I went to my co-working place is because I checked HN in the morning, and figured my internet was down at home because HN didn't work.
I heard from an ISP owner that a common support call was when kids deleted the browser icon on the desktop, the parent thinks internet is down and calls up support. These days almost everyone is using a phone or tablet so it happens less.
Right, HN is the site I check when I'm not sure my connection is working, because it loads so fast. It HN doesn't load, then I'm more likely to walk away from the computer than to check another site.
Or there's some hot topic with more than 2 replies that's making the server pull its hair out...
I worked a place that used to use ‘is HN up?’ as a proxy for ‘does the internet go?’
Time was I would use sun.com to check if the network was up. These days, like you, I use HN.
For decades now pings to yahoo.com to check connectivity have been my only direct interaction with them. If they go away or start responding to pings I could use news.ycombinator.com, but the URL is longer.
I was playing with my DNS settings and Hackernews was one of the sites I used to test. I picked the wrong day...
Ditto! It took me a while to realise my DNS settings were ok.
What percentage of HN readers are doing recreational DNS configuration at any given time?
It was terrible, had to resort to reading CLRS in the meantime.
I had to resort to lobste.rs. Still, it gave me a chance to revisit all those stories I read on HN a fortnight ago.
I've wanted to re-visit StumbleUpon before it became an ad-ridden shell of its former self. Push button, get random website based on your interests.
The last few weeks I’m continuously getting white pages on dynamic pages: login, profile, comments. What is going on? (Safari, iOS)

Things seem to work on Firefox on mobile. Something with internet relay perhaps?

worst seconds of my life
Browsing to HN was the first thing I tried to do this morning after accidentally sleeping in. When it failed to load, my panicked thought was: "Is this how it all ends? Did Putin finally press the red button?"
I noticed! My internet felt slow so I pinged HN since it's my baseline. Bad luck lol.
I am at gmt-5. I cannot sleep, my wife is out of town, I lost mi kindle the day before yesterday and HN was down. If I turn in the lights to read a book I will not sleep for the rest of the night so I listened to our cats fighting over what I assume was some big moth as my only distraction.
Sounds horrible, my condolences

Good that we are past the bump now

Reading a book will keep you awake for the night, but staring at a screen won't?
Probably in dark mode... Whether this really helps is another matter...
Awww, I missed it.

  >Awww, I missed it.
It was a day that will live in infamy.
The last time we had a hard drive failure, we were down for a couple days. That was early 2013.

After that, in the spirit of 'never again', Nick (kogir) set up the failover server system that we still use today. That's why you missed it tonight; you wouldn't have back then. Sorry, I guess?

Edit: by pure coincidence, Nick was in town tonight and we met up for the first time in years. Two hours later HN goes down. almost as if the server was overcome by nostalgia

A+ for Timing! - Quantum Entanglement? - The Mind and The Machine.
Not to worry; you just got a second chance ;)
(comment deleted)
I just automatically assumed I was having internet issues instead of checking other connections like I usually do when other sites have issues.
ᶦ ᶠᵉˡᵗ ˢᵒ ˡᵒⁿᵉˡʸ ᵇᵘᵗ ᶦᵗ'ˢ ᵇᵉᵗᵗᵉʳ ⁿᵒʷ
Looks like a hard drive stopped working. We switched to the failover server.

Sorry everyone!

Well, that's ok... thanks for being up to fix it.

It's not an actual spinning hard drive, is it?

Probably a 2.5 MB one-platter Diablo hard disk drive cartridge running on a restored Xerox Alto.

https://en.wikipedia.org/wiki/Xerox_Alto

Diablo Systems Incorporated Series 30 Disk Drive Maintenance Manual

http://bitsavers.org/pdf/diablo/disk/model_30/81503-02_Serie...

Restoring Y Combinator's Xerox Alto, day 4: What's running on the system (righto.com):

https://news.ycombinator.com/item?id=12197591

http://www.righto.com/2016/07/restoring-y-combinators-xerox-...

Xerox Alto Restoration Part 16 - our disk goes down, the Alto connects to Google and draws fractals

https://www.youtube.com/watch?v=adEr2aRwHnI

Our Diablo disk goes on the fritz, but who needs a disk when you can netboot? Ken demonstrates the Alto network capabilities, connects to Google, and has the Alto calculate and display a Mandlebrot set. Ken's in-depth blog entry including the fractal demo source code is found here:

http://www.righto.com/2017/06/one-hour-mandelbrot-creating-f...

Xerox Alto Restoration Part 1 - power supply restoration, disk drive surprise

https://www.youtube.com/watch?v=xPyqQXFC2yw

We begin our very gentle and progressive power up of the seminal Xerox Alto. No magic smoke, but one power supply is faulty. Opening it up reveals that it had a tough life, having suffered a catastrophic short of some sort, hastily repaired, and some traces almost entirely corroded through. But the source of the malfunction seems to be a somewhat classic case of bad electrolytic capacitors, way too far gone for any hope of reforming. After replacing them and repairing the supply, we turn our attention to the Diablo disc drive and cartridge, and have a bit of a surprise.

Many thanks to my CHM restorers colleagues Ron Crane, Ken Shirriff, Carl Claunch and Luca Severini.

See previous video introducing this historically significant machine:

https://youtu.be/YupOC_6bfMI

For much more details and references, see Ken Shirriff's blog entry corresponding to this video here:

http://www.righto.com/2016/06/restoring-y-combinators-xerox-...

A 1970s disk drive that wouldn't seek: getting our Xerox Alto running again

http://www.righto.com/2018/03/a-1970s-disk-drive-that-wouldn...

Identify It Challenge for 7-26-2012 Answer

https://reinventingscience.wordpress.com/tag/diablo-systems-...

ARTIFACT DETAILS: Series 30 disk drive

https://www.computerhistory.org/collections/catalog/10266694...

Description: "Not working cards mis...

There is a good chance that it is (or was!) an actual spinning hard drive. Whatever it is, it lives in one of our boxes at M5 and it's in their hands for the moment.
It was an SSD. A 1.6TB SAS3 SSD. (M5 CEO here)
Stop making stuff up guys, I just know that someone at the YCombinator HQ tripped over the power cable of the Raspberry Pi you're hosting this on.
(comment deleted)
If you had an auto scaling kubernetes cluster with multiple redundancies using rust and 3 JS frameworks this wouldn't happen. ;)
... but many other horrible things might
That being said, is the actual production infra archi for HN described somewhere ? Curious how simple it can afford to be.

We laugh at people piling layers and layers and artifacts on their sites, all in the hope of adding redundancy, handle "webscale" load, and avoid an outage (ironically increasing the chances that _something_ will break).

However, if a single hard drive crashing somewhere can cause your site to be down for minutes or hours, some non-tech people (managers, shareholders, customers) will wonder if the site is "professionnal" enough - and I can sympathize with them.

From 2018: <https://news.ycombinator.com/item?id=16076041>

> We’re recently running two machines (master and standby) at M5 Hosting. All of HN runs on a single box, nothing exotic: CPU: Intel(R) Xeon(R) CPU E5-2637 v4 @ 3.50GHz (3500.07-MHz K8-class CPU) FreeBSD/SMP: 2 package(s) x 4 core(s) x 2 hardware threads Mirrored SSDs for data, mirrored magnetic for logs (UFS) We get around 4M requests a day.

(comment deleted)
Three.js on HN? Interesting thought.
Raymarched SDF pyramids implemented with GLSL shaders for the voting arrows! They could be so shiny!

(It's a wonder that anyone lets me near the frontend of their websites really.)

It wouldn't surprise me if someone managed to implement a 3d voting arrow in less than the 407B transfer size of the existing .gif
huh, all this time I thought the arrow was just a unicode character
Imagine a Beowulf cluster of those!!111!!eleven!!1!
Irony aside, what's the point? In theory, yes, it could work better. In practice though, HN with its two baremetal boxes has better uptime than 99,99% of the Web, including the biggest ones - just because complexity has its price.
(comment deleted)
> If you had an auto scaling kubernetes cluster with multiple redundancies using rust and 3 JS frameworks outages like this wouldn't surprise your users anymore.

FTFY

You need to rewrite Hacker News in Rust to prevent these sorts of things!
And just as I was on the train to work. Worst. Website. Ever.
If this site being down was that impactful to your commute, it seems like you actually feel like it has a lot of value.
I’m curious as to what time zone you’re in. Or if there’s multiple people behind your account. It’s pretty impressive how omnipresent you are :).
No apology necessary, but I'm curious how a hard drive failure caused an outage. No RAID or mirroring? No hot spares? No clustering or distributed systems?
It was part of a mirror of identical SSDs on an LSI MegaRAID RAID card. We see occasional "spectacular" drive failures that take the machine down with a single disk failure. Usually it's just a reboot to come back up, and a disk replacement, then some hours of time to rebuild the array and get back to situation nominal.
For two hours I thought I was finally blocked.
This caused me a small but concerning amount of anxiety that I'll probably meditate on as I fall asleep now that HN is back up.