HN will be down Saturday morning while we switch servers • Hacker News Board

Re: HN will be down Saturday morning while we switch servers

#101

Post by tomschlick » Sat, Feb 16, 2013, 8:47 AM UTC

GeorgeTirebiter wrote: post #100

Hi. Could someone please explain to me why it is necessary for a service (e.g. HN) to go down while people play with the (increasingly amorphous and abstract) back end? Is it 1990? Sorry, just hit a nerve. Like doing some OS updates (Windows) and then needing to reboot to "complete the installation". I'm sorry. That totally sucks.

If you're using a mac http://selfcontrolapp.com/ works pretty well.

Re: HN will be down Saturday morning while we switch servers

#102

Post by JoachimSchipper » Sat, Feb 16, 2013, 9:17 AM UTC

ghshephard wrote: post #78

robbiep wrote: post #39

Do you think you could go down for a week? That's generally how long I need to break an addiction

Add me to the list - I'm in for $100 if we can make it a one week outage. $250 if we can stretch it out to two weeks. I'm sorry, but 10 minutes on a saturday morning does nothing for me.

You can lock yourself out easily with noprocrast and minaway set to a week or two (say 60 x 24 x 10). At least one prominent HN'er claims to have done this when needing a break.

Happy to have been of assistance; please send the $250 to GiveWell. ;-)

Re: HN will be down Saturday morning while we switch servers

#103

Post by jbverschoor » Sat, Feb 16, 2013, 9:43 AM UTC

pitchups wrote: post #91

pg wrote: post #55

Earlier quoted context omitted.

We currently get just over 200k uniques and just under 2m page views on weekdays (less on weekends).

Does a single server support this level of traffic? That is pretty impressive - it would be great to learn about the server's hardware and software configuration.

Around 40-60 requests per second, of which almost anything an be cached. Any single server should suffice.

Re: HN will be down Saturday morning while we switch servers

#104

Post by tomjen3 » Sat, Feb 16, 2013, 9:47 AM UTC

GeorgeTirebiter wrote: post #100

Hi. Could someone please explain to me why it is necessary for a service (e.g. HN) to go down while people play with the (increasingly amorphous and abstract) back end? Is it 1990? Sorry, just hit a nerve. Like doing some OS updates (Windows) and then needing to reboot to "complete the installation". I'm sorry. That totally sucks.

HNs backend isn't very abstract. The site doesn't really use Javascript and resides on a single server.

And I agree with you on the windows issue -- except that ten plus years ago I did in fact reboot and/or shut down my computer.

Re: HN will be down Saturday morning while we switch servers

#105

Post by tomjen3 » Sat, Feb 16, 2013, 9:49 AM UTC

d4vlx wrote: post #47

JoshTriplett wrote: post #12

Earlier quoted context omitted.

And the shiny new server replacing it: http://upload.wikimedia.org/wikipedia/commons/f/f1/Ibm_pc_51...

Ah the good old days, when computers were so simple and yet so mysterious and exciting at the same time.

I don't know, I still find them exciting -- although PC hardware isn't as fun as it used to be.

Re: HN will be down Saturday morning while we switch servers

#106

Post by tomjen3 » Sat, Feb 16, 2013, 9:50 AM UTC

cubix wrote: post #64

pg wrote: post #7

Earlier quoted context omitted.

http://imgur.com/VuSH2j5

I would have thought: http://en.wikipedia.org/wiki/File:LISP_machine.jpg

That would have been really cool, and a true hackers machine, but I imagine that the days when LISP needs (or benefits from) custom hardware are long gone.

Re: HN will be down Saturday morning while we switch servers

#107

Post by tomjen3 » Sat, Feb 16, 2013, 10:00 AM UTC

MaysonL wrote: post #89

pg wrote: post #7

Earlier quoted context omitted.

http://imgur.com/VuSH2j5

Not one of these? http://ed-thelen.org/comp-hist/vs-bendix-g15.jpg [one of the first computer models I ever wrote for - a three address machine: first a source operand, second destination, third address of next instruction. The memory was a rotating drum]

Mel?

Re: HN will be down Saturday morning while we switch servers

#108

Post by anu_gupta » Sat, Feb 16, 2013, 10:36 AM UTC

Wheee - currently much faster, although the true test will be on Monday, around 5pm GMT.

Re: HN will be down Saturday morning while we switch servers

#109

Post by markokocic » Sat, Feb 16, 2013, 11:18 AM UTC

jbverschoor wrote: post #103

pitchups wrote: post #91

Earlier quoted context omitted.

Does a single server support this level of traffic? That is pretty impressive - it would be great to learn about the server's hardware and software configuration.

Around 40-60 requests per second, of which almost anything an be cached. Any single server should suffice.

But you would need at least 30 Heroku dynos to handle that reliably ;)

Re: HN will be down Saturday morning while we switch servers

#110

Post by InclinedPlane » Sat, Feb 16, 2013, 12:55 PM UTC

spoiledtechie wrote: post #20

iamwil wrote: post #16

Earlier quoted context omitted.

And I recall the last time I looked at the source (when it was released with Arc), it didn't use a database either. All the data is stored as files.

Files???? That's a joke right???

I find it interesting how easy it is to criticize a functioning system based on some aspect of non-conformance with some hypothetical ideal. The idea that HN wouldn't run on a database is no less astounding than the idea that much of facebook runs on PHP. Design of real systems is often messy and imperfect and deviates from the ideal due to necessity of optimizing one or another factor that may not be obvious.