Help Slashdot Test Our New Data Center

After many years of living in California, Slashdot is preparing to move to a new data center in Chicago, and we need your help. We have our new site running a dump of our database from a few days ago. You can hit it at beta.slashdot.org. Please go there, post comments, submit stories, and do whatever you do normally. Or maybe abnormally — run crawlers, write poll spamming robots or something. If you find any crazy issues, please submit them to our sourceforge tracker. If you're curious, the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM.

Comments

well done slashdot

[nih]

it's working gre

Wow...all that computer power!

[Anonymous Coward]

Wow...all that power just to push a few kilobytes of text to my computer.

Re:Wow...all that computer power!

[Anonymous Coward]

Well, with a mysql and perl combination... it's not hard to see why they need all that power.

Obligitory

[BluffBlank]

But does it run linux?

Re:Obligitory

[tkid]

Did you read the history of Slashdot when it hit 10 year anniversary mark? I think someone should go read and stop asking silly questions.

Re:Obligitory

[brxndxn]

Imagine a Beowulf cluster of those!

Re:Obligitory

[krgallagher]

"But does it run linux?"

I know that was meant to be a joke, but that is my question. What is the OS? What is the Database engine? What is the web server? Are you using a load balancer? Details! I want details!

Re:Obligitory

[__aanjtz122]

$ lynx -dump -head http://beta.slashdot.org/ | grep Server

Server: Apache/1.3.41 (Unix) mod_perl/1.31-rc4

So that's one question answered :-)

Re:Obligitory

[Hatta]

iblis% sudo nmap -A -T4 beta.slashdot.org

Starting Nmap 4.53 ( http://insecure.org/ ) at 2008-05-23 15:00 CDT
SCRIPT ENGINE: rpcinfo.nse is not a file.
SCRIPT ENGINE: Aborting script scan.
Interesting ports on beta.slashdot.org (216.34.181.45):
Not shown: 1704 closed ports
PORT STATE SERVICE VERSION
21/tcp open tcpwrapped
25/tcp filtered smtp
80/tcp open http Apache httpd 1.3.41 ((Unix) mod_perl/1.31-rc4)
135/tcp filtered msrpc
136/tcp filtered profile
137/tcp filtered netbios-ns
138/tcp filtered netbios-dgm
139/tcp filtered netbios-ssn
443/tcp open ssl/http Apache httpd 1.3.41 ((Unix) mod_perl/1.31-rc4)
445/tcp filtered microsoft-ds
Device type: general purpose
Running (JUST GUESSING) : OpenBSD 4.X (91%)
Aggressive OS guesses: OpenBSD 4.1 (x86) (91%), OpenBSD 4.0 - 4.2 (90%), OpenBSD 4.0 (88%), OpenBSD 4.0 (x86) (86%)
No exact OS matches for host (test conditions non-ideal).
Network Distance: 10 hops

TRACEROUTE (using port 21/tcp)
HOP RTT ADDRESS
1 0.25 beta.slashdot.org (216.34.181.45)

OS and Service detection performed. Please report any incorrect results at http://insecure.org/nmap/submit/ .
Nmap done: 1 IP address (1 host up) scanned in 32.605 seconds

Looks like they're running OpenBSD of some sort.

Re:Obligitory

[moderatorrater]

That's impossible. I thought BSD was dead, or at least dying.

Re:Obligitory

[livingdeadline]

Netcraft thinks it's an F5 BIG-IP device, which probably means that they've got a big, fancy load balancer. If my memory doesn't fail me, F5's appliance OS is BSD based.

Re:Obligitory

[Schemat1c]

Did you read the history of Slashdot when it hit 10 year anniversary mark? I think someone should go read and stop asking silly questions. Whoosh!

oh yeah

[stoolpigeon]

my cowboy neal bot has already gained complete control of all survey functionality.
 
A Natalie Portman statue and a 10 gallon tub of hot grits or I wont give control back. Oh - and I want the ponies theme back. Permanently.

Abnormally?

[FooAtWFU]

Run abnormally? So, throw a few of these out there?

ab -n 10000 -c 10 http://beta.slashdot.org/

Re:Abnormally?

[Anonymous Coward]

So it was really a typo: Run ab normally.

Re:Abnormally?

[FooAtWFU]

Apache benchmarker. If you have the right packages installed (e.g. 'apache2-utils' on Debian) it certainly exists on any Linux distro worth its salt.

L8r D00dz

[FurtiveGlancer]

I'm off to ride the new iron!

Tracerouted

[Anonymous Coward]

3 fewer hops to 11, latency tripled to 85ms

Re:Tracerouted

[Joe The Dragon]

I get 10ms

Re:Tracerouted

[mnslinky]

wow, 5 fewer hops to me, with about half the latency

Re:Tracerouted

[Odin_Tiger]

5 fewer for me, about 2/3 latency.

Re:Tracerouted

[Anubis350]

I'm in NYC, dropped latency rather impressively from 96ms to 36ms, dropped hops from 16 to 13, all depends on where you live and what networks your traffic is going over

Re:Tracerouted

[the_one(2)]

I get 16 hops from sweden on both. latency is about 40 ms less

Re:Tracerouted

[bram]

try mtr

Re:Tracerouted

[DerekLyons]

It's not as if on a text based site anyone can really notice latency below a couple of hundred microseconds.

Re:Tracerouted

[GumphMaster]

Latency! That ain't latency:

• Current 14 hops 236ms
• New 16 hops 317ms

Figures courtesy of tracepath. The packets have paddled across the Pacific.

"dump"...

[MadFarmAnimalz]

I can't explain it (oh okay, maybe I can) when the database in question contains /. comments and old Jon Katz stories (quite apart from Taco's nicely spelled and eloquent posts), calling it a dump somehow seems appropriate in the more scatological sense of that word...

you're done

[amazon10x]

I have access to the storm botnet. Should I try to DDoS the new data center?

Re:you're done

[idiotwithastick]

Alternatively, we could always just try the Slashdot effect.

Re:you're done

[Culture20]

Does having your computer be a zombie node equate to "access"? ;-)

Re:you're done

[compro01]

so we're gonna slashdot slashdot?

new account

[altair1230]

I couldn't create a new account on beta.slashdot.org

Re:new account

[Anonymous Coward]

Funny, *mine* still works...

Re:new account

[kesuki]

okay i've tried 4 of my slashdot accounts, in opera for linux, and none of them lets me log into beta.slashdot.org...

mine was low 6 digits..
the rest were in the high 6 digits...

tried with wand asking, enabled and disables, as well as 'public terminal' checked, no dice, can't log in with opera at all.

I think....

[ForestGrump]

I think slashdot slashdotted themselves.

Re:I think....

[Idbar]

They just wanted to coin the term: "recursive slashdotting".

Slashdot hardware specs...

[madskyllz]

"If you're curious, the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM" What do you mean, "If" ?

Re:Slashdot hardware specs...

[Talen317]

Does this mean that "If I'm not curious" the system runs on some other hardware?
Perhaps an Apple IIe?

What if I'm curious one day but not curious the other day?
Will that dramatically affect my user experience?

Seems recursive

[linuxwrangler]

So let me get this straight. You want us to Slashdot....Slashdot?

Re:Seems recursive

[Bovius]

Stop it, you're freaking me out.

Re:Seems recursive

[Dachannien]

Here's a Coral Cache link, just in case. ;)

Power consumption?

[Anne_Nonymous]

> If you're curious, the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM.

What's the power consumption on this set up?

Re:Power consumption?

[AKAImBatman]

What's the power consumption on this set up?

Probably less than you'd expect. Intel's latest line of Quad-core Xeon processors use 45nm fab technology. If these are 2.33GHz "Harpertown" chips, then each chip (not core!) uses only ~50 watts. That means these machines could be running in as little as 150 watts for the dual processor machines, and 250 watts for the quad processor machines. (Though the quad core machines are most likely drawing a lot more power than that for a full RAID array.) Which really isn't too shabby.

Re:Power consumption?

[hansamurai]

Three puppies and a hair from CmdrTaco's head every 108 minutes.

Re:Power consumption?

[sYkSh0n3]

Now if we can just find a way to get the datacenter up to 88mph...

Re:Power consumption?

[linuxpyro]

That would explain the dupes.

Re:Power consumption?

[legoman666]

And that's only if they're at 100% load, which they won't be except in rare cases.

Re:Power consumption?

[zdzichu]

This is /.. If there were 2.33 chips they would write 2.33. 2.3 GHz clearly show it's Barcelona Opteron. They've TDP of 137W (or 74W in case of High Efficiency chips).

Re:Power consumption?

[omnipresentbob]

They're all quad core machines. 18 of 'em have 2 quad core processors, and the other 4 lack the extra processor but get more RAM.

It's sorta like, there's 22 guys. 18 of 'em ain't missing nothing, but the other four only have one nut but have more endurance. Though the analogy is a bit off.

I'm in downtown Chicago right now

[lpangelrob]

If you want me to try to walk out with your servers, just post the address. I'll let you know if I'm successful, but odds are you'll know long before then...

Not a problem

[dedazo]

18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM.

twitter will create a few more sockpuppets and bring this rig to its rack-mounted knees.

More info about the hardware?

[Joe The Dragon]

AMD or INTEL?

What is the raid setup?

Hardware raid? raid 5? raid 6? raid 50? others?

SAS? scsi?

MB chipset?

How are they linked to each other Dual or more teamed gig-e

10 gig-e?

others?

Re:More info about the hardware?

[Precision]

On the Web servers, the specs are

Dual Quad Core Intel E5345 with RAID 1 of 2xSATA drives

On the Database boxes

Single Quad Core Intel E5345 with RAID 5 of 7xSAS drives w/ a hot spare

All machines are 2xGigE connected to cabinet switches which are 10G connected to our cores and our Dual 10G uplinks

Re:More info about the hardware?

[felipekk]

Yeah, what version of IIS are they running?

And SQL 2008 I suppose?

*hides*

Re:More info about the hardware?

[allscan]

Only dual 10G uplinks?

Only on Slashdot

[FurtiveGlancer]

do they intentionally slashdot the beta.slashdot.

What is the current hardware?

[llZENll]

Your FAQ is from 2000 http://slashdot.org/faq/tech.shtml#te050

Re:What is the current hardware?

[Odin_Tiger]

http://meta.slashdot.org/article.pl?sid=07/10/18/1641203

Slashdot currently has 16 web servers all of which are running Red Hat 9. Two serve static content: javascript, images, and the front page for non logged-in users. Four serve the front page to logged in users. And the remaining ten handle comment pages. All web servers are Rackable 1U servers with 2 Xeon 2.66Ghz processors, 2GB of RAM, and 2x80GB IDE hard drives. The web servers all NFS mount the NFS server, which is a Rackable 2U with 2 Xeon 2.4Ghz processors, 2GB of RAM, and 4x36GB 15K RPM SCSI drives. (CT: Just as a note, we frequently shuffle these 16 servers from one task to another to handle changes in load or performance. Next week's software story will explain in much more detail exactly what we do with those machines. Also as a note- the NFS is read-only, which was really the only safe way to use NFS around 1999 when we started doing it this way.)

Besides the 16 web servers, we have 7 databases. They currently are all running CentOS 4. They breakdown as follows: 2 Dual Opteron 270's with 16GB RAM, 4x36GB 15K RPM SCSI Drives These are doing multiple-master replication, with one acting as Slashdot's single write-only DB, and the other acting as a reader. We have the ability to swap their functions dynamically at any time, providing an acceptable level of failover.

2 Dual Opteron 270's with 8GB RAM, 4x36GB 15K RPM SCSI Drives These are Slashdot's reader DBs. Each derives data from a specific master database (listed above). The idea is that we can add more reader databases as we need to scale. These boxes are barely a year old now -- and still are plenty fast for our needs.

Lastly, we have 3 Quad P3 Xeon 700Mhz with 4GB RAM, 8x36GB 10K RPM SCSI Drives which are sort of our miscellaneous 'other' boxes. They are used to host our accesslog writer, an accesslog reader, and Slashdot's search database. We need this much for accesslogs because moderation and stats require a lot of CPU time for computation.

Ah, the Data Center robbery center of America!

[snarfies]

How soon we forget.

http://it.slashdot.org/article.pl?sid=07/11/03/2054208

WISE choice, Slashdot. Wise choice.

Re:Ah, the Data Center robbery center of America!

[hakr89]

They're not being hosted out of CI Host, they're being hosted out of Savvis.

great choice

[ILuvRamen]

Nobody should have any sort of datacenter in CA cuz it's too hot and the power to cool it is too expensive. Chicago was a far better choice! Plus it's closer to Wisconsin where I live so the response time will be better. That's even more important ;)

I'm sure I'm not alone...

[Anonymous Coward]

"the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM"
I just got an erection

Re:I'm sure I'm not alone...

[StarfishOne]

You just gave my mind a way to fast association with "uptime"... ;o

And to be ahead of obvious replies: has Netcraft confirmed it? :P

Weaker Databases?

[TheNinjaroach]

If you're curious, the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM. I read this as "two quad-cores per webserver, and a single quad-core per database." Is this correct?

I was always under the impression that database servers needed more CPU horsepower than webservers.

Re:Weaker Databases?

[bol]

Depending on their architecture they very well be more web server bound than DB.

If they are more web server bound than DB I would suspect that they are using a high degree of server side caching to remove as much dependency on the database as possible. The DBs are also likely scaled out across the web servers and load balanced as well with possibly a distributed data set. This would require the front end web servers to do more work in querying the databases and displaying the results.

For example, a mod on the article ID may determine which database that article is stored on distributing the load evenly across N database nodes.

That combined with front end caching will give you almost unlimited scalability and is the foundation for all high traffic web sites.

Re:Weaker Databases?

[jpbelang]

I remember a USENIX talk about Altavista's servers and they had many more web servers than DB servers (like 3 to 1).

So I'm not surprised.

Oblig

[felipekk]

I, for one, welcome our new beta Slashdot overlords.

Tempus Interruptus!

[FurtiveGlancer]

Please, kill the 2-minute comment delay timer on beta.slashdot. We really want to test it!

Re:Tempus Interruptus!

[Mastadex]

In other words:

"First Post"
Re: "First Post" - "Second Post"
Re: Re: "Second Post" - "Third Post"
etc...

I know your kind.

Morbo crush your puny data center

[davidwr]

If you're curious, the new system features 18 2x quad-core 2.3 GHz webservers each with 8 gigs of RAM, and 4 quad-core 2.3 GHz databases with 16 gigs of RAM. Morbo crush your puny data center with his mighty story submissions!

Best of luck,

[hwyengr]

Hopefully it's just downtown and not at 64th and Oakley. http://maps.google.com/maps?f=q&hl=en&geocode=&q=6436+S+Oakley+Ave,+Chicago,+IL+60636&sll=37.0625,-95.677068&sspn=28.611123,60.292969&ie=UTF8&ll=41.780049,-87.680275&spn=0.00656,0.01472&t=k&z=16&layer=c&cbll=41.776747,-87.681011&panoid=R4fqWJEJkVCRco2uu3GK9Q&cbp=1,248.03517600888696,,1,8.383780676689486 Security's pretty tight down there.

Re:Old content

[larry bagina]

Those aren't old stories, they're dupes.

Re:Just one question

[Omestes]

Then why do you guys keep moving to my state, and messing it up?

We have no beaches, we have 2 cities, the weather is NEVER mild, but yet we have an endless stream of Californians coming to to mess everything up.

Living in AZ, sometimes I wonder if Californians are like locusts, they all swarmed to CA, killed it, and now have to move on to somewhere else.

Pictures?

[Matt Perry]

Some pictures of the new setup would be great.

It's all about the Pizza, isn't it?

[Spinlock_1977]

Been to California (4 weeks), New York (7 years), and Chicago (2 weeks). The pizza in California sux. NYC is better, but I think Chicago gets the nod. Good move Slashdot!

Can't log in using Opera

[RobertB-DC]

I can't log in to the new site using Opera. IE works just fine. But with Opera, the system returns me to the screen it should, but not logged in. The login doesn't work in any access technique -- main page, sidebar, even the Ajax-y popup thing.

It's validating, because if I use invalid data, I get "Danger Will Robinson!" But if I use my actual login/pwd, it silently fails.

I'm not about to create a Sourceforge account just for this, though... I suspect they'll find enough bugs without my direct help.

Comcast Routes are Screwy

[slifox]

I live just south of the loop (downtown Chicago). Using Comcast cable, I get routed to Boston, back to Chicago, then to New York, then back to Elk Grove (near Chicago)!

Comcast from a suburb just north of Chicago does nearly the same thing:
    Chicago, to Chartford CT, back to Chicago, to New York Savvis, then to Elk Grove, and finally to beta.slashdot.org.

That doesn't make a whole lot of sense... Who the hell setup Comcast's routing??

For comparison, on the Illinois Century Network (an Internet2-capable ISP for many universities and research labs), I get routed through Chicago Cogentco, then right to Chicago Savvis, and finally to beta.slashdot.org

Can anyone explain Comcast's weird routes?

IPv6 please

[ptudor]

Can you please start listening on a v6 address?