digg

Join DiggAbout
See the original image at vgpro.com

What Digg does to Servers [1016% Server Load!]

vgpro.com — This is a screenshot of the output from running the Top command on my CentOS box before it completely crashed due to the traffic from an article that made it to the homepage. The httpd.conf file had been misconfigured and was allowing far too many apache processes to spawn which ate up all the memory so the box switched to Swap memory (kswapd).

Bury
show profanity settings
expand all
  • AhrenBa, on 05/28/2008, -17/+316Hmm....how ironic....four diggs and this site is already not responding...
    • kidcodea, on 05/28/2008, -10/+22its linux shouldnt it be better than sliced bread?
      • arjie, on 05/28/2008, -3/+31It is, bread sliced under 1016% load is awful
    • ligyron, on 05/28/2008, -19/+7Your comment was made 11 minutes ago and this submission was made popular 11 minutes ago. I somehow doubt it made front page at 4 diggs
      • asskicker32, on 05/28/2008, -3/+14god forbid someone chekc the upcoming section...
        • ligyron, on 05/28/2008, -0/+2What're you talking about? If his comment was made 11 minutes ago and the article hit front page 11 minutes ago, that means that the submission had well over 4 diggs at the time he made his post (I'd imagine at least 100)

          There's a difference between submission time and popular time. It was submitted 19 hours ago
        • Yonson, on 05/29/2008, -1/+1God forbid someone submitted a comment after their initial observation.
    • asskicker32, on 05/28/2008, -2/+7The first word that popped in my head was "ironic"
    • t3rmv3locity, on 05/28/2008, -9/+18Here is a mirror: http://tinyurl.com/4ha32k
    • tewstep, on 05/28/2008, -6/+2wow, how Ironic...the page is...DAMNIT, unoriginal post on my part.
    • berryhill, on 05/28/2008, -0/+1Did you really have to spoil the whole point of your story in the title?
    • Jookly, on 05/28/2008, -2/+1That is actually exactly what I would expect to happen so it's not ironic at all, maybe even the opposite. Of course as Americans we don't need to know the actual meanings of words as long as we use them. Eh?
    • AsSubtleAsABrik, on 05/29/2008, -0/+1Seriously, this ***** is like 10,000 spoons when all you need is a knife.
  • AnotherThought, on 05/28/2008, -9/+218digg effect FTW
    • nhayes, on 05/28/2008, -6/+7FTL
    • SilenceGold, on 05/28/2008, -13/+1I have a solution for absorbing the digg effect.

      http://digg.com/mods/Someone_built_a_home_based_Li ...
      • jp12380, on 05/28/2008, -0/+4Won't do anything to get rid of the digg effect if it's hooked up to home DSL.
    • nullx42, on 05/28/2008, -1/+27Weird how digg never suffers from he digg effect..
      • 333cale, on 05/28/2008, -3/+17not really

        the severs load up because there is a sudden rush of processes waiting to access the load, all within a confined time period (i.e. when the digg gets made popular)

        Whereas Digg.com has an even capacity of processes and there is no cause for a sudden rush of visitors to the server. So Digg can always handle the smooth amount of traffic without crashing.
        • 30thElement, on 05/28/2008, -1/+16If people are going to a site through digg, then the number of page loads on digg and on the other site are (roughly) equal. The reason digg doesn't succumb to the digg effect is because they know they are going to get a lot of traffic and so have spent a lot of money on servers and bandwidth. Digg has rooms of servers, many of the sites on digg are just shared hosting, many sites on less servers.
        • repentantfan, on 05/29/2008, -0/+2@30thElement:
          333cale's point was the number of users to digg is generally constant, yet the proportion of those users visiting a referenced site varies greatly. The only variation digg might see is users who heard about a posting on digg from somewhere else, but this change is most likely quite small in relation. Yes, digg probably has many servers (or "big" servers), but less than we might expect I suppose.
      • redwallhp, on 05/28/2008, -0/+20What if Digg was Stumbled, Reddited, and covered on CNN at the same time...?
        • Acglaphotis, on 05/28/2008, -1/+7Nothing.
        • seventoes, on 05/28/2008, -1/+7Dont forget the seemingly-faded-out-of-existence Slashdot
        • D4rkDrago0n, on 05/28/2008, -0/+2See I always thought the world would end.

          tomatoe tomahto
        • t0xen, on 05/28/2008, -0/+3Then the internet will have divided by zero.
      • awesometastic1, on 05/28/2008, -0/+4it used to. it is only recently that digg wasn't slow as molasses
    • benologist, on 05/29/2008, -0/+0It's not the digg effect it's the consequence of ***** websites and/or cheap hosting. If people want to make 50 database calls to generate a page every time that page is viewed, on their $4 a month hosting account that's on a server shared by 600 other idiots making 50 database calls per page view, then their sites die.
  • chaos386, on 05/28/2008, -7/+449That's not 1016% load, it's 101600% (!!) divided by the number of cores in the system. In other words, the server needed to be 1016 times faster in order to cope with the load it was getting.
    • JlmAWP, on 05/28/2008, -5/+106So, rape, basically.
      • BoneheadFarker, on 05/28/2008, -3/+44With a bit of flogging thrown in for good measure.
        • DrVanilla, on 05/28/2008, -1/+48And digg wasn't even kind enough to give it a reach around.

          Terrible.
        • tunafizzle, on 05/28/2008, -3/+23Used the broomstick too. Thats never pleasant in the anus.
        • HonoredMule, on 05/28/2008, -0/+16Time to invest in a broomstick less prone to splintering.
    • reggplant, on 05/28/2008, -3/+52Server load is basically the number of processes waiting to access the process. I'm surprised the server managed to maintain such a load!
      • Nowaiman, on 05/28/2008, -0/+15That's what… she said?
      • gahal, on 05/28/2008, -0/+6It's linux.
        What amazes me is the swap daemon 12% cpu. That is some insane memory activity.
      • repentantfan, on 05/29/2008, -1/+4I used to manage Solaris 2.6 servers running sendmail for a well-known ISP. We would see this about once a week, but Solaris is great at handling traffic like this. We saw load averages of 5000 a few times also.

        And will people quit referring to load average as a percentage...

        @gahal:
        It didn't handle the load. The admin had to make configuration changes to handle the traffic. Don't be so ignorant as to think it handled the load, because "It's linux." Linux is great for small data requests, but not data heavy traffic. My company uses it for DNS resolvers, and it works great. On the flip-side, we tried using Linux as mail servers, in which it crumbled.
        • shakin, on 05/29/2008, -2/+2Linux is easily as good as Solaris, AIX or other "big" Unixes. In this case what's likely to be a poorly-written web app (and obviously not written to handle a high load) and a low-end server combined to bring it down under high load. I have yet to meet an OS that can swap nearly a gig and a half of data and while it's still being hammered run well.

          Linux runs many large mail servers as well as other types of servers with no touble at all.
        • gahal, on 05/29/2008, -0/+2@repentantfan
          My point was that It was still running.
          The machine was obviously on its knees, however it was still moving enough to run top and take a screenshot. Despite the punishment it was taking, it was still kicking.
          I have no idea what happened just before or after the shot.

          I agree that there are better tools for the job.
    • khail250, on 05/28/2008, -13/+5right or wrong, 101600% is a cooler number than 1016%, and i do not think computers are accurate enough to determine 1% as in the 1016%...
      • twiztidsinz, on 05/28/2008, -1/+20well either way you're wrong... its 1016.02 or 101602 which means that it IS showing down to 1% (or 0.01% depending on the numbers you use).
        And I do think that computers are accurate enough to determine 0.01%
    • tmoney578, on 05/28/2008, -16/+189it's OVER 9000!!!!
      • nationalist, on 05/29/2008, -8/+1118 digg and no responses???!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! :'(
        • LegendRoad, on 05/29/2008, -0/+7Some comments are better left without responses.
      • giantsfan134, on 05/29/2008, -0/+8WHAT 9000!?!
        • TomTruelle, on 05/29/2008, -4/+2hahahahhaa I always thought it was 1/9000th? which didnt make much sense.
    • FeartheKnighted, on 05/28/2008, -12/+6Thats no moon...
    • DCLXVI, on 05/28/2008, -4/+72The load average has nothing to do with percentages.

      that number 1016 is the sum of the run queue and and the number of jobs currently running, on average over the last 1 minute.

      In this case it basically means that apache has a lot of jobs queued up it can't handle at that given moment.
      • mentor, on 05/28/2008, -0/+9Thank you.

        Also, "interactivity" is essentially something different to load, and is, further, not necessarily a useful metric on a server.
      • SteveMax, on 05/28/2008, -0/+13Thank you. The comments above were destroying my faith in geekkind.
      • KevinAce, on 05/28/2008, -0/+5I was going to post a dork reply explaining load but you already took care of it - thanks.
    • noseeme, on 05/28/2008, -0/+17So for us UNIX geeks, that means the load average was 1016.0. :)
    • coder_cotton, on 05/28/2008, -1/+13Actually, it's the load average, not CPU utilization. It is the average amount of processes waiting for CPU time at certain intervals (last minute, last 5 min, last 15 min IIRC). That being said, a 1016 load avg is... bad.
      • whodathunk, on 05/28/2008, -0/+4Not 'waiting for CPU', simply 'waiting', which can also be for disk or network data, in which case it will be blocked, not using CPU but still adding to the loadavg (otherwise you could never see a loadavg > 1 and still have idle cpu cycles).

        And in this case, it was likely waiting for disk, due to the high memory usage: The waiting was probably waiting for the next needed page to be read back in from swap.

        So, for this case, (besides better software/settings), the loadavg probably would drop much more by adding RAM than by adding CPU cores (not really sure, because top says '0% wa', but I've seen top report that wrong more often (use vmstat and iostat instead. according to 'top' here the CPU is busy in 'sy', which would be a driver... a badly broken harddisk or network card driver maybe, or, more likely, a version of top that doesn't read /proc correctly for the kernel that is running...)...
    • ammut666, on 05/28/2008, -4/+0Just run an "apachectl stop" followed by an "apachectl start".

      durrr.....
    • adamward1985, on 05/28/2008, -1/+4That's a very good method to use. I recommend checking out this article for a full explanation: http://adamstechblog.com/2007/12/04/load-average-e ...
    • MCleveland89, on 05/29/2008, -0/+0That's what she said.
    • linuxpenguin, on 05/29/2008, -0/+1No, that's not true. Generally you want this load average to be under the number of CPUs you've got - so if he has a 2000 CPU server, he's set. The load average isn't a percentage per se - on a standard desktop machine, it's basically impossible to get a load average that high because anything near or above 1.0 is going to feel super-slow since you've probably only got one processor (if you have a hyper-threaded processor, it'll still feel slow, but not as much as with a normal single-core processor).

      The fact that this guy's machine was still running when it hit this point tells me he's not just using an old Dell in the basement. . . if his server's got 500 CPUs, a load average of 250 might be quite the norm.
  • rockrodger, on 05/28/2008, -10/+98I love you digg
  • borez, on 05/28/2008, -7/+30 David Lightman: People sometimes make mistakes.
    Joshua: Yes, they do.
  • juckman, on 05/28/2008, -89/+3What Porn does to Me [1016% Ejaculatory Load!]
    • follower64, on 05/28/2008, -4/+45Everyone is this thread just got dumber.
      • macfan93, on 05/29/2008, -2/+3I find it extremely ironic that you said "is" instead of "in" Pot calling the kettle black?
        • follower64, on 05/29/2008, -0/+5Pot saying *****
        • Jennefah, on 05/29/2008, -0/+4Or maybe juckman's comment caused follower64 to get dumber too, thus proving his (follower64's) point.
        • macfan93, on 05/29/2008, -0/+1That is the aforementioned "irony"
        • Jennefah, on 05/30/2008, -0/+1@macfan93... Are you sure that's irony? I thought it was just coincidental.
      • rockrodger, on 06/03/2008, -0/+1is or in? I think you sir made me dumber on your own
    • t2t2, on 05/28/2008, -7/+3Being on subject of Porn and Servers, here's something from Joomla documentation:

      "Top 10 Stupidest Administrator Tricks
      10. Use the cheapest hosting provider you can find

      Preferably use a shared server that hosts hundreds of other sites, some of which are high-traffic porn sites. Don't check the list of recommended hosting providers."
      http://docs.joomla.org/Top_10_Stupidest_Administra ...
    • astronautje, on 05/28/2008, -0/+2yeey! an alternative for the "report"-button. dumb *****
  • olithebumm, on 05/28/2008, -20/+7Dugg because I want this site to get the digg effect
  • estvir, on 05/28/2008, -29/+7Yes, Digg, Reddit, /. and other sites can cause a performance problems. We've known this for a bloody long time.

    Buried.
  • Mariasha, on 05/28/2008, -17/+4http://www.youtube.com/watch?v=XLLRsn_nr6s
    • senorcool, on 05/28/2008, -2/+14Are you suggesting we overload YouTube's servers or are you just posting an irrelevant, mediocre video?
  • brendanryder, on 05/28/2008, -10/+3woa. i wonder how much that all costs?
  • konkourteev, on 05/28/2008, -12/+2DIGG this!
  • ifonly, on 05/28/2008, -8/+20Server load figures are a completely inaccurate way to judge the performance of a system; you can have a load as high as 200 and not see any issues with performance, e.g. non I/O intensive applications, and not to mention this figure needs to be divided by the number of cores on the system.
    • adamward1985, on 05/28/2008, -0/+5I wrote a quick article a while back on reading the iostat command to see what percent of time your processors spend on waiting for disk I/O.

      http://adamstechblog.com/2007/12/20/how-to-read-th ...
    • HamNCheese, on 05/28/2008, -6/+0Wrong - the number of cores has absolutely nothing to do with the load average.
      • ifonly, on 05/28/2008, -0/+6Err you need to divide by the number of cores to see the real figure ...
      • adamward1985, on 05/28/2008, -0/+2HamNCheese - the number of processors recognized by the server is important. Load average is the number of processes in the run queue. When this number exceeds total recognized processor count performance will suffer.
  • chantron, on 05/28/2008, -15/+4OMFG TOP ON VISTA?
    • whereiseljefe, on 05/28/2008, -5/+4OMFG PUTTY!

      Dugg down for FAIL
    • chantron, on 05/28/2008, -0/+5give me an ASCII graphic showing something going WAY OVER SOMEONES HEAD.
      • GliTCH82, on 05/28/2008, -0/+2............................................________........................
        ....................................,.-‘”...................``~.,..................
        .............................,.-”...................................“-.,............
        .........................,/...............................................”:,........
        .....................,?......................................................,.....
        .................../...........................................................,}....
        ................./......................................................,:`^`..}....
        .............../...................................................,:”........./.....
        ..............?.....__.........................................:`.........../.....
        ............./__.(.....“~-,_..............................,:`........../........
        .........../(_....”~,_........“~,_....................,:`........_/...........
        ..........{.._$;_......”=,_.......“-,_.......,.-~-,},.~”;/....}...........
        ...........((.....*~_.......”=-._......“;,,./`..../”............../............
        ...,,,___.`~,......“~.,....................`.....}............../.............
        ............(....`=-,,.......`........................(......;_,,-”...............
        ............/.`~,......`-...................................../...................
        .............`~.*-,.....................................|,./.....,__...........
        ,,_..........}.>-._...................................|..............`=~-,....
        .....`=~-,__......`,.........................................................
        ...................`=~-,,.,......................................................
        ................................`:,,...........................`..............__..
        .....................................`=-,...................,%`>--==``.......
        ........................................_..........._,-%.......`...............
        ...................................,
    • teh_techie, on 05/28/2008, -0/+2It's likely NOT on top of vista. It's a terminal window, accessing a remote server.
    • chantron, on 05/28/2008, -2/+1Holy ***** give me TWO ASCII graphics showing stuff going WAY OVER SOMEONE'S HEAD. One for not knowing that top is the program and the other for not knowing I'm KIDDING. EGADS.
  • svensksvamp, on 05/28/2008, -3/+92Mirror: http://img519.imageshack.us/img519/7457/97248c0ca8 ...
    • DefaultGen, on 05/28/2008, -1/+3The site is back up but I dugg for the sheer irony.
  • diggit08, on 05/28/2008, -13/+6But will it blend?
  • adamward1985, on 05/28/2008, -3/+33A quick word about load averages: http://adamstechblog.com/2007/12/04/load-average-e ...
    • daok, on 05/28/2008, -0/+11Your website learn me something more than the principal article over here. Nice job
    • demonic, on 05/29/2008, -0/+1yeah cheers for that, was helpful, but it didn't show a solution on how to keep the processes down, anyone know where the find out howto?
  • Testies, on 05/28/2008, -7/+225WE NEED MORE COAL!!!!!!!!!
    • dajuggernaut, on 05/28/2008, -1/+35forget the coal, the hamster is having a heart attack.
      • FadieZ, on 05/28/2008, -4/+2forget the hamster, where's the greater mole rat?!
      • H1tchh1k3r, on 05/28/2008, -0/+3Where is my supply of whale blubber? my lamps are running low.
    • Tryptomine, on 05/28/2008, -0/+10"Shovel on more dogs, would you?"

      "Aye, more dogs!"
    • digitalarcanum, on 05/28/2008, -2/+7CONSTRUCT ADDITIONAL PYLONS
      • melonade, on 05/29/2008, -0/+1Some more vespene gas wouldn't hurt, either.
    • EpochOne, on 05/29/2008, -0/+0God damn digg... god damn.
  • geekevolved, on 05/28/2008, -20/+3ITS OVER ONE THOUSANDDDDDDD
  • dabdinoor, on 05/28/2008, -5/+37Now this site has crashed, oh irony.
    • xB4R7x, on 05/28/2008, -2/+5beat me to it... damn you.
    • spectre_25gt, on 05/28/2008, -2/+4I hate when people point this out, but irony means that there was an unexpected outcome. Is it really unexpected to see a server go down when it's serving up a screenshot of ridiculous load averages under the same circumstances?

      I propose a modification to the definition of irony.
  • destron, on 05/28/2008, -5/+14Uh, how ironic... the site is down.
  • clondike7, on 05/28/2008, -6/+95Killed by irony overload.
  • tachistoscope, on 05/28/2008, -4/+28This baby can take temperatures up to 9000 degrees.
  • Bluemotion, on 05/28/2008, -7/+42Down!
    IRONY... anyone.

    or was it always dead, and i'm missing the point?
  • reisrocks, on 05/28/2008, -5/+9is this the article which contains the most retarded comments so far?
  • breezytrees, on 05/28/2008, -3/+10can someone explain what I'm looking at?
    • CarStan, on 05/28/2008, -0/+22It's Goatse in Linux.
    • arjie, on 05/28/2008, -1/+6Okay, the numbers right after the words 'load average' are the number of processes that are using or waiting to use the processor. The numbers are for 1, 5 and 15 minutes respectively. So that particular server needed to be 1016 times as fast (or have 1016 processors) as it was to handle all the processes running in that minute perfectly. Similarly, over the last five minutes, the server was under a load that 415 processors could have handled perfectly.

      Also of note is the memory and swap. If you look at the fourth and fifth line of the terminal output, the server's memory is full and so it is paging large amounts of data to disk. That means that things tend to run slower because the data takes that much longer to get to the processor. kswapd is the kernel swap daemon, the program that handles the swapping.

      That's my understanding of it, but I'm not an expert.
    • t2t2, on 05/28/2008, -0/+8Computer Screen
  • msamjr, on 05/28/2008, -3/+104The Digg Effect: Now with 80% more irony
  • Mononuclear, on 05/28/2008, -0/+7the idiot's server gets dugg to death so he posts another article about it on the same server and submits it to digg?
    • michelspc, on 05/28/2008, -0/+6It's a great way to test changes you made to your httpd.conf file.
  • elislider, on 05/28/2008, -6/+149WE REQUIRE MORE VESPENE GAS
    • noseeme, on 05/28/2008, -6/+23***** YOUR VESPENE GAS, WE MUST CONSTRUCT ADDITIONAL PYLONS!!!
      /cruisecontrol
    • frostbyt, on 05/28/2008, -2/+34***** YOUR PYLONS! WE NEEDS MORE OVERLORDS.

      countdown to starcraft 2
    • dominikkom, on 05/28/2008, -0/+5You made my day
    • MaxisReed, on 05/28/2008, -0/+15EN TARO ADUN, EXECUTOR.
    • 3amboo, on 05/28/2008, -2/+3SHOW ME THE MONEY!!!!!!!!!!!!!!!!!!!!
    • microphonix, on 05/28/2008, -1/+3there is no cow level
      • Decoy84, on 05/29/2008, -0/+2that's because you killed the cow king
    • Oea420, on 05/28/2008, -0/+3too soon executus, you have awaken me too soon!


      o crap wrong game
    • linuxpenguin, on 05/29/2008, -0/+1operation CWAL
      show me the money

      Problem solved.
  • EllisAshbrook, on 05/28/2008, -13/+3there is no ***** irony in this.

    http://sc.tri-bit.com/Irony
    • andypop481, on 05/28/2008, -0/+1i believe your man blends sarcasm and irony together. There is irony in a website explaining the severity of the digg effect being overwhelmed by the digg effect.
  • D14BL0, on 05/28/2008, -0/+7Wow, what an expected turn of events.
  • zyl0x, on 05/28/2008, -8/+8Ironies.

    Let me show you them.
  • built2spill, on 05/28/2008, -8/+2About what I expected:

    Unable to connect

    Firefox can't establish a connection to the server at www.vgpro.com.

    * The site could be temporarily unavailable or too busy. Try again in a few
    moments.

    * If you are unable to load any pages, check your computer's network
    connection.

    * If your computer or network is protected by a firewall or proxy, make sure
    that Firefox is permitted to access the Web
  • HydrogenOxide, on 05/28/2008, -0/+3It seems the source of the fire were these stress release candles
    • sublimemm, on 05/28/2008, -0/+1whats that from? I cant remember, its on the tip of my tongue and its driving me nuts!
  • Zammo, on 05/28/2008, -1/+7fork() FTL

    the system is swapping to death here. noted by the lack of free memory, swap usage, kswapd being the #1 process. I've gotten a machine to a load average of 700, but never over 1000 :)

    nice job
    • dood, on 05/28/2008, -0/+3Back in the day I worked at a place hosting a bunch of sites on IRIX 6.2 (I think... might have been 5.3). At one point the load got so high that it wrapped to negative numbers.
      • mossblaser, on 05/28/2008, -0/+3If you are not joking then man is that impressive :D What the hell were you doing?!
        • dood, on 05/28/2008, -0/+2This was a shared hosting server with thousands of customers on it, and hundreds or thousands of interactive logins. Basically, it was doing something it was not really designed for.

          I forget what exactly went wrong that caused the load to climb so low, though. It ended up requiring a NMI to proceed (we couldn't afford to let it continue to try to work itself out, heh.)
  • karlw, on 05/28/2008, -0/+6clogging the tubes.
  • mjones77, on 05/28/2008, -1/+9This is what digg does in terms of mbps too a server when a image gets dugg http://uploadingit.com/view/637159_pxo0d
  • PhattyPhattMatt, on 05/28/2008, -6/+1i think i just made it 1017%, it wont load for me...
  • Akraz, on 05/28/2008, -6/+1ROFL the irony.. even this webserver went down


    The page cannot be displayed
    Explanation: The request timed out before the page could be retrieved.
  • frostbyt, on 05/28/2008, -1/+29Build more farms.
  • Zammo, on 05/28/2008, -7/+2http://en.wikipedia.org/wiki/Fork_bomb
  • InuX, on 05/28/2008, -7/+2[insert irony joke here]
  • galvo, on 05/28/2008, -2/+7Noob server.
  • ribo, on 05/28/2008, -1/+12ITT: people who don't understand system load. lrn2unix
    • skyshock1, on 05/29/2008, -0/+3What gave it away? The windows vista "terminal" session?
  • endlessoul, on 05/28/2008, -1/+2*****
  • Eze2k, on 05/28/2008, -10/+5ALL YOUR BASE ARE BELONG TO US.
  • Ocelot13, on 05/28/2008, -2/+34to protect yoursekf from digg... here are some simple steps
    1.) Dont host lolcat pictures
    2.) Dont host Obama articles
    That should be about it, follow those to keep your server safe
  • DrDustbunny, on 05/28/2008, -4/+13 Zombie?!?

    WTF?

    • adamward1985, on 05/28/2008, -0/+8A Zombie process is a process that fails to stop successfully. Generally these processes do not respond to the "-9" flag for kill. "-9" is the highest amount of priority for killing a process. Rebooting or waiting it out can help :-)

      Adam
  • Andyschism, on 05/28/2008, -0/+1Divide by it . . . see what happens.
  • somedirtbag, on 05/28/2008, -0/+4I've always said that a real computer needs a hole in the front to shove the trees into and an exhaust for all the black smoke to pour out.
  • srodolff, on 05/28/2008, -0/+4MORE STEAM!!!!......
  • Fetching more discussions...
    Show 51 - 86 of 86 discussions
  •  

    Login or register to add a comment

    Create a new account or login to join in the conversation on Digg. You'll also be able to Digg stories to help promote things you like.


© Digg Inc. 2008 — Content posted by Digg users is dedicated to the public domain.
DIGG, DIGG IT, DUGG, DIGG THIS, Digg graphics, logos, designs, page headers, button icons, scripts, and other service names are the trademarks of Digg Inc.