diveintomark/.org

I noticed earlier tonight that mark's blog was down... didn't think too much of it until I later found this. I can't help but wonder if it was wordpress or his hosting that was unable to handle it.

Scott Hanselman blogged earlier this year on how dasBlog was able to handle the load just fine when he got /.'ed. - but of course that was just an individual page, and not a whole book...

Hopefully Mark will get some more book sales out of it, and not just a down site...

TrackBack URL for this entry: http://www.hutteman.com/scgi-bin/mt/mt-tb.cgi/174
Comments

His site's been down for me for quite a while... months, even. No idea why, as I was even unable to access it when others must have been able to. I could see from Google's cache that he's been updating, but haven't been able to see it myself at all.

Posted by Ryan at September 9, 2004 12:20 PM

Well, its the hosting. Once your website is /.ed, your web server better be upto it or it'll kneel down. Its got nothing to do with WP, as it sure can handle large amounts of hits. Not that you as an MT user will understand. ;)

As for Ryan not being able to get to Mark's website, well Ryan, you better check his website using another ISP b'coz as far as I know, its a problem with your current ISP & not Mark's webserver.

Posted by Amit Gupta at September 11, 2004 1:46 PM

A few months ago, every wordpress installation had about 2, 3 seconds to just _generate_ the front site. That step is that painful rebuild process on a MT installation - painfull for the webmaster. Users encounter no build process. I checked on a few Wordpress users and theyr loading time measured 2 times. This is the build process time as written at the bottom of the page. Amit Gupta prevents that build process time to be published.

http://www.onefinejay.com/ (0.2sec, 0.3sec)
http://photomatt.net/ (0.2sec, 0.5sec)
http://www.treygivens.com/ (8sec, 5sec)
http://www.mindofmog.net/ (3sec, 3sec)

you do the math on cpu usage if you have real traffic.

Posted by RaphaŽl Balimann at September 14, 2004 5:34 AM

I do have real traffic, and I host plenty of other sites that do as well. My experience in /. situations has been Apache has reached its maxclients far before WP was causing any noticable load on the system. This was also before the Staticize plugin which cuts out all database accesses from a cached page. WordPress by itself is very fast so often when you see a slower installation it's because it's on a terrible host or the person has installed a lot of heavy plugins. Also the build time on MT isn't just on the webmaster, it's on everybody who leaves comments as well. I've left comments on MT sites before and it's been *minutes* before my browser refreshed. I don't know the details of Mark's hosting, but I imagine the traffic of a Slashdot review would be trying for any system.

Posted by Matt at September 14, 2004 10:55 PM

I didn't want to insult WordPress in any way. I think the code quality is pretty high, and I love it's website. My point really is on it's dynamic nature. Rebuilding a site for each PageView makes sense by this formula:

Builddynamic if AveragePageAffectedPerModification > AveragePageViewsUntilModificationOccurs

I hope this makes sense to you. If i'm wrong, please tell me so - it's more of an essay I just thought up. I supressed details like PageFileSize for the sake of clarity. 'Apache maxclients vs WP load' - I see a bigger problem in usability of a average surfer (adding 3 seconds to every page view time) than in rare /.tted Websites. I just surfed on about five WP sites measuring load time. What will the rare really long times be? What is the average commenting time (a rare action) on a MT installation?

Anyway, MT commenting really freaks me out ;) But that is more of a implementation problem..

Posted by Raphael at September 15, 2004 6:07 PM
This discussion has been closed. If you wish to contact me about this post, you can do so by email.