Rebooting the Titan Network server

Hey all, you're beating our server up pretty badly today.

To try to alleviate the memory drain that the traffic is causing, we've added some more RAM to the system. However, for it to take effect, I need to reboot the server. Therefore, the Titan Network server, including all of the sites such as City Info Tracer, the Paragon Wiki, Faces, and the planners, will be offline for a couple of minutes until it comes back up.

Sorry for the inconvenience, but this should help get access back up to speed.

Noooooooooooooooooooooooooooooooo!













... well, ok. But just this once.

See my Guides and Tutorials

It's already back up. It might still be slow for a few minutes while everything loads and reconnects.

Were my microsoft framework errors and game crash when using glycerine related to that ?

All that traffic and still no one will update the Demon Summoning page. A little too much "Gimme!" and not enough "Hey, I should put some effort into this" by the users foor my tastes.

It's soooo sloooowww... We're still tweaking and kicking stuff now and then, so if you get any kind of server error, apologies in advance.

One thing that's obvious is that this release is like no other in history. I can tell you that we've never even been close to having this many connections happening constantly. I'm absolutely boggled at how busy the server is. This is bad for us, but really good for the game, methinks. People are obviously EXTREMELY excited about Issue 17.

I can tell already that when Going Rogue hits, I'm going to need lots and lots of aspirin.

Hey all, quick update. I found a setting that I think was a bit out of whack and tuned it to handle the massive number of connections we've been getting with the launch of Issue 17. I've been monitoring the server since I made the change a little while ago, and the performance difference is night and day. Everything is running, well, like you'd expect a web server to.

I really apologize for the long delays that some of you have been experiencing, especially during the peak hours, and especially right as Issue 17 launched. I'm not going to be so overconfident as to say that we 100% have it nailed down. ("Nothing could possibly go wrong now!") But I'm really optimistic that we have a much better handle on it.

For any other Apache web admins out there, read this. Much heartache and grief may be avoided, and it may help you greatly mitigate the so-called "Slashdot effect." Our KeepAliveTimeout was 15 seconds, the default when the server was installed, which optimistically assumes that your server isn't going to be balls-to-the-wall busy constantly pegged at the MaxClients setting. Now it's 2 seconds, which should be plenty of time after you get a page for your browser to request any additional resources (images, css files, javascript files) it needs. The number of active requests was literally cut in less than half of what it was.

Oh well, live and learn, I say! I'm just glad that I can actually pull up pages in a reasonable amount of time now, as I hope that most of you are.

Enjoy!