We were down for about an hour plus at about 6 PM Eastern time tonight. Our host, Bluewho said it was because of increased HTTP traffic that burdened the server. Despite having a lot of pages today (at least for us) – looks like 11,000 plus we are no where near our bandwidth limits for the month (we did about 750 megabytes yesterday – less than 1 gig and I don’t think we did more today actually).
I need to figure out whether it was just a lot of visitors today or whether something is goofy with POSTpn that is overtaxing the server. Only thing is I didn’t do anything “under the hood” with this most recent website change so I can’t think of anything that would be different between last month (when we were fine) and this month (when Bluewho cut us off). Any advice, speculation, diagnostic approaches, etc is appreciated.
Once upon a time, my site went down because I had used “all my webspace,” when it turns out all my webspace was being taken up by huge log files the host generated that I had no control over.
That clearly bites. I hope you gave that host the boot!
This isn’t a space thing – it seems to be a CPU thing – too much activity taxing the CPU.
The fucking web – love it, hate it. Not a ton of traffic from this exploit but someone may have been hacking or trying to hack into Comixpedia’s server. I don’t see any damage so I’m assuming (and keeping my fingers crossed) that it was a no go. Still I tracked it back to this strange website where this forum thread seems to be on point. I can’t tell if this website is evil or had same evil perpretrated on it.
Anyone care to translate this into old person english 🙂
Also this thread seems to have someone accusing the Visual Coders of trying to hack.
I guess some of this crap was some kind of PHPBB worm (I’m guessing) – they got this site:
and i guess it started going after Comixpedia. Luckily, we patched the forums, eh?
11,000 pageviews and 750 megabytes shouldn’t be anywhere near enough to crash your server. Hell, I do over 71,000 pageviews a day and over 5 gigs in transfers and I’m sharing a server with over 10,000 other webcomics and I’ve not had my site crash in about a year (thank you very much Kisai).
I know everyone always goes for the software in a case like this but it could be the hardrive is starting to crap out on you. That’s what caused the last great Keenspace Crash.
That thought has occurred to me. I don’t really know how many accounts are on the machine i’m on. Comixpedia has a virtual host account not a dedicated machine. But apparently Comixpedia was the “culprit” today according to the host.
I can’t imagine post pn is particularly friend on the server in regard to CPU usage, especially under higher visitor loads.
11,000 page views shouldn’t be a problem, have you thought about disabling the unessacary post pn features? Or maybe using a less bloated content management system?
Something like MoveableType would be good because it generates STATIC pages rather than pulling all the content from the database everytime.
Most large sites tag the approach of servering as much static content as possible, as in general its the database queries that start to eat the server (alive).
Sounds like the Santy Worm. I had this problem a few weeks back. It isn’t that you’re infected: it’s that infected servers are scouring the web trying to infect others. At one point, I had about 400 infected servers trying to infect talkaboutcomics.com. Not being infectable is nice, but it’s still a pain in the ass having all those fake http connections.
There’s some conversation about ways to block infected servers from making http requests on your server here:
Comments are closed.