Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Instability Mystery Solved
#1
So, I just found the reason for our random periods where the site starts to grind to a halt when usage is normal...

Apparently certain search engines that will not be named when spidering us, do in fact run a 'show all posts by this user' on every post.

An explicit "do not do this" rule has been added to our robots.txt and the offending spiders have been blocked at the firewall for the next 24 hours until they have time to notice (and hopefully obey) the rules.
It's not having what you want - It's wanting what you've got.
Reply
#2
Damn you..

..Google
...Bing
..Yahoo!
..Dogpile
..Brainboost
...YaCy
..WebCrawler


etc.
Reply
#3
Good Eye.
Reply
#4
Oh, that explains a lot of things.
Reply
#5
Locked Wrote:etc.


Except for MSN most of them weren't even US.
It's not having what you want - It's wanting what you've got.
Reply
#6
Eos Wrote:Except for MSN most of them weren't even US.

Damn those Chinese spider crawlers. Rolleyes
Reply
#7
They still misbehaving? Cause SP is running so slow right now.
Reply
#8
SP was just flatout unreachable for me for a little while there... I'm assuming your "solution" didn't slow them down any Sad
Reply
#9
Rather certain ones of them are blatantly choosing to ignore their own stated guidelines.

And since they're assinine enough to use dozens of different IP addresses the only fix effectively blocks entire subnets.
It's not having what you want - It's wanting what you've got.
Reply
#10
Ugh, look at this crap:

June's logs for this one search agent: (number after colon is how many hits from specific bot that day)
 Spoiler

November's logs:
 Spoiler

And we know damn well Southperry's traffic hasn't picked up that significantly.
I've firewalled it, robots.txt'd it, and sent en email to it's support address with the 8mb of logs from the past 72 hours of just it re requesting the same threads once per post. Tongue
It's not having what you want - It's wanting what you've got.
Reply
#11
An average of ten THOUSAND hits a day? Sweet jesus.
Reply
#12
Justin Wrote:An average of ten THOUSAND hits a day? Sweet jesus.

And thats just from one search agent...no wonder things grind to a shuddering halt.
Reply
#13
Holy fucking pomegranate eos.
that's a lot of traffic
Reply
#14
Could it be some kind of attempt at DoS?
Reply
#15
Shidoshi Wrote:Could it be some kind of attempt at DoS?

A search spider that is also a DoS is counterintuitive, counterproductive, and just plain stupid.

A search spider that tries to bring other people to Southperry when the spider itself is making the site unreachable... that's a logic clusterf'uck.
Reply


Forum Jump:


Users browsing this thread: 1 Guest(s)