The Market Ticker
Rss Icon RSS available
2021 NY Resolution: Take your mask and shove it up your ass.
You are not signed on; if you are a visitor please register for a free account!
The Market Ticker Single Post Display (Show in context)
User: Not logged on
Top Login FAQ Register Clear Cookie
User Info Covid-19 -- A White Paper - To @RealDonaldTrump and @CDC; entered at 2020-12-02 17:47:26
Tickerguy
Posts: 171473
Registered: 2007-06-26
Yep.

They "re-fetch" the header information EVERY TIME someone looks at their timeline; the software here sees that as an attack since the same IP address is repetitively hammering the **** out of the server for the same document, and when it detects it it drops a permanent IP ban on the source address.

Since I have no idea what IP address they're coming from I can't go look for and remove it either, not that it would do me any good since it would just go right back as soon as they do it again.

What they're doing is complete horse**** -- the correct approach is to grab it once when the post is composed and store it on THEIR END. If they want to check it once every hour or every day to see if it's changed and update it, that's fine. But to hammer the source on every VIEW is obscenely abusive and were I able to send a W88 back down the wire to them I would. The load they are putting on other people by doing this ought to get them blackballed by content providers EVERYWHERE.

The way the AKCS code is written one of the very first things it does is check that table which is VERY fast; if it finds a blackball entry you get bounced immediately before it even attempts to figure out what you want.

I gave some contemplation a while ago to modifying the code to instead stick a blackhole entry in the IP forwarding tables. That would reflect the abuser's bull**** back on THEM in that it would hang the process on their end for 30 seconds or so for each attempt until their SYN times out; if their code handles that sort of thing poorly it could conceivably cause the abuser's machine to hang or become unusably slow. I decided against it, but it's not very hard to do; I'd just have to write a small wrapper to handle it in a secure fashion with some guards in the code against it doing bad things because access to that table is privileged.

Last modified: 2020-12-02 17:55:18 by tickerguy

2020-12-02 17:47:26