-7

From Update on the ongoing DDoS attacks and blocking Tor exit nodes

My IP address was banned on March 13. But I wasn't using Tor (and I'm not a Tor node.)...

Too many requests error page

Of course, my first reaction was to sent mail as instructed, but as this happened on Saturday, I've received back quickly a message explaining that there are no support on week-end....

For my understanding: I was working on SO the day before (March 12, but I don't remember the clock time precisely) when SO was coming down. I was writing a post... After some manual retries, I've initiated a little Bash loop running wget on Stack Overflow until the standard SO's title could be read, something like every 120 seconds... So I would imagine that this did not look like a regular human interacting with the site and thus the ban was legitimate.

But how long is this going to last?

I've finally used ssh -D somewhereelse as a proxy SOCKS in order to be able to terminate my post. But I was surprised about the duration of the ban. After I stopped my wget loop, test from browser after one hour, then one day was unsuccessful.

After more than one day, I've finally sent my question by e-mail, to team At SE.com, but that was on Sunday... Do some test on Monday, and finally I've stopped my SOCKS proxy Tuesday, maybe something like 72 hours after I was banned.

Is this correct, or did the ban system use a progressive ban time regarding number of tries I submitted?

9
  • 1
    I have a proper uptime monitor polling 2 SE chat servers, mse, Su and so every minute with no issues. I don't think doing so every 120s would cause a too many requests rate limit Commented Mar 18, 2022 at 15:46
  • 8
    ...Have you tried following the advice in the note there that says "if you believe this to be in error, please contact us at [email protected]"? I had this happen to me a bit ago (caused by a request-heavy userscript), and they got back to me within a few hours with detail about the block. I imagine that this issue isn't really something that anyone but staff can help with.
    – zcoop98
    Commented Mar 18, 2022 at 15:57
  • 4
    These general rate-limit IP blocs typically last 10 to 15 minutes. If your use is/was legitimate and the general IP block rate-limit lasts substantially longer than that, then you are best off using the "Contact"/"contact us" link to contact Stack Exchange. There have been times in the past where they have misconfigured the rate limiting such that it never expires for an IP once triggered. You're way past the normal timeframe for this, so contact them.
    – Makyen
    Commented Mar 18, 2022 at 16:01
  • 11
    A single request every 120 seconds shouldn't be sufficient to trigger the general IP rate-limiting block. It usually takes substantially more traffic than that to trigger these blocks.
    – Makyen
    Commented Mar 18, 2022 at 16:05
  • @zcoop98 Of course, unfurtunely, This happen at saturday, I've become quickly a message explaining: no support on week-end... Commented Mar 20, 2022 at 6:31
  • Are you on some kind of public network? (School / work?)
    – Cerbrus
    Commented Mar 21, 2022 at 8:20
  • What is the total number of requests? Opening 50 pages with a 3 seconds delay has never been a problem (daily). The same with 2,000 pages (Wget'ting) with a 16 seconds delay (9-12 months interval). Commented Mar 21, 2022 at 15:10
  • @This_is_NOT_a_forum Interesting! It's maybe consecutive to SSL fails and some bug in my script... I realize when connection fail because of security fail, the 120 second delay won't be executed... I have to review my shell and try to build a fake server to make some tests... Commented Mar 21, 2022 at 16:25
  • @F.Hauri This is why every BASH script I ever started eventually got redone in Python as the complexity grows. It's not a full fledged programming language and always seems to have weird errors. Bash is just more convenient for running system commands, but you can do that in Python with a little finagling. Commented Mar 29, 2022 at 19:55

1 Answer 1

10

Mea culpa

From Log in to a site (Stack Overflow) by Bash, I've created a Bash script with some fun features: getSo.sh.

I just corrected a bug now: my endless loop (while true) made a 120 seconds delay, but only when a correct read was done.

The bug was: when connection fail, script loop immediately in order to try again. As SO was restarting, requests were unsuccessful and the script did a lot of retries.

The more: once banned for the first grace period, as I was sleeping, the script did continue to make non temporised tries...

I understand now the reason I was banned for a long time.

3
  • Its a bit more work, but have you considered setting up a 'proper' monitoring thing? :D I've been running sourcegraph.github.io/checkup for some semi internal services. In theory I've upgraded to uptime kuma github.com/louislam/uptime-kuma which is nice but checkup is easier/better for linking to other people since its dashboard has graphs . Needs a web server and stuff but Its a much better setup Commented Mar 25, 2022 at 7:48
  • 1
    Maybe just don't create such scripts to begin with? Stack Overflow isn't a game. Using script is always a risk, unless you really need it for some reason, better not take such risk. Imagine SO didn't have such good protection. Your script would have been actual DDoS attack on SO. Funny, right? Well, no. Commented Mar 27, 2022 at 7:09
  • 4
    @ShadowWizardSaysNoMoreWar No, just a plain DoS attack. Not distributed. And it should have generated so little traffic that SO would barely even notice it. It could be a very different matter if many other people also used that script, but in that case it would have a more rigorous development process, and someone would quickly point out the bug, right? Commented Mar 29, 2022 at 15:51

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .