mass-join flood detection + prevention

[angang]

hi,

in sentinel.tcl, do we have any mass-join flood protection ?


Definition of mass-join - massive joins in a very short time interval. ie 10 joins in 3 seconds.

I know there'll be conflict if there's a netsplit happens, since when the server is reconnect back , a similar situation will occur.

So if the script is intelligent enough to detect , when there's a netsplit , turn off the mass-join protection. When the split is over, the script will turn it on again.

I dont know about tcl scripting, and not sure if it's hard to achieve.

Anyway it's just my suggestion Thanks for reading it.

[slennox]

hi,

in sentinel.tcl, do we have any mass-join flood protection ?

This is the purpose of channel limit scripts. Limiting is ideal because it's robust--simple and relatively non-reactive. At least until ignorant people start setting 30 second timers on the +l interval (not my fault).

One thing they are missing (chanlimit.tcl at least) is a "roof" that puts an upper limit on the +l setting.

[angang]

yes, i did try limit tcl , but at the end i still remove it bcos things happen when your bot is down.

And when there's a netsplit, the chanserv will always set the +l according to the number of users on the bot's server ... blah i dont think i can explain this situation well , in short, the limit.tcl causes alot of problem

[Alchera]

When chanserv returns from a netsplit it always resets the limit to whatever it was when the split occurred. If the script is properly configured all users will reenter without any dramas as the bot re-adjusts the limit.

[greenbear]

not all networks have services, though

[Alchera]

And when there's a netsplit, the chanserv will always set the +l according to the number of users on the bot's server