Text 5385, 100 rader
Skriven 2012-12-26 19:32:18 av mark lewis (1:3634/12.42)
   Kommentar till text 5378 av Ward Dossche (2:292/854)
Ärende: Z4 is dead
==================
ml>WD> On a minute-by-minute basis my system calls Ioram Sete and Renato ml>WD>
Zambon (ex-Z4C) for over a month now with zilch results.
 WD>  
ml> if you were hitting my system like that, you'd end up in my perimeter
ml> firewall's block list pretty quickly...
 WD>  
 WD> I've never gotten that far even once ... the calls simply don't
 WD> terminate because the systems aren't there.

you are not understanding what i'm saying...

the first attempt to connect, whether successful or not, starts the count and
the timer... each connection attempt after that increments the count while the
timer ticks... at X connection attempts within Y time, an alert is raised and a
block rule is generated... from then on, you won't connect at all even if the
system is up and ready to receive... why? because your system is blocked with a
"drop" rule which means "no response at all" so all attempts wil garner the
same apparent "nothing is up on the remote end" failures...

i tried to point this out to you previously when you posted a "connection was
refused" log entry... that was when i tried to point out that "connection
refused" message is not always indicative of the remote not being "alive"...
  
ml> ... especially since all those connect attempts could easily lead 
ml> to a DOS from logging them which could lead to running out of drive 
ml> space...
 WD>  
 WD> Don't become too melodramatic ... 

i'm not... i'm extremely serious and speaking from an IDS/IPS POV... this has
nothing to do with FTN stuff at all... it has everything to do with preventing
one's system(s) from being abused by others on the outside trying to force
their way in, one way or another...

 WD> my system gets ten of thousands of calls on a daily basis 

tens of thousands of inbound connections daily? really?? the shear number leads
to a huge doubt about that... sorry... while you may run a busy FTN system, it
is very doubtful that it is that busy with legitimate FTN connections...

 WD> and the daily log for all that is on average 1 megabyte. 

1Mb, is that all?? geez, my system is nowhere near as busy as yours and i see
FTN logs at least 2 or 3 times that size each month for just binkp
connections...

 WD> At that pace my system would run out of drive-space in 111 years. 
 WD> I'm certain similar parameters apply for you.

ummhumm... what are you running over there? multiple terabyte drives? my main
system, which handles all of my FTN traffic, has only 30Gigs of drive space...
please don't fall into the trap of thinking that everyone else is playing that
"keep up with the dossches" game ;) my main system was just recently upgraded
from a PII 300mhz 112M RAM to a PIII 800mhz with 512M RAM...

ml> ... if that were to happen, one might also be justified in filing a 
ml> policy complaint for excessivly annoying behavoir...
 WD>  
 WD> That would be fun ...

i'm sure it would be... just like when the one fidonet system figured out that
they could bring down another fidonet system by file attaching $CLOCK... that
even made the snooze...
  
ml> who are you to say at what "level" a system operates? you Z2 folks and
ml> your infatuation with points is appalling :?
 WD>  
 WD> A system listing "ICM" should live up to that. Nodes can be down
 WD> for short periods for whatever reason ... maintenance, back-up,
 WD> power out, software upgrade ... but it shouldn't last a month ...

agreed... to a point... see above... as the operator of my system(s), it is up
to me to decide what is abuse and what is not... polling my system more than
once an hour without delivering any mail is abusive in my world (right now)...
especially since my system only processed mail once an hour...

 WD> So if said node is only available during intervals, then the 
 WD> nodelist needs to reflect that, otherwise it is a point ...

bullshit... i have a point that is CM 24x7 unless the software it is running
shits the sheets for some reason that my scripts don't currently catch and
recycle... it should be a node by your reconing...
  
ml> you'd think but i have communicated with some latinos on various subjects
ml> and many of them are notorious for leaving stuff out that one would think
ml> would be included automatically...
 WD>  
 WD> I've also communicated with Gamey Garcia ... that was 10 years ago.

i don't know who that is... but i also wasn't talking about only fidonet
stuff... in fact, what i was speaking of was completely network security
related... you wouldn't believe the hen's teeth pulling that has to be done to
get all the necessary information from a system admin complaining about
problems with something... even with working crystal balls, (e)WAGs are only
successful < 10% of the time...

)\/(ark

--- 
 * Origin:  (1:3634/12.42)