Subj : binkd connecting to own AKA
To   : Alexey Fayans
From : mark lewis
Date : Mon Jan 13 2020 11:48 am

  Re: binkd connecting to own AKA
  By: Alexey Fayans to mark lewis on Mon Jan 13 2020 19:10:10


 ml>> how do you prevent binkd from connecting to one of its own AKAs?

 ml>> i see this from time to time and it results in a loop of the traffic
 ml>> being attempted to be sent...

 AF> The easiest way is to not create outboud for own AKA. Or if it is really  
necessary, create it with _hold_ flavor.

not really viable options... especially when it may be due to what may be a 
routing problem on one or both sides of the connection...

in the specific case i'm looking at, the problem is a netmail from a node's 
point to the node but the node forwarded it here... apparently there is no 
automatic routing in place with the software i'm using that routes netmails 
from points to their boss node... in this instance, there was a routing problem 
on the boss node which resulted in the point netmail being sent here... that's 
been fixed but my system was following its routing table and sending all mail 
for that boss' points to the NC for forwarding to the node... since i'm also 
the NC for that net (and a few others) binkd was calling itself which just 
seems weird...

so i hope to draw a simple diagram for what happened...



point system (1:123/xxx.y)
  |
  V
boss system (1:123/xxx.0)
  |
  V
my HUB in different net (1:3634/12)
  |
  V
my tosser on 1:3634/12 routes to net123 NC 1:123/0
  |
  V
my binkd is 1:3634/12, 1:3634/0, 1:123/0, 1:18/0
  |                                  ^
  V                                  |
my binkd calls itself over and over --



so far, i've fixed this case by deleting the ?ut file and adjusting my routing 
to specifically routes this node's point netmails to the node but to have to 
specifically set routes for all nets' nodes' points to avoid this if i'm also 
the NC of that net is painful to say the least :/

it may also be a tosser problem not recognizing messages that may be routed to 
one of its AKAs...

it just seems that a mailer should recognize mail destined to itself and refuse 
to make the attempt with an appropraite log notice... possibly in the .try file 
that monitoring software may use to display the last attemptted connection 
status...

i can imagine what a single node system on POTS would do trying to call 
itself... a multinode POTS system calling one of its other lines might be 
noticible due to modem sound but this is TCP/IP so there's no modem sound to 
possibly hear...

i'm still researching prevention possibilities... routing, tosser AKA 
recognition, both(?)...


)\/(ark
--- SBBSecho 3.10-Linux
 * Origin: SouthEast Star Mail HUB - SESTAR (1:3634/12)

.