Subj : Binkley - no modem & BinkD
To   : Bob Jones
From : Paul Lentz
Date : Mon Jan 14 2002 01:49 pm

Bob Jones wrote in a message to Paul Lentz:

 PL> The only problem I have with sharing things 
 PL> that way is that I sometimes send out uncompressed 
 PL> packets as I toss mail on my system in stages with 
 PL> steps in between and it goes out before the "SQUASH 
 PL> PACK" step gets done.

BJ> I don't see this "problem" because I run squish in single pass mode.
BJ> Why are you needing to run in multi-pass mode?  [Just curious....] 
BJ> Yes, BinkD can start up fairly quickly.

I run binkd as a full server, so it doesn't even need to start. It grabs the
..out file and fires up a session before I can blink.

On a typical mail processing session I run...

an areafix check
a raid check
then a SQUISH OUT for anything I night have processed in those two then a 
normal SQUISH IN OUT to process the mail packets I received (I used to
    run GMD between the in and the out... got to where I hated it) then I run a 
Bad Mail checker that strips ECHOMAIL messages out of my netmail
   area (I've found that of all the things that some misconfigured link can
   throw at you, this is the most devasting and can cause terrible-terrible
   loops and it only take a few seconds to check)
another areafix check
another raid check
then I finally do a full SQUISH SQUASH PACK to send everything out.

I think that's how I do it (????) :-)

BJ> As a suggestion, change your mode for the nodes in question from
BJ> crash to direct or hold, and then in your last squish operation use
BJ> the change command in squish's route.cfg to change the packets back
BJ> to crash.  This might solve your "problem"....

I actually thought about that... but really the 5 or 10 raw packets I send to 
my upstream link are usually small with a few messages each, so it didn't seem 
that important to deal with. Maybe...


*Paul*
--- timEd/386 1.10.y2k+
 * Origin: Dumb Guy's!!! (1:124/5025)

.