[geeks] ADMINISTRIVIA: Changes to mail delivery policies

[Greg A. Woods]

[ On Thursday, March 14, 2002 at 08:04:39 (+1100), Scott Howard wrote: ]
> Subject: Re: [geeks] ADMINISTRIVIA: Changes to mail delivery policies
>
> Where the hell are you getting this from?

The "de facto standard" BSD TCP stack.  Though it seems I have an
off-by-one error in my reading of the code.....  :-)

It's actually four SYN packets before a connection times out....  Horror
of horrors -- another whole 48 bytes sent seconds after the past 48 bytes.

# tcpdump -vvv -n host 10.10.10.10 &
[1] 29163
# tcpdump: listening on tlp0
# telnet 10.10.10.10 25
Trying 10.10.10.10...
16:14:23.136364 204.92.254.15.49195 > 10.10.10.10.25: S [tcp sum ok] 2854565586:2854565586(0) win 16384 <mss 1460,nop,wscale 0,nop,nop,timestamp 0 0> (DF) (ttl 64, id 48218)
16:14:28.648997 204.92.254.15.49195 > 10.10.10.10.25: S [tcp sum ok] 2854565586:2854565586(0) win 16384 <mss 1460,nop,wscale 0,nop,nop,timestamp 11 0> (DF) (ttl 64, id 48296)
16:14:40.649512 204.92.254.15.49195 > 10.10.10.10.25: S [tcp sum ok] 2854565586:2854565586(0) win 16384 <mss 1460,nop,wscale 0,nop,nop,timestamp 35 0> (DF) (ttl 64, id 48470)
16:15:04.650494 204.92.254.15.49195 > 10.10.10.10.25: S [tcp sum ok] 2854565586:2854565586(0) win 16384 <mss 1460,nop,wscale 0,nop,nop,timestamp 83 0> (DF) (ttl 64, id 51503)
telnet: Unable to connect to remote host: Connection timed out
# 

Note that Postfix does exactly the same thing as telnet for the normal
case of a domain with no or one MX.

Still even if there are 44 SYN packets sent by some lame TCP stack whos
authors think they know better than the combined experience of all the
IETF, the total "bandwidth" used will still be nearly unmeasurable
unless it's Bill's ISP that gets disconnected.  You're not looking at
the whole picture!


> It's a mailing list. Get over it. (Repeating just to keep Amy happy :)

It's not just a mailing list -- it's all of Bill and Amy's e-mail, _all_
of the mailing lists they run, and all the outgoing maile for any other
users they might be hosting too.  While the reliability of lists like
geeks and even rescue might be irrelevant to almost everyone concerned,
I'm sure some of the other subscribers on the other lists will not be
quite so happy about the potential of their mail from those other lists
bouncing prematurely due to some mistaken idea Bill's got about how he
can improve his postmaster duties.

In any case the whole point here is that Bill has adjusted at least one
parameter(s) incorrectly and for all the wrong reasons and he will not
get the results he so dearly desires (and deserves, in this case).

-- 
								Greg A. Woods

+1 416 218-0098;  <gwoods at acm.org>;  <g.a.woods at ieee.org>;  <woods at robohack.ca>
Planix, Inc. <woods at planix.com>; VE3TCP; Secrets of the Weird <woods at weird.com>