[plug] This Mailing list

[John Summerfield]

> On Fri, 28 Jan 2000, John Summerfield wrote:

> ---
> Sendmail sends the message to one reciepient at a time ... waits for each
> message to be accepted for delivery (or return an error) and then moves onto
> the next reciepient.
> 
> <Snip>
> Jan 28 08:50:27 pandora sendmail[6544]: IAA06544
> Jan 28 08:50:28 pandora sendmail[6544]: IAA06544
> Jan 28 08:50:28 pandora sendmail[6546]: IAA06544
> Jan 28 08:51:43 pandora sendmail[6546]: IAA06544
> Jan 28 08:51:57 pandora sendmail[6546]: IAA06544
> Jan 28 08:52:04 pandora sendmail[6546]: IAA06544
> Jan 28 08:52:04 pandora sendmail[6546]: IAA06544
> Jan 28 08:52:53 pandora sendmail[6546]: IAA06544
> Jan 28 08:53:26 pandora sendmail[6546]: IAA06544
> Jan 28 08:53:27 pandora sendmail[6546]: IAA06544
> </Snip>
> 
> As you can see the last person on the list was mailed 3 mins after the messag
> e
> was recieved by the mailserver.

Delays are likely caused by DNS lookups (both ends) and ident discussions. 
If you want the checking done, you have to pay the time.

 
> I can't show you the addresses, BUT you can take my word that serveral of the
>  
> list members are on the same host (infact 2 list members are on the same
> machine as the mailserver itself)

Read the tuning info beginning on p22 of the operations guide (sendmail 
8.9.3). There are delivery options that are likely to improve performance. 
One of the options is 'interactive delivery' which is sure to be slow.


I'm not so much defending sendmail as trying to educate people to 
disbelieve assertions unsupported with evidence. When evidence is produced 
(as you have, Tony) then we can evaluate it.

I don't think sendmail is as bad as people claim, and I don't believe many 
on this list will gain anything at all by replacing it; the act of 
evaluating alternatives itself is a pretty hefty investment (especially if 
it's not being undertaken by a hobbyist for personal use); recovering that 
investment would take quite some time. otoh I I don't believe it's perfect 
either, or necessarily better in any particular respect than those 
alternatives.
-- 
Cheers
John Summerfield
http://os2.ami.com.au/os2/ for OS/2 support.
Configuration, networking, combined IBM ftpsites index.