|
Neohapsis is currently accepting applications for employment. For more information, please visit our website www.neohapsis.com or email hr@neohapsis.com |
Re: Increasing retry time for a single domain
From: Ace Suares (lists
suares.an)
Date: Mon Dec 11 2006 - 22:15:23 CST
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]
Excuse me for appearing not smart. I am trying to figure out what the best
thing is to get rid of this problem. There are a lot of things I do not
know about Postfix, hence the stupid questions.
To (partly) answer Wietse Venema's question:
> Why don't you follow the advice posted here twice (!) to
> set up a dedicated mail delivery transport for broken servers.
> I'm not going to repeat that URL another time.
First of all, I didn't get that the slow transport limits the number of
processes to only one. (How could I have missed that!).
Now I understand that setting the concurrency_limits still spawns 50
processes and thus lots of connections.
Secondly, it seems that I need to add the slow transport for quite a lot
of domains - the omroep.nl servers hosts I don't know how many domains,
at least 20 or so. And there are other mailservers, that host several
other domains, too, so I guess with the 'slow' transport solution
advertised here indeed more than once, I need to check the logfiles and
add slow transports to many domains, and then if new members subscribe to
the list, some of them again will need to be added to the slow transport.
Is that right ?
I kinda had hoped that there would be some universal or automagical
setting that solved this recurring problem. Apparently, there is no such
solution, and I need to update the slow transport for each domain that
needs it.
Another issue with the slow transport seems to be that if a certain
mailserver changes policies, I would never know of it. So maybe today I
add 20 domains to 'slow' and then in a month, they reconfigure their
mailserver and I will still send the mail very slow, instead of nice and
fast like Postfix is supposed to do and does do for many other hosts.
Also, I see that in the advertised URL, one part of the proposed solution
is to increase the destination_recipient_limit, which seems not to be
working for that specific mailserver.
Once again, forgive me for being non-smart, I totally missed the benifits
of the slow transport but now I see it clearly.
Maybe I should not mind the entire problem, and put all settings back to
normal, like it was, and let those mailservers defer mail for hours and
hours and just go with the flow.
Thank you all for your great help and explanations! I learned a lot, and
if anyone comes up with other ideas about this, I'd be interested to hear
them.
Cheers,
ace
PS I am *not* being sarcastic.
- Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]