[sdiy] the list, the new machine, etc
Magnus Danielson
magnus at rubidium.dyndns.org
Mon Mar 22 21:47:46 CET 2010
Ben Lincoln wrote:
> FWIW, I'm getting the messages in the correct order and without any
> more delay than prior to the change.
>
> I know that Graham Atkins checked with his ISP and they claimed this
> wasn't the case, but a lot of the behaviour sounds like what I'd
> expect from a misbehaving spam filter. E.g. in the case of not seeing
> your own messages until someone else has replied, I could see a filter
> assuming that if a message came from a return address purporting to be
> its own domain, but sent through a mail system on another domain, then
> it must be spam due to the forged From: address.
>
> In the case of messages coming in out of order, could that not be due
> to something like "greylisting", where the recipient's mail server
> refuses the first n connection attempts by untrusted sending systems?
> When the connection attempt was finally allowed, I could see the
> sending system pushing through the newest messages, then checking its
> queue to see if anything else was pending and then sending that too.
Not if the grey-listing is behaving properly.
The one thing that one can do to analyze the situation is to check the
headers, where you can see the time-stamps as the mail has jumped
between the servers. That could be a good lead on where the problem is.
> A *lot* of the larger email providers (Google, etc.) seem to have
> stepped up their filtering of mail from small(er) domains as well. If
> I send an email to friends with GMail accounts and it includes a link,
> there's a reasonable chance it will end up in their spam bins.
SPAM-filtering is a war. There will be some casualties.
> Anyway, just a guess - but since I make sure there is no ISP-side
> filter in front of this account *and* I get the messages correctly,
> that's where I'd put my money.
I put both grey-listing and spam-filtering in front of my email, and it
seems to work well enough. With increased levels of paranoia oddities is
know to occur from time to time.
Cheers,
Magnus
More information about the Synth-diy
mailing list