By any chance do you have a firewall (Cisco ASA for example) that you block all or most ICMP?
A few years ago, I experienced issues with a few particular remote sites and their erratice mail delivery to us. We had blocked most ICMP types at the firewall for PCI compliance. We relaxed the rule and blocked just a few specific ICMP types (the time query ones) and all of a sudden those issues went away. It must have been breaking path mtu discovery. ...Todd On Thu, Feb 23, 2012 at 9:24 AM, Scott Neader <[email protected]> wrote: > Thanks, David, I'll send it to you direct. > > My concern on the timeouts is: > > 1) I have seen in the past that all of my Exim sockets can be consumed by > misbehaving mail servers (or spam zombies) and thus we defer mail. I'm > open to discussion on this, if I'm doing something wrong, or > misunderstanding. > > 2) The far-end customer (using EdgeWave) is reporting SOME fatal errors. > Most messages are getting through, but the reason I found the problem is > after being contacted by their ISP asking why we aren't accepting some of > their mail. > > 3) We have rate limits set up for misbehaving mail servers, and these > timeouts are counted toward the rate limit. I will need to research to > find out how to stop counting timeouts toward rate limits, if I am to start > ignoring these timeouts as non-issues. > > 4) It seems most servers with this timeout problem are either EdgeWave mail > servers, or spam zombie home computers. I'm hesitant to ignore these > timeouts, but if the Exim community feels that I should, then I will. > > Thanks!! > > - Scott > > On Thu, Feb 23, 2012 at 1:49 AM, David Woodhouse <[email protected]>wrote: > >> On Wed, 2012-02-22 at 10:36 -0600, Scott Neader wrote: >> > Are you willing to look at the cap file from their side, to see if they >> are >> > doing things right? I'd like to be able to tell them... look, RFC XXX >> says >> > after we send the 250 OK, you should send a QUIT but your cap shows you >> are >> > not..." (or whatever) -- but I'm just not knowledgeable enough. >> >> By all means, send it my way. Note that the only "problem" this causes >> is an extra line in your log and a small amount of memory used while >> Exim is waiting to die, right? >> >> -- >> dwmw2 >> > -- > ## List details at https://lists.exim.org/mailman/listinfo/exim-users > ## Exim details at http://www.exim.org/ > ## Please use the Wiki with this list - http://wiki.exim.org/ -- SOPA: Any attempt to [use legal means to] reverse technological advances is doomed. --Leo Leporte -- ## List details at https://lists.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://wiki.exim.org/
