Hello Jeremy / Phil,
Thank you so much for your replies. I've got answers below for all your questions and comments.
1) The missing log file error shows up for any of the items that show up in the exinext command such as the ones I showed below. It appears items that are in delivery mode are not showing up in the exinext command ( I verified by picking a few message ids from exiwhat and looking for them in exinext ). Items that are visible in the exiwhat command do have a message log I can visit using exim -Mvl <msg id>. It appears the connection from google is getting terminated.
Here's some sample output from exinext
Transport: alt1.gmail-smtp-in.l.google.com [173.194.193.27/69.160.42.53]:1cath8-000KiO-0r error -18: H=alt1.gmail-smtp-in.l.google.com [173.194.193.27]: Remote host closed connection in response to end of data
first failed: 02-Feb-2017 03:50:41
last tried: 02-Feb-2017 03:50:41
next try at: 02-Feb-2017 03:50:41
here's what I see for a message using Mvl that shows up in exiwhat
2017-02-02 14:03:55 H=gmail-smtp-in.l.google.com [2607:f8b0:400e:c04::1b]:25 Connection timed out
2017-02-02 14:13:58 H=gmail-smtp-in.l.google.com [74.125.28.26] Remote host closed connection in response to pipelined end of data
2017-02-02 14:15:01 H=alt1.gmail-smtp-in.l.google.com [2607:f8b0:4001:c0f::1a]:25 Connection timed out
2017-02-02 14:25:03 H=alt1.gmail-smtp-in.l.google.com [173.194.193.26] Remote host closed connection in response to pipelined end of data
2017-02-02 14:26:06 H=alt2.gmail-smtp-in.l.google.com [2607:f8b0:4002:c03::1b]:25 Connection timed out
2017-02-02 14:36:08 H=alt2.gmail-smtp-in.l.google.com [173.194.219.26] Remote host closed connection in response to pipelined end of data
And if I do an exinext on the msg id then I get a response that there is no retry data there. This is confusing because it clearly tried to deliver the message - since the connection timed out I guess it never recorded it? I'm not sure. This seems to be at the heart of the matter.
2) The VM platform being used by my hosting company doesn't allow me to update the clock - it has to come from the root server. I've put in a request to have it updated but my other idea is to add a time differential and update our items from the queue.
3) Yes I am concerned about inserting the items directly in the spool such as we are. I read a bunch of articles on how to migrate messages from one server to another which seemed acceptable and I didn't have any other ideas how to skip the overhead of SMTP protocol for each message over the network from our backend server to the exim mta servers ( tests running to get rid of everything and immediately put stuff in the queue still took about an hour and a half to upload an hours' worth of email deliveries - even at a low rate of 4,000 emails per day per ip ).
I've looked at Mike's recommendations for command line options but I can't find examples to help me see how to implement them so it's been a little slow going. Are there any examples out there of how to use the -bm, -bs and -bS options? ( right now I'm leaning towards -bs option to do buik SMTP inserts ). I've read
http://www.exim.org/exim-html-current/doc/html/spec_html/ch-smtp_processing.html and so I'm thinking I could just create these files with the appropriate SMTP commands and route them through standard input via the -bs option. That being said, one of the very important things is to make sure we can route it via a specific ip address on the machine. I'm not sure how we would do that with these approaches. My current thought is to use the authenticate_domain which would be tied to the RCPT FROM: header I believe.
Would this be a better option to get the items sent? I assume the routers will still be employed and remote_smtp transport will be used ( that's where I have my logging now ).
4) When we add something to the queue there are two current commands we try. Right now I kick off a new queue process ( -q option ) for each ip address that is getting new emails in it's queue. I've also tried exim -odb -M <msg id> to try to do an immediate delivery. We seem to get in this situation with the retries in either case.
Thanks, so much everyone!
Dan
-----Original Message-----
From: Exim-users [
mailto:exim-users-bounces+danliles72=gmail.com@exim.org] On Behalf Of Jeremy Harris
Sent: Thursday, February 2, 2017 6:59 AM
To: exim-users@???
Subject: Re: [exim] Retry rules don't seem to work in 4.88
On 02/02/17 09:09, Dan Liles wrote:
> I seem to be having a weird problem with retry rules in exim 4.88
>
>
>
> Right now when I send emails and I run the exinext command I see where
> they say the failure was the same time as the next retry.
Where there was a deferral or when it delivered ok?
> * lookup F,4h,1h
>
> * refused F,1h,15m
>
> * rcpt_4xx F,4h,15m;
>
> * * F,2h,15m; G,2h,15m,1.5;
Did you flush the retry DB after modifying the rules?
> The time on my servers running exim is way off - I can't set it
Why? That, IMHO, is a very foolish way to operate a system. Many many things make deeply hidden assumptions about things like time-and-date being correct and meaningful,
> Could that affect it? Remember I'm taking our messages and bundling
> them to be inserted into the spool for now so I set the timestamp in
> the file.
Exim assumes it is in control of its spool. You are very likely to confuse it. Don't do that.
--
Cheers,
Jeremy
--
## 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/
