On Wed, Jul 30, 2014 at 04:38:39PM -0700, Todd Lyons wrote:
> On Sun, Jul 27, 2014 at 8:02 AM, Russell King <rmk+exim@???> wrote:
> > I know that 4.69 is an old version of exim, but... I'm seeing some
> > weird behaviour with it.
>
> TL;DR, I agree with your hunch that it's an old-exim-version problem.
> A fair amount of things have been touched WRT retry logic since 4.69.
>
> > The machine in question acts as a backup machine for another computer.
> > It's setup such that each night, it powers itself on, transfers the
> > data, archives it, sends a mail and powers off. Once a week, it
> > remains on for a 24 hour period.
>
> Can you verify the clock accuracy during the duration of each of the
> periods? Running ntp on it? Setting the time before exim starts?
I'll re-check this Sunday. I think I did set NTP up on it to ensure that
the RTC remains synchronised - in my experience, unsynchronised PC RTC
clocks tend to have fairly horrid drift.
> I looked through the ChangeLog to see if there was something that
> sounded like it addressed this. I found the following in the git
> history that seemed slightly relevant:
>
> 4.82 release: Fix ultimate retry timeouts for intermittently
> deliverable recipients.
>
> ...which sounds very strongly like it was a bug that was fixed in 4.82
> last year.
Thanks - I'll try putting 4.82 or later on it depending on what I can
back-port to F14 (the 4.82 packaged in F21's RPM requires systemd, and
the system doesn't have systemd on it).
> It took me a few reads to grok this. The intermittent connectivity is
> the DSL line that the host running Exim 4.69 uses for internet
> connectivity. So that's the reason for the DNS failures mentioned
> above.
Sorry, and your interpretation is correct.
Todd, many thanks for the response, this is exactly the information I
was hoping for.
--
Russell King
