Re: [exim] attempting to change smarthost

On 2008-03-10 at 14:56 -0700, dpalom@??? wrote:
> With exim4 pointed at the new smarthost, I see the EHLO from my machine, and
> code 250 (OK) response from the smarthost. However, right after that my
> machine sends a "STARTTLS" command to the smarthost, the smarthost sends
> back a code 220 (Ready to start TLS), at which point it just sits there with
> no further progress. I assume it times out eventually.

Known problem with Exim/GnuTLS on Debian with entropy exhaustion; see
/usr/share/doc/exim4/README.Debian.gz -- using hosts_avoid_tls = * on
the remote_smtp transport, as Renaud suggested, works around this.

If you never want TLS to the new server, on that server find the
"tls_advertise_hosts = *" and remove it.

If Exim is stalling, I believe that suggests that it's trying to
generate the Diffie-Hellman parameters itself, which you can avoid by
having them created externally. According to README.Debian, you just
need to install the "gnutls-bin" package and then the daily cron-job
which was nuking the files to let Exim re-create them will instead
re-create the fresh parameters itself.

-Phil