Messages queued; SMTP relay failed; sendmail not running as daemon

[A. Daniel King_1]

Hi, folks.

Here's a theoretical question for you...

1) Sendmail is not running as a daemon.
2) Messages sent via sendmail through a SMTP relay work fine.
3) The SMTP relay fails, and the messages queue up on HP-UX sendmail.
4) HP-UX is rebooted while the SMTP relay is down.

What happens to the messages waiting in the queue? Do they continue attempting to send (until or after the reboot)?

Is there a way to keep sendmail trying on these messages without running as a daemon?

Are there variations in behavior between the 8.x versions of sendmail in this regard?

Discussion is welcome.

Thanks.

[Jordan Bean · Accepted solution]

They stay in the queue until a queue processing session is invoked (sendmail -q).

You may find this thread useful:

http://forums.itrc.hp.com/cm/QuestionAnswer/1,,0x94c40bce6f33d6118fff0090279cd0f9,00.html

[Misa]

Well, I made your theoretical question practical and tried it with a 10.20 server (sorry, it was what I could get at the moment).

I turned off sendmail (/sbin/init.d/sendmail stop and vi /etc/rc.config.d/mailservs). I then generated some normal outgoing mail, watched it process through, and verified delivery.

I ifconfig'ed all the network interfaces down and generated some more normal mail. It queued and stayed there. After I rebooted the machine, the messages were still queued. I then put in a cron job to process the queue (/usr/lib/sendmail -q) and the messages were delivered as expected. Frequency of the cron job would depend on how critical the data is, how large your /var is, and the like.

HTH,
M????a