hubcap@hubcap.UUCP (Mike Marshall) (11/13/87)
Greetings...
I am trying to analyze what is going on with the mailer on the machine
I administer (a 780 running Ultrix).
Although mail in general to and from my machine seems to go fine,
it seems to me that some mail is backing up in /usr/spool/mqueue for no
good reason. So, I've been "twiddling with the knobs" to see what
happens.
There were three or four pieces of mail sitting in mqueue last night,
all had lock files associated with them. I had reason to belive that
the lock files had been there for some time and that they were causing
sendmail to ignore these mails whenever it woke up (every ten minutes).
We have Internet connectivity with the hosts that these mails are
destined for (I can "telnet host.name 25" to these places most of
the time with no problem), but the mail sits on the queue marked
"(Deferred: Connection timed out)".
So, I deleted the lock files and waited for sendmail to wake up.
Sendmail woke up, put a lock on one of the pieces of mail, failed
to deliver it, left the lock file there and went back to sleep.
Ten minutes later, sendmail woke back up, locked the next piece of
mail, failed to deliver it, left the lock file there and went back
to sleep. The point is that pretty soon these pieces of mail all
had lock files and sendmail was ignoring them.
I can understand why, with the variable load of traffic on the Internet,
that SMTP connections with a particular host would time out sometimes
and be successful other times. But it seems to me that sendmail (or someone)
ought to get rid of lock files after a failed delivery attempt. Plus, the
time stamps (that I see when I say ls -l lf*) on the lock files don't make
sense. Last night I deleted all the lock files around midnight, and this
morning there were some lock files that were time stamped around 8:00pm
last night.
If you have read this far and think you might have some suggestions (nothing
lewd :-) pointers or info that might enlighten me, please drop me a note!
advTHANKSance...
Mike Marshall hubcap@hubcap.clemson.edu ...!hubcap!hubcap