error from wall command

[Fred Martin_1]

Consider the following root crontab entry:

40 00 * * 2-6 /usr/sbin/wall < messagefile

Intent is to display the contents of messagefile to screens of logged-on users.

Problem is, occasionally (every few months) root will get the following mail:

From: root
Subject: cron

sh: 18465 Alarm call

Anyone know what would cause this? Why would it not happen every time the thing runs?

[Brian M. Fisher]

There is a process hanging around on your system that wall(1)
identifies as a user process. Wall tries to send it a message,
covered by an alarm call so it does not hang. The message fails, the
alarm goes off, and you get the alarm call error message.

You should reboot your machine to get rid of the defunct process.
Sometimes reading from or writing to the offending pty fixes it,
for example, "cat /dev/pty/ttyp2 >/dev/null" or
"cat /etc/hosts >/dev/pty/ttyp2".
The above info is from Document#MM01081735
http://us-support2.external.hp.com/cki/bin/doc.pl/

Brian
<*(((>< er

[Alex Glennie]

An alarm call message is normally the result of a "wall(1)" command. Wall expects a
response from certain processes (such as login's). It starts a timer,
and if the timer times out before any response is received from the
process, you will get the "Alarm call" message. The five-digit number
is the Process ID . This is nothing to be concerned about - as to why you get it intermittently, zombie login process, high network load some thing like that perhaps ?