"Killed" Message On Linux (Running From BASH)

[Hal Vaughan]

I've tried Googling this and get a high noise-to-signal ratio.

I'm running Perl scripts on Linux (Debian, Sarge), from the command line (in
bash), and I get a "Killed" message. I can't find enough to verify this is
coming from Perl or bash or where. Sometimes the program runs fine,
sometimes it doesn't. When I run it with one data file that didn't work
before, it runs just fine the 2nd time. The only module I'm using that
isn't mine is Term::VT102, and I've grep'ed and found NO occurance of the
word "Killed".

I'm used to getting crashes that tell me the line numbers and errors, but
not just a simple "Killed."

At the least, I'd like to know what is issuing this "Killed" message (is
this what happens when another program kills a Perl script?), or what could
be causing this.

Thanks for any help!

Hal

 

[Lars Kellogg-Stedman]

At the least, I'd like to know what is issuing this "Killed" message (is

this what happens when another program kills a Perl script?), or what could
be causing this.

Some shells (e.g., bash) print this message if a process dies due to
receiving a SIGKILL. Try this:

In xterm 1, run:

sleep 30

In xterm 2, run:

killall -9 sleep

You'll probably see the message "Killed" in xterm 1. I'm not sure what
would be causing this -- it is unlikely that something is sending
spontaneous KILL signals to your processes, unless (a) you're running
extremely low on memory and (b) the Linux OOM killer kills processes in
this fashion.

-- Lars

 

[Gregory Toomey]

I've tried Googling this and get a high noise-to-signal ratio.

I'm running Perl scripts on Linux (Debian, Sarge), from the command line
(in
bash), and I get a "Killed" message. I can't find enough to verify this
is
coming from Perl or bash or where.

Large process running out of virtual memory?

gtoomey

 

[Hal Vaughan]

Large process running out of virtual memory?

gtoomey

That was my first thought, but it's one Perl program, on 1.5 GB of memory,
that had run before, a number of times, on a system with 1/2 that amount
(and this system has fewer daemons and such running). I would also think
if that were the case, running it under the same conditions every time
would produce the same results. It doesn't.

Hal

 

[Hal Vaughan]

Some shells (e.g., bash) print this message if a process dies due to
receiving a SIGKILL. Try this:

In xterm 1, run:

sleep 30

In xterm 2, run:

killall -9 sleep

You'll probably see the message "Killed" in xterm 1. I'm not sure what
would be causing this -- it is unlikely that something is sending
spontaneous KILL signals to your processes, unless (a) you're running
extremely low on memory and (b) the Linux OOM killer kills processes in
this fashion.

-- Lars

Aha!

That's it. With that info, I tracked down a program that wasn't supposed to
be running (I'm not finished testing it yet), that tracks down and kills
certain programs if they don't register in a database first. Somehow a
stray instance was still in memory, killing other programs.

Thanks!

Hal