If memory serves me, cron jobs do not get their own tty.
tty based file handels can get flaky in cron jobs.
I have run across situations in Solaris and AIX,
where the the last filehandle accessed --redirected to /dev/null or not-- by a cron script is the one that gets the output.
I am not sure why or what causes it. 

I would force the STDOUT and STDERR of each command in the job to $0.$DATESTAMP log files until the error(s) reproduce.
that way it can "hopefully" be determined where it is comming from and suppression or remediation can
take place.

Mike

On Tue, Jan 27, 2009 at 1:19 PM, <mmcgrath@redhat.com> wrote:

On Jan 27, 2009, at 12:23 PM, Ricardo Ichizo <n1ghtcr4wler@gmail.com> wrote:

Hi,

2009/1/27 Ricky Zhou <ricky@fedoraproject.org>:
On 2009-01-27 09:01:55 AM, Mike McGrath wrote:
...... done.
...... done.
This is just a random guess, but the text reminds me of:

receiving file list ... done

from rsync.  I wonder what could cause this to be printed if it were
rsync...


Yes, rsync always do that:

building file list ... done
...
receiving file list ... done

This makes me wonder if rsync is sending part of that message to stderr. I tried to reproduce it but have been unable to.  It just sortof happens sometimes.

   -Mike


_______________________________________________
Fedora-infrastructure-list mailing list
Fedora-infrastructure-list@redhat.com
https://www.redhat.com/mailman/listinfo/fedora-infrastructure-list