[rhn-users] Tar problem

Richard Riley rriley at procuri.com
Wed Apr 4 21:20:24 UTC 2007


I have a nightly backup process that uses tar writing the output to a
file on an NFS mount.  It had been working fine for weeks and suddenly
on Tuesday started failing with error code of 141.  Anyone know where I
can find out what the error code means.  I have tried a google search
and looking at tar source code, both to no avail.  I have another
machine running the exact same script also writing to NFS mount and is
experiencing no problems.  The machine that has started failing is RHEL
3.  The second machine is running RHEL 4 and the NFS server is running
RHEL 4.  They are manually kept up to date using up2date command.  Here
is the script.

day=`date +%a`
if [ "$day" == "Sun" ]
   then
        STAMP="Wk`date +%U`-$day`date +%m%d%H%M`"
   else
        STAMP="$day`date +%m%d%H%M`"
fi

logfile=/var/log/db-bkup/bkuplog-$STAMP
db_bkup_file=$bkup_hold_dir/`hostname`-db_bkup-$STAMP.tar.gz

# Stop mysql process to start backup, force flush to disk
/sbin/service mysql stop

echo "backup started at `date`" >$logfile
cd /home/

# backup /var/lib/mysql to $db-bkup-file directory
tar -czv --ignore-failed-read --totals mysql >$db_bkup_file 2>>$logfile
CODE=$?
if [ $CODE == 0 ]
then
   echo "backup completed at `date`" >>$logfile
else
   echo "Backup failed at `date` with error code ($CODE)" >>$logfile
fi

# Start mysql service again, will resync with master at this point
/sbin/service mysql start


The tar command is failing in the exact same place every time,
regardless of when run from cron or kicked off manually.  I know this by
the logfile which shows a list of all the files written.  It appears to
be part way through a directory.  If I run tar on that directory alone,
it completes fine.  I can find no other error indicators.



Richard Riley
System Administrator

Procuri Inc.
www.procuri.com








More information about the rhn-users mailing list