DSM-G600, DNS-3xx and NSA-220 Hack Forum

Directory entries also take space, so if you have many files, it's perfectly possible that the overhead will be over 20M. If you leave out the --link-dest option then every directory will become a full backup, which takes a lot more space. If you still want to keep incremental backups, I recommend using rdiff-backup (http://www.nongnu.org/rdiff-backup/). It provides convenient high-level features for managing incremental backups which will save you from doing all of that manually.

Last edited by adambyrtek (2011-03-22 10:15:49)

And a lot depends on how often the files change.  In the extreme case of the files never changing once they are backed up, you'll have nothing but the small overhead of directory links for each backup level.  OTOH, if a bunch of big files change as fast as the backup frequency, most of your space will be taken up by versions of the large files.

It will take some tuning to adjust things to your pattern of file updates.  But the advice to delete the older backup dirs is good for when you start to get full.

Just for another example, I use rsnapshot:  (The "hourly's" run every four hours)

653G    /mnt/HD_b2/Snapshots/hourly.0/
159M    /mnt/HD_b2/Snapshots/hourly.1/
155M    /mnt/HD_b2/Snapshots/hourly.2/
155M    /mnt/HD_b2/Snapshots/hourly.3/
155M    /mnt/HD_b2/Snapshots/hourly.4/
155M    /mnt/HD_b2/Snapshots/hourly.5/
614M    /mnt/HD_b2/Snapshots/daily.0/
619M    /mnt/HD_b2/Snapshots/daily.1/
618M    /mnt/HD_b2/Snapshots/daily.2/
617M    /mnt/HD_b2/Snapshots/daily.3/
617M    /mnt/HD_b2/Snapshots/daily.4/
618M    /mnt/HD_b2/Snapshots/daily.5/
618M    /mnt/HD_b2/Snapshots/daily.6/
924M    /mnt/HD_b2/Snapshots/weekly.0/
600M    /mnt/HD_b2/Snapshots/weekly.1/
593M    /mnt/HD_b2/Snapshots/weekly.2/
648M    /mnt/HD_b2/Snapshots/weekly.3/
117G    /mnt/HD_b2/Snapshots/monthly.0/
777G    total

My link overhead is about 155MB.
I pull down a database that is 450MB each night from a web site I run and so that is why the daily snapshots are about 615MB.
If I get to the point where the disk is nearly full, I would clobber the monthly.0 to get a quick 117GB.

This disk has mostly media on it, and some backups of other systems.  It is a fairly static disk.

adambyrtek wrote:

karlrado wrote:

Just for another example, I use rsnapshot:  (The "hourly's" run every four hours)

I've never used rsnapshot, but it looks pretty similar to rdiff-backup. Do you know how do they compare?

OK, I've already found the difference: rsnapshot uses hardlinks while rdiff-backup maintains incremental deltas.

Last edited by adambyrtek (2011-03-22 23:16:37)

From your ffp/packages directory, just type:

funpkg -i rsync-3.0.7-1.tgz

Then type 'rsync --help' then pick the options that you want in your environment.

To update your FFP packages, this is what I run in a script

echo "`date` Rsyncing FFP from inreto.de to /ffp/0.5"
/ffp/bin/rsync -auv --stats --delete inreto.de::dns323/fun-plug/0.5/ /mnt/HD_a2/ffp/0.5/ >/mnt/usb/logs/rsync_inreto.log
if [ -f "/mnt/usb/logs/rsync_inreto.log" ]; then
  echo -n "`date` "
  /ffp/bin/grep "Number of files transferred" /mnt/usb/logs/rsync_inreto.log
fi

I also perform a backup of a few key directories every night to my usb stick:

for _f in "etc" "ffp" "opt"; do
  echo "`date` Rsyncing /${_f} to /mnt/usb/backup/${_f} log:/mnt/usb/logs/rsync_${_f}.log"
  /ffp/bin/rsync -auv --stats --delete /${_f}/ /mnt/usb/backup/${_f}/ >/mnt/usb/logs/rsync_${_f}.log
  if [ -f "/mnt/usb/logs/rsync_${_f}.log" ]; then
    echo -n "`date` "
    /ffp/bin/grep "Number of files:" /mnt/usb/logs/rsync_${_f}.log
    echo -n "`date` "
    /ffp/bin/grep "Number of files transferred:" /mnt/usb/logs/rsync_${_f}.log
  fi  
done

The output of the script itself is redirected to a log file. The echo -n does indeed print the date without a carriage return. I use that as a "header" for each line in the log. Remeber, when you 'grep' it prints the entire line that matches the text you are looking for, not just a portion of the line. A sample of the output is

Wed Mar 30 00:00:28 EDT 2011 Rsyncing FFP from inreto.de to /ffp/0.5
Wed Mar 30 00:00:30 EDT 2011 Number of files transferred: 0
Wed Mar 30 00:00:30 EDT 2011 Rsyncing /etc to /mnt/usb/backup/etc log:/mnt/usb/logs/rsync_etc.log
Wed Mar 30 00:00:30 EDT 2011 Number of files: 112
Wed Mar 30 00:00:30 EDT 2011 Number of files transferred: 0
Wed Mar 30 00:00:30 EDT 2011 Rsyncing /ffp to /mnt/usb/backup/ffp log:/mnt/usb/logs/rsync_ffp.log
Wed Mar 30 00:00:37 EDT 2011 Number of files: 7915
Wed Mar 30 00:00:37 EDT 2011 Number of files transferred: 2
Wed Mar 30 00:00:37 EDT 2011 Rsyncing /opt to /mnt/usb/backup/opt log:/mnt/usb/logs/rsync_opt.log
Wed Mar 30 00:00:37 EDT 2011 Number of files: 33
Wed Mar 30 00:00:37 EDT 2011 Number of files transferred: 0

There are actually numerous logs. A summary that contains the info above and also the detailed rsync info should I wish to go review it. I check the summary log every day to monitor the units. It ultimately contains more than just what I posted above.

Sample rsync_inreto.log

receiving file list ... done

Number of files: 229
Number of files transferred: 0
Total file size: 278651585 bytes
Total transferred file size: 0 bytes
Literal data: 0 bytes
Matched data: 0 bytes
File list size: 8698
File list generation time: 0.001 seconds
File list transfer time: 0.000 seconds
Total bytes sent: 90
Total bytes received: 8738

sent 90 bytes  received 8738 bytes  3531.20 bytes/sec
total size is 278651585  speedup is 31564.52

Last edited by FunFiler (2011-03-30 07:35:43)

Yes, the summary gives me a quick and easy way to ensure all the processes have run, and have run correctly as well as the general health of the box. It is particularly useful for the SmartCtl log as it is large, yet I pull out just the number of "pass" entries to ensure both drives are in good shape.

LOG_FILE="/mnt/usb/logs/smartctl.log"
if [ -f "${LOG_FILE}" ]; then
  echo "`date` Clearing old SmartCtl log ${LOG_FILE}"
  rm ${LOG_FILE}
fi
for _f in "sda" "sdb"; do
  echo "`date` Dumping SmartCtl data for /dev/${_f} to ${LOG_FILE}"
  echo "`date` SmartCtrl Data for /dev/${_f}" >>  ${LOG_FILE}
  /ffp/sbin/smartctl -a -i -d marvell /dev/${_f} >> ${LOG_FILE}
done
if [ -f "${LOG_FILE}" ]; then
  echo -n "`date` SmartCtl log Diagnostic Pass count = "
  /ffp/bin/grep "self-assessment test result: PASSED" ${LOG_FILE}|/ffp/bin/wc -l
else
  echo "`date` WARNING: SmartCtl log ${LOG_FILE} does not exist"
fi

I rarely go and look at any of the detail now but all logs get archived so I can review right back to day one if I need to.

#!/ffp/bin/sh
echo "`date` Running script $0" 
export hostname=`hostname`
export cnd=`date +"%Y%m%d"`
/ffp/bin/tar -cz -f /mnt/usb/archive/${cnd}_${hostname}_Logs.tar.gz /mnt/usb/logs/* >/dev/null 2>&1
echo "`date` End of script $0 "
# eof

Similarly with the rsync logs, if I see a high number of files being transfered, I can review the detail log to see what happened.

Last edited by FunFiler (2011-03-31 03:55:27)

---

Attachments:
daily.sh, Size: 3,177 bytes, Downloads: 230