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

You are probably using a debian chroot environment, since the
uptime binary is not included in the stock firmware.

The uptime data is kept in the file /proc/uptime.  The first number is (on Unix systems)
the number of seconds since the system reboot, and the second number is the number of
seconds the system processor is idle

for my DNS-323 this is
#cat /proc/uptime
1169682368.36 54908.72

The first number seem to be the current system time in seconds from the epoch (Jan. 01, 1970)
The second number could be the number of seconds since reboot, or could be idle processor time.

I wrote a shell script to read these values

#!/bin/sh

# formatted output function
print_time()
{
  t_sec=$1
#  echo $t_sec

  # constants
  min=60
  hr=`expr $min \* 60`
  day=`expr $hr \* 24`

  # comupte days, hrs, min
  t_days=`expr $t_sec / $day`
  t_hrs=`expr \( $t_sec \% $day \) / $hr`
  t_min=`expr \( $t_sec \% $hr \) / $min`

  # output
  echo -n : $t_days days", "
  echo -n $t_hrs hrs", "
  echo $t_min min
}
# end of formatted output function


UPTIME=/proc/uptime

# echo $UPTIME

if [ ! -f $UPTIME ]
then
    exit 1
fi

# read uptime
read up_sec idle_sec < $UPTIME

#output uptime
if [ -n "$up_sec" ]
then
  echo -n uptime
  print_time $up_sec
fi

# output idle time
if [ -n "$idle_sec" ]
then
  echo -n idle_time
  print_time $idle_sec
fi

for the preceding values of /proc/uptime the script output is:
uptime: 13537 days, 23 hrs, 46 min
idle_time: 0 days, 15 hrs, 15 min

I rebooted the system about  Wed Jan 24 08:30:29 and I ran the script at Wed Jan 24 23:46:07
the DNS-323 was idle all day while I was at work ~15 hrs between boot and running script, so I'm
not sure if the second number is idle or uptime.

The first number (without the decimal digits) gives Wed, 24 Jan 2007 23:46:08 when entered in
http://www.onlineconversion.com/unix_time.htm  this is ~the time I executed the cat /proc/uptime command

I think the kernel is writing the wrong values into /proc/uptime.


//Mig

i have compiled uprecords
http://upit.jtw.dk/dl/8r9/

you need to change the
SBINDIR=${FUNPLUGDIR}/bin
to
SBINDIR=${FUNPLUGDIR}/sbin
in fonz latest fun_plug

saltisol wrote:

sweet, thanks!

<snip>

seems to work like a charm

yes i have coded it in to my webinterface so i have it on the login page

Looks like this was written for ffp, and an older version. I am using a chroot debian install and I'd like to have the correct uptime. Can someone point me in the right direction to get this working?

Try this. it's a shell script funplug drop-in - no compiling necessary.

It sets an alias for uptime so if it doesn't work first time, then either read the README (gasp) or log off the NAS and back on again to source your profile. It also assumes that you have fonz's bash shell installed.

---

Attachments:
uptime-0.1-1.tgz, Size: 1,602 bytes, Downloads: 1,028

with uptime-0.1-1.tgz, I got some strange readout under chroot debian lenny:

11:19:38 up 0 days, 14:08, 2 users, load average: 2.05, 2.76, 2.83

11:27:23 up 0 days, 14:09, 2 users, load average: 1.99, 2.18, 2.49

11:31:20 up 0 days, 14:09, 2 users, load average: 3.15, 2.42, 2.49

result of cat /proc/uptime around 11:26 is:
1232969208.78 50947.06

I am pretty sure is 1st number the current time, but have no idea how the 2nd number is being updated/used.

Last edited by toolbox (2009-01-26 22:29:42)

Toolbox, if I understood your post, the output from uptime.sh was wrong.

you got this:
13:25:21 up 0 days, 14:11, 2 users, load average: 2.49, 2.65, 2.42

What's wrong in the above? What should it have been?

Here's a detailed thread I just started on the kernel bug that causes these issues: http://dns323.kood.org/forum/viewtopic.php?id=4006.

No real fix without a custom kernel; but I think you could adapt the scripts here to get the initial boot time by subtracting values from the cpu0 line in /proc/stat from the initial sampled uptime. It wouldn't make much difference but for purests it should help compensate for the missed time before fun_plug kicks in.

-Jeff

capone wrote:

I installed uptime-0.1-1.tgz... then rebooted the DNS... waited about 30mins then ran uptime and got the following output....

/ffp/bin/uptime.sh: line 70: users: command not found
11:21:47 up 0 days, 00:00, 0 users, load average: 5.46, 5.19, 4.05


Any ideas??

cd /mnt/HD_a2/funpkg/
wget http://www.inreto.de/dns323/fun-plug/0.5/packages/coreutils-6.12-1.tgz
funpkg -i ./coreutils-6.12-1.tgz

Hi!
This is how I solved it on my DNS-313 running Debian Lenny:
(Linux DNS-313 2.6.15 #79 Thu Jul 30 10:38:31 EDT 2009 armv4l GNU/Linux)

I created below script and put it under my home directory (/home/testuser) and made it executable:

#!/bin/bash
#
#   uptime for DNS-313 Debian 5 (Lenny) -
#   05122013 - created by Nagy Péter
#

# values
# read out uptime values from `cat /proc/uptime` then swap them...
days_hours=`cat /proc/uptime | awk '{print $2}' | awk '{printf("%d days, %02d:%02d:%02d",  int($0/86400), int(($0%86400)/3600), int(($0%3600)/60), int($0%60))}' | strings`
w_0=`w`
w_1=`echo $w_0| head -1 | awk -F" " '{print $1,$2}'`
w_2=`echo $w_0| head -1 | awk -F" " '{print $6,$7,$8,$9,$10,$11,$12}'`
echo "$w_1 $days_hours, $w_2"

then I renamed the old uptime binary in /usr/bin so that it won't be used unattendedly, and created a symlink as new "uptime" for my script
mv /usr/bin/uptime /usr/bin/uptimeold
ln -s /home/testuser/uptime.sh /usr/bin/uptime

You can see below that it works! uptimeold is the old deprecated uptime, while uptime is the new one...

root@DNS-313:~/testuser# uptimeold
11:25:36 up 16044 days, 10:25,  1 user,  load average: 0.06, 0.15, 0.17
root@DNS-313:~/testuser# uptime
11:25:42 up 0 days, 02:05:14, 1 user, load average: 0.14, 0.17, 0.17

Even Webmin shows normal uptimes after implementing this script ...

Best Regards,

aFoP

Last edited by aFoP (2013-12-05 13:01:56)