[SOLVED] BASH: How to write script that checks other script and controls it?

[aizul357]

Hi,

I have written a script that runs in background. This script runs a command and while the command is running it has to wait before it can perform following commands. Let's say this command takes varying amount of time to complete. What the script does is, once the command is completed it checks how long it took to complete the command and if over a certain length of time, it issues a restart command to a certain service. This service restart will solve the lag to that command.

The problem with the script is, for as long as the command hasn't completed ( be it an hour or two) the script wouldn't be able to restart the service.

So is there a way to write a script to check if that script, or better that command, has run for over a specified length of time then issue a stop command and restart the service without having the script to wait for that command to complete?

[atreyu]

You can use ps to track the elapsed time of the process. get the pid of the process (your script), then use ps to look for that time. you can use pidof to get it. you might have to format the output of the etime value to make sense of it:

ps -o etime --no-heading -p $(pidof -x script.sh)

[aizul357]

Thanks atreyu,

Your code works. I have been experimenting with it until I came across a command called timeout which does exactly what I want. I have to drop the original direction and use the new one. Are you familiar with timeout? I'm having a problem with its output, when I typed this:

/usr/bin/time -f %e timeout 0.1 ./script.sh

it outputs:

Command exited with non-zero status 124
0.10

Now, I don't want the status message, I just need the time. Any idea?

[atreyu]

Hmm, never even heard of timeout before! So after poking around on one of my systems...are you sure the timeout command is even available on your system?

which timeout

It is in /usr/bin/timeout on my system, and belongs to the coreutils package.

Even if it is there, i don't think it takes floating point numbers for duration (e.g., 0.1). this works on my system:

[root@localhost ~]# /usr/bin/time -f %e /usr/bin/timeout 1 sleep 5
1.00

[aizul357]

Unless I misunderstand it..here's from manpage :

NAME
       timeout - run a command with a time limit

SYNOPSIS
       timeout [OPTION] DURATION COMMAND [ARG]...
       timeout [OPTION]

...

DURATION is a floating point number with an optional suffix: `s' for seconds (the default), `m' for minutes, `h' for hours or `d' for days.

But that isn't important .

I tried the same commands you provided:

[root@localhost ~]# /usr/bin/time -f %e /usr/bin/timeout 1 sleep 5
Command exited with non-zero status 124
1.04

Strange that it output that message..

I'm using Arch Linux by the way..

[atreyu]

Interesting...not the same on my system - Fedora 15 and timeout (GNU coreutils) 8.10

[root@localhost ~]# /usr/bin/timeout --help
Usage: /usr/bin/timeout [OPTION] DURATION COMMAND [ARG]...
  or:  /usr/bin/timeout [OPTION]
Start COMMAND, and kill it if still running after DURATION.

Mandatory arguments to long options are mandatory for short options too.
  -k, --kill-after=DURATION
                   also send a KILL signal if COMMAND is still running
                   this long after the initial signal was sent.
  -s, --signal=SIGNAL
                   specify the signal to be sent on timeout.
                   SIGNAL may be a name like `HUP' or a number.
                   See `kill -l` for a list of signals
      --help     display this help and exit
      --version  output version information and exit

DURATION is an integer with an optional suffix:
`s' for seconds(the default), `m' for minutes, `h' for hours or `d' for days.

If the command times out, then exit with status 124.  Otherwise, exit
with the status of COMMAND.  If no signal is specified, send the TERM
signal upon timeout.  The TERM signal kills any process that does not
block or catch that signal.  For other processes, it may be necessary to
use the KILL (9) signal, since this signal cannot be caught.

Anyway, it seems that you'll just have to remove the portion of the output that you don't want. Here's some simple examples:

/usr/bin/time ... timeout ... command 2>&1|tail -n1

/usr/bin/time ... timeout ... command 2>&1|grep ^[0-9]*

[aizul357]

Cool! never thought of tail.

tail solves the problem. grep also but I think I stick with tail

Thank you very much

It seems that Arch coreutils is version 8.15.

www archlinux org/packages/core/i686/coreutils/

[atreyu]

It seems that Arch coreutils is version 8.15.

man, i thought Fedora was bleeding edge. I just checked and Fedora 16 is only on coreutils 8.12...nice Arch!

glad its sorted.