UNIX for Advanced & Expert Users

I have written a "simple" shell-script (BASH) to monitor the logs of our Resin-applicationserver. Whenever Resin runs out of available heap-space, the script then tries to restart the Resin-server.

My problem is that my "autopilot"-script fails when it's run from crontab.

Here's what the script does:

1. Check available heap-space. If heap-space < 15%, then restart server.

2. I created a function to stop the server, which is called next in this event.

Code:

func_emma_stop()
{
        if [ $test -eq 0 ]; then
                /etc/init.d/resin stop
        else
                func_eventlog "Resin should now be stopped!"
        fi
        if [ $debug -eq 1 ]; then
                func_eventlog "emma_stop was called, result=Stopping!"
        fi
}

3. Next I verify if all processes are stopped. I cannot use the command "/etc/init.d/resin restart" because the server will sometimes take too long to stop (handling user sessions) and I cannot depend on the sleep-command in the init-script.

4. Next I try to start the server using this function:

Code:

func_emma_start()
{
        if [ $test -eq 0 ]; then
                /etc/init.d/resin start
        else
                func_eventlog "Resin should now be started!"
        fi
        if [ $debug -eq 1 ]; then
                func_eventlog "emma_start was called, result=Starting!"
        fi
}

5.Then again check for successful start. Retry when failed or exit when done.

When run from crontab, it tries to execute "/etc/init.d/resin start" over and over again, without result. I know this, because I made the init-script mail me when it's called. There's no output at all, no errors in logging, nothing...

Do I need a special way to call the init-script? (Instead of just executing it)

One more problem for me is that we just acquired the company who runs this application and I am its admin. However, I only have a production-environment and cannot "experiment" too much.

Any help/hints would be much appreciated.

Also, I run this from root's crontab. Does this mean I automatically use root's environment-settings?

crontab-entry below:

Code:

*/5 7-18 * * 1-5 /home/beheer/scripts/logchecker-autopilot.sh 1>/dev/null 2>/dev/null

check your cron environment-settings, it is ok?

you maybe find the answer from :
https://www.unix.com/answers-to-frequently-asked-questions/13527-cron-crontab.html
https://www.unix.com/unix-for-advanced-and-expert-users/10674-not-running-cron.html?s=

Change the script sh to csh & check

I finally found out why things weren't working... I ran the env-command from crontab and did the same for root and found out that crontab's path was only /bin;/usr/bin. The init-script "inherited" these environmentsettings and couldn't find the Debian /sbin/start-stop-daemon. From there on, it was simple...

Thanks guys!

We use SUSE SLES Linux at my office a great deal and I run into this issue a great deal. One of the things I find that helps is doing a sudo -c to run commands from crontab. It seems to apply the real user environment to the cron shell.