脚本在crontab中不起作用，但在终端中起作用

Question

I'm attempting to run a bash script in crontab on Centos 5.5 but i don't see it running as a pid (pid -e). The script itself uses an awk command to monitor a running log for a keyword then sends an email, then once the server is back online it sends a confirmation email:

tail -fn0 user | awk '
    !server_down && /disconnect_tcp_conn/ { 
        system("mail -s \"server down\" mail@gmail.com </dev/null") 
        print "server down"
        server_down = 1 }
    server_down && /daemon apps started/ {
        system("mail -s \"server's back!\" mail@gmail.com </dev/null") 
        print "Server's back!"
        server_down = 0 }' 

Originally, I had trouble with running the script while the terminal was closed. Someone suggested using setsid in the following way:

setsid server_mon.sh >/dev/null 2>&1 < /dev/null &

And this works great.. until it doesn't. For some reason, after some time like a day or so, the script no longer sends notifications. So, I thought to create a cronjob for the script.

Not being familiar with bash shell and not much of a programmer i tested the script in cron every minute to see if i was using it right:

* * * * *  /var/log/server_mon.sh

Didn't work. It would only say:

"/tmp/crontab.XXXXmu99ii" 19L, 760C written
crontab: installing new crontab

Tried to use rc.local.. can't. I think the centos version is too old (5.5), plus, the crontab might be the way to go because i might want to find out at what point the notifications seems to stop working and set up a cron job accordingly.. at least until i have time to fix it.

tried adding the path to the awk command in the script: /usr/bin/awk Still no pid.

tried adding the following changes from what i found on the web:

setsid bash -lic 'cd /var/log && exec SLR_reaper.sh -start' </dev/null &>/dev/null &

didn't work, and not sure if it's right.

my expected result is to get this:

setsid server_mon.sh >/dev/null 2>&1 < /dev/null &

somehow to be triggered by crontab:

@reboot /var/log/server_mon.sh

       *or*

* * * * * /var/log/server_mon.sh

Just to see it work.

Any direction or help would be great. If you need more info please let me know also. Thanks!

A screen shot of the log sent to mail: enter image description here

Answer 1

So the solution was pretty simple. I just had to provide the path to the file the script was monitoring, in the script itself:

tail -fn0 /var/log/user | /usr/bin/awk '
    !server_down && /disconnect_tcp_conn/ {
        system("mail -s \"server is down XXX.XXX.XXX.XXX\" mail@gmail.com </dev/null")
        print "server is Down XXX.XXX.XXX.XXX"
        server_down = 1 }
    server_down && /daemon apps started/ {
        system("mail -s \"Server is up XXX.XXX.XXX.XXX\" mail@gmail.com </dev/null")
        print "Server is Up XXX.XXX.XXX.XXX"
        server_down = 0 }'

The log file reported that it couldn't find the file which clued me to it being a path related issue.

Previously the script read:

tail -fn0 user | /usr/bin/awk '

Now I added the full path:

tail -fn0 /var/log/user | /usr/bin/awk '

Kudos to @EdMorton for telling me to track the log and then provide the full path in the script. I didn't realize this was necessary because the script was in the same directory as "user".