Shell Programming and Scripting

We are trying to design a flow so that an ETL job shouldn't start until the previous job completes. The script we have written is

Code:

while [ `ps -ef|grep "phantom DSD.RUN job_jobName"|grep -iv -e "grep" -e "SH -c"|wc -l` -ne 0  ]; do sleep 2; done

The loop however exits even when the process is actually running. Why could this be happening?

If you have pgrep, I would recommend using it instead:

Code:

while pgrep -f "phantom DSD.RUN job_jobName"
do
   sleep 2
done

Note that -f is used to match pattern against process name. If you know the actual process name, you can get rid of this option and pass the actual process name as argument.

We do not have pgrep. How does pgrep help? Why could this be happening?

The reason why pgrep is much preferred is because you don't have to worry about excluding grep command itself.
It is difficult to say without seeing the command output. So can you please post the output of below command:

Code:

ps -ef|grep "phantom DSD.RUN job_jobName"

Or you can debug your script by setting xtrace and verbose to understand what exactly is going on..

Code:

#!/bin/your_shell -xv

Where can I see the trace logs after running the script with the -xv tags?

---------- Post updated at 02:50 AM ---------- Previous update was at 01:52 AM ----------

The trace doesn't seem to help much. We saw the point where the loop exited. Went back to the shell and fired a ps and we could still see the same process running.

The trace often does not help that much, in my experience. Too much output, and too hard to wade through. On the other hand, it basically does show everything, if you know how to use it.

The first step is to get your script into normal appearance, instead of all glommed together on one line. The second step is to add some targeted diagnostic statements in your script.

ps combined with grep, as you are doing, will work fine, once you find the glitch.

Can you show the output of:

Code:

ps -ef | grep "phantom DSD.RUN job_jobName"

I've tried this simplified version of your command, and it works perfectly:

Code:

while ps | grep '[p]hantom DSD.RUN job_jobName'; do sleep 1; done

The grep command will give an exit status of 1 or 0 on found/not found that while can evaluate; the [p]... will find exactly p... but rule out the grep with its parameter line itself