I have a worker process that dies sometimes. The parent (when it works right) will respawn another child process. Doing a 'ps -ef' type command will then tell me when the new worker spawned, and this way I've been able to track down the death of the now defunct worker process.

But sometimes the parent can't respawn the worker. In these cases a "ps -ef|grep defunct| will give me a listing of the defunct processes, including the most recently dead one. That listing only tells me in minutes how much cpu time it used and the date it started. Is there any way to get more info out of it? I'd like to know exactly when (hour/minute/second date) it died. Ideas? thx.

I also have a number of logs that kick out. It's a bear to open them one by one (less) to search for errors. What's a good way to take that info and pass it through a program like awk or grep or something to search for just certain key words?

ß

Perhaps a bash script to look at several files and print the result to one file?

#!/bin/bash
cd /var/log
cat <log file name> | awk '/<search word>/ {print $0 }' <log file name> | tee logs.txt
cat <log file name> | awk '/<search word>/ {print $0 }' <log file name> | tee >> logs.txt
cat <log file name> | awk '/<search word>/ {print $0 }' <log file name> | tee >> logs.txt

where {print $0} gives the filename; <search word> is the word to search for; <log file name> is the name of the log to search; and tee sends the result to a text file for reading. Just keep adding log filenames, with appropriate 'cd' commands to change directory when necessary.

The first tee command initiates the file, the subsequent >> commands append to that file.