UNIX for Dummies Questions & Answers

May be not i am missing.

We have a alarm which will take more than 70% CPU process and send mail with details. This script works fine for some time, after that suddenly it occupied more CPU. The details shows this awk command. Then i asked helop in the forum for more details. Even i too wonder that this awk wont take more more CPU and more time as this is the only command that complete script has.

I need some help from experts that to implement the logic when i found the error exit from the loop. dont need to check the entire 1000 lines of log file.

thanks
senthil.

Do you want your script to stop as soon as any single exception is detected ?

yes i need to stop in the first occurance.

This will exit your awk script as soon as an exception is detected.

Code:

tail -n 1000 /logs/trace.log | awk 'BEGIN{OOM = 0; ScE = 0; NaE = 0; Jms = 0}
/OutOfMemoryException/{OOM = 1;exit}
/StaleConnectionException/{ScE = 1;exit}
/NamingException/{NaE = 1;exit}
/JmsTimeOutException/{Jms = 1;exit}
END{printf " %-7s%-7s%-7s%-7s",OOM,ScE,NaE,Jms;}'

However, I still think you should investigate elsewhere. Having your script running between thousands times slower than expected when you believe it is ok, and hundreds of thousands slower when you complain about a problem is the main issue.
Perhaps is there an I/O or hardware issue with your vxfs disk.

How big is the log file ?
At what rate does its size increase ?
Isn't the problem happening when you rotate that log file ?