One grep, multiple arguments and variables ? 1

[TSch]

Hi folks,

another problem.

I'm using the following command to monitor the states of print jobs:

chk1=$(find /data/ -name job.history -exec grep -l FAILED {}\; | wc -l)
chk2=$(find /data/ -name job.history -exec grep -l WAITING {}\; | wc -l)
chk3=$(find /data/ -name job.history -exec grep -l ABORTED {}\; | wc -l)

The Output of each grep is stored in a variable. As soon as the variable's value becomes greater than 0 there's a message being displayed on my monitoring box saying "Some jobs have failed" or "Some jobs have been aborted" and so on ...

However because there's an enormous directory structure existing under /data/ each find-grep combination takes a significant amount of time to complete.

Now the question is: Is there any way to get rid of 2 of the 3 find-grep commands and accomplish the same goal using only 1 find-grep ?

Regards
Thomas

 

[feherke]

Hi

I would use this :

find /data/ -name job.history -exec grep -o 'FAILED\|WAITING\|ABORTED' {} \; | sort | uniq -c > /tmp/printjobstatussummary

To produce a temporary file containing information like this :

      4 ABORTED
      1 FAILED

Producing that temporary file should be abit faster and generating alerts from its content should be fairly simple.

Tested with GNU [tt]grep[/tt]. See if your [tt]grep[/tt] implementation supports the -o ( --only-matching ) switch.

Feherke.

http://free.rootshell.be/~feherke/

 

[TSch]

Hi feherke,

thanks a lot !
That's perfect !

However: There was no -o option for grep under AIX. But here the "egrep" command did the trick.

Regards
Thomas

 

[p5wizard]

or grep -E (egrep is deprecated, but still exists in AIX)


HTH,

p5wizard