Contributed by Wes Morgan
While tuning our mail system, we needed to take a "snapshot" of the users' mailboxes at regular intervals over a 30-day period. This script simply calculates the average size and prints the arithmetic distribution of user mailboxes.
#! /bin/sh # # mailavg - average size of files in /usr/mail # # Written by Wes Morgan, [email protected], 2 Feb 90 ls -Fs /usr/mail | awk ' { if(NR != 1) { total += $1; count += 1; size = $1 + 0; if(size == 0) zercount+=1; if(size > 0 && size <= 10) tencount+=1; if(size > 10 && size <= 19) teencount+=1; if(size > 20 && size <= 50) uptofiftycount+=1; if(size > 50) overfiftycount+=1; } } END { printf("/usr/mail has %d mailboxes using %d blocks,", count,total) printf("average is %6.2f blocks\n", total/count) printf("\nDistribution:\n") printf("Size Count\n") printf(" O %d\n",zercount) printf("1-10 %d\n",tencount) printf("11-20 %d\n",teencount) printf("21-50 %d\n",uptofiftycount) printf("Over 50 %d\n",overfiftycount) }' exit 0
Here's a sample output from mailavg:
$mailavg
/usr/mail has 47 mailboxes using 5116 blocks, average is 108.85 blocks Distribution: Size Count O 1 1-10 13 11-20 1 21-50 5 Over 50 27
This administrative program is similar to the filesum program in Chapter 7. It processes the output of the ls command.
The conditional expression "NR != 1" could have been put outside the main procedure as a pattern. While the logic is the same, using the expression as a pattern clarifies how the procedure is accessed, making the program easier to understand.
In that procedure, Morgan uses a series of conditionals that allow him to collect distribution statistics on the size of each user's mailbox.