scripting help

[Vito Sarducci]

I want to pull the last line out from a the sar report im running on a daily basisin cron. I want to put this data to another file and concatenate each day for thirty days.

I want to take the average line and totals and add each days average line to this file so that I have approximately a file with 30 days of my sar data in it?

I have a cron job that produces this report each day. The output file it creates each day is called hostname_monthday. example would be:

system1_1029 for yesterdays output of this report.

21:45:00 3 3 2 92
22:00:01 3 3 2 92
22:15:00 9 7 53 32
22:30:00 3 4 91 1
22:45:00 3 5 91 1
23:00:00 3 5 91 1
23:15:00 3 4 91 1
23:30:00 3 4 91 1
23:45:00 3 4 90 2

Average 21 6 24 50

How can I accomplish this?

Please advise?

Vito

[Rodney Hills]

how about

tail -1 hostname_month >>collect_file

tail -1 gets the last line and >> will append it to collect_file

-- Rod Hills

[Santosh Nair_1]

How about something like this:
LOG=$(date +%m).Summary
echo "$DATE
\c ">> $LOG
sar -f |grep "Average" >> $LOG

This would put a date and the average line from sar into a file called $LOG, i.e. monthy.Summary. Just have to define the DATE variable.

-Santosh

[Craig Rants]

Vito,

Based on your output here is how you would get the last line from your file

cat FILENAME | awk '{ print $5 }' > OUTPUT

To use mathmatical formulas look at the expr command

A simple example would be:

AVG=`expr TOTAL/30`

These should get you started down the right road.

[Sachin Patel]

Hi Vito,
If I understand properly you need last line only. easy
#tail -1 filename >> outputfile

But I think you that is not what you want because it is seems too easy.

Sachin

[Craig Rants]

oops, didn't read that well. The awk statment is for the last column, not the last line.

[Vito Sarducci]

Yes, all these examples are great, but were missing one step. I need to take the data from 30 of these files per month and tail them all, which im not exact on how to do this?

I have 30 of the system1_1029, system1_1028,system1_1027 yada, yada, down to system1_1001...

So if i tail all of them, how do i get all the average lines into one output file?

Please advise?

Vito

[Rodney Hills]

If you want to process on multiple files, then use "for"

for hostfile in system1* ; do
last -1 $hostfile >>collect_file
done

This will extract the last line of each of the files and append onto collect_file

-- Rod Hills

[Alan Riggs · Accepted solution]

for i in system1_10*
do
tail -1 $i
done |awk '{T2+=$2} {T3+=$3} {T4+=$4} {T5+=$5} {print} END {printf"AVE: %f %f %f\n",T2/NR,T3/NR,T4/NR,T5/NR}' >> 10.sum



Play with formatting etc. as you desire.