would like to see how others would do this...



i get an output from a script like:

rlink: rlk_ecims-dr_lloravol_rvg
state: active
status: has 1833 outstanding writes, occupying 66328 kbytes (0%) on the srlis behind by 00:00:08 hours

flags: write enabled attached consistent connected asynchronous
protocol: udp/ip


if i wanted to keep an eye on either the percentage or the time, how would you parse this output? more specifically, this is the output of veritas replicator. i would like an email if either a) the percentage is higher than 50% or b) more than 3 hours behind.
i would love to see how someone else would do this with a .sh script...
thanks,

here's an awk implementation. modify to suit your needs

code:

awk 'begin { threshold=50;
             timetheshold=3;
             sysadmin="root";
             email="/usr/bin/mailx "
     }
     /status/{
               for(i=1;i<=nf;i++) {
                  if ( $i ~ /^\(/ ) { gsub(/[(%)]/,"",$i)
                       percent=$i
                  }
                  if ( $i ~ /behind/) { time=$(i+2) }
               }
     }
     end{
           n=split(time,timearr,":")          
           if ( percent > threshold ||  int(timearr[1]) > timethreshold) {
               cmd = email sysadmin   
               print "body text " | cmd  
               close(cmd)
               print "email sent"
           }

wow, that was fast! i'll give this a whiz today and let you know. thanks ghostdog74!

praveen,
coincidently, my solution is very similar to yours.
the issue that i see in your solution is the fact that you are loading
the entire file into one string.
this can cause problems in the performance for large files.
here is my solution not loading the entire file into a string:

code:

typeset -i mpctlimit=50
typeset -i mtimelimit=3

minpfile='b'
mtimere='[0-9][0-9]:[0-9][0-9]:[0-9][0-9]'
mhhre='[0-9][0-9]'
mmmssre='[0-9][0-9]:[0-9][0-9]'

mperctline=`sed -n '/.*%.*/p' ${minpfile}`
mperctvalue=`echo ${mperctline} | sed 's/.*(\([0-9].*\)%).*/\1/'`
echo "mperct = "$mperctvalue" limit = "${mpctlimit}
if [ ${mperctvalue} -gt ${mpctlimit} ];then
echo "above the percentage limit"
fi

mhoursline=`sed -n "/.*${mtimere}.*/p" ${minpfile}`
mhoursvalue=`echo ${mhoursline} | sed "s/.*\(${mhhre}\){mmmssre}.*/\1/"`
echo "mhours = "$mhoursvalue" limit = "${mtimelimit}
if [ ${mhoursvalue} -gt ${mtimelimit} ];then

a variation from praveenkumar_l's solution where precentage and hours are read from file with one sed command.

code:
usage=50
time_limit=3

sed  -n 's/.*\([0-9]\+\)%.*\([0-9][0-9]\):[0-9][0-9]:[0-9][0-9].*/\1 \2/;t;p' out.txt | \
echo "${percentage}% ${hours}h"

if [[ $percentage -gt $usage ]] ;then
   echo "disk usage beyond $usage%"
   #send mail
fi

if [[ $hours -gt $time_limit ]] ;then
   echo "time limit beyond $time_limit hours"
   #send mail
fi

thank you all for the quick responses. i'm going to be working on this later this afternoon and i'll see what i end up with, if wanted i'll share my code when finished.

execute your script with -x option and post the trace please.