iostat monitor

If everybody who has done similar hacks can send me their magic, then maybe
we can fold it into the code, and create a more universal script.
We will just have to get past the ksh/bash difference of opinion. :-)

Cheers
     V

On Wed, Sep 15, 2010 at 6:19 AM, Isaac W Traxler <user-4dfb0dbf036e@xymon.invalid> wrote:

Hi,

Jason and I have hacked at the iostat and got something that seems to work
on Linux. We have not cleaned it up near enough or fixed all that we need.
Along with changing the script, the graph definitions need to also be
changed. Here is the what we have done with iostat so far:

#!/bin/bash
OS=$(uname -o)
PID=$$

if [[ ${OS} == "solaris" ]]
 then
   IOSTAT='/usr/bin/iostat -xrn'
 else
   IOSTAT='/usr/bin/iostat -x'
 fi

TEMPFILE=${BBTMP}/diskstat.tmp.${PID}
SHOW_NFS=no   # Set this to yes on server side clientlocal.cfg to change it
             # DISKSTAT:SHOW_NFS=yes
DURATION=270   # The duration of the iostat sample
             # This can be updated in the same way as above

# Now we redefine some variables, if they are set in clientlocal
LOGFETCH=${BBTMP}/logfetch.$(uname -n).cfg
if [ -f ${LOGFETCH} ]
   then
      grep "^DISKSTAT:" ${LOGFETCH} | cut -d":" -f2 \
                                    | while read NEW_DEF
                                    do
                                       ${NEW_DEF}
                                    done
fi

 ${TEMPFILE}  # Make sure it's empty

TEMPFILERAW="${TEMPFILE}.raw"
${IOSTAT} $DURATION 2  > ${TEMPFILERAW}  # And collect some data to work
with.
# We have to collect 2 sets, because the first set is the average since
boot.

# Define where the second set of data starts
LINE=$(cat ${TEMPFILERAW} | grep -n "^Device:" | tail -1 | cut -d":" -f1)
# take the second set, and massage it into usable data
TEMPFILEDATA="${TEMPFILE}.data"
if [[ ${OS} == "solaris" ]]
 then
   cat ${TEMPFILERAW} | awk "NR>${LINE}" \
                      | sed "s/,/ /g" \
                      | awk '{ print $NF" "$0 }' \
                      | awk '{ $NF="";print }' > ${TEMPFILEDATA}
 else
   cat ${TEMPFILERAW} | awk "NR>${LINE}" \
                      | awk '{ print $0" "$1 }' \
                      | awk '{ $NF="";print }' > ${TEMPFILEDATA}
 fi
rm ${TEMPFILERAW}
# Now we format the data and send it off to the server
if [[ ${OS} == "solaris" ]]
 then
   COLUMNS="reads writes kreads kwrites wait actv svct pw pb"
 else
   COLUMNS="rrqm wrqm r w rsec wsec avgrq-sz avgqu-sz await svctm util"
 fi
count=1
for subtest in ${COLUMNS}
do
  ((count=count+1))
  echo "" >> ${TEMPFILE}
  cat ${TEMPFILEDATA} | cut -d" " -f1,${count} \
                      | while read DEVICE VAL
                       do
                          echo "${DEVICE}" | grep ":/" > /dev/null
                          if [ $? -eq 0 -a "${SHOW_NFS}" = "no" ]
                          then
                             break
                          else
                             DEVICE=$(echo ${DEVICE} | tr : - )
                          fi
                          echo "${DEVICE}:${VAL}" >> ${TEMPFILE}
                       done
                       echo "" >> ${TEMPFILE}
                       ${BB} ${BBDISP} "data ${MACHINE}.diskstat-${subtest}
$(echo; cat ${TEMPFILE} ;echo "" ;echo "ignore this" )"
                       # Without the last echo "ignore this", it seems to
not graph the last entry.
                       # Odd really, but that seems to fix it.
                       rm ${TEMPFILE}
done
rm ${TEMPFILEDATA}


--
Isaac Traxler                             AIX,Linux Admin
Louisiana State University                user-4dfb0dbf036e@xymon.invalid
High Performance Computing                XXX-XXX-XXXX
LONI AIX Clusters
AIX, Linux Support

On Tue, 14 Sep 2010, Daniel Bourque wrote:

 Date: Tue, 14 Sep 2010 16:22:13 -0500

From: Daniel Bourque <user-a141068964db@xymon.invalid>
Reply-To: xymon at xymon.com
To: xymon at xymon.com
Subject: [xymon] iostat monitor


Sorry I can't reply to the thread for some reason i quit receiving the
emails, I checked the archive and noticed the replies to my former thread. (
thanks ! )

Vernon, since I don't run solaris here, only linux and some tru64, the -r
( csv output ) and -n ( friendly names ) options makes it hard to use your
shell script since they either don't exists or don't work the same. Can you
perhaps provide a same output of  "iostat -xrn" and along with formated text
you pass to hobbit in your check.

I can then provide a snippet of code for linux, which would provide the
equivalent output. So you could just add a case in the shell script.


case `uname` in
 Linux)
   /usr/bin/iostat -x $DURATION 2 | wonderful stuff > $TEMPFILE.raw
    ;;
 SunOS)
   /usr/bin/iostat -xrn $DURATION 2 > $TEMPFILE.raw
   ;;
esac


--
Dan