# print usage
usage ()
{
echo ""
echo "<<USAGE>> ${PROG} -i <iterations> -t <target filesystem> -b <block count> -s <block size>"
echo ""
echo " function run current dd write and read tests to a SAS filesystem to test thruput"
echo " where results will be stored in ${PROG}.results.<iterations>"
echo ""
echo " where iterations are the number of concurrent tests to run"
echo " target filesystem is the SAS read/write filesystem to test"
echo " block count is the number of block you want dd to process (eg: 20000)"
echo " block size is the size of the blocks you want dd to process in K (eg: 64)"
}
# validate input parameters for iterations and writeable target directory
validateinput ()
{
UNAME=`uname`
TMPNAME=tmpfile.$$
TMPFILE=${TARGETDIR}/${TMPNAME}
rm -f ${PROG}-readtest.out.* ${PROG}-writetest.out.* >/dev/null 2>&1
if [ -z "${ITERATIONS}" ]
then echo "<<ERROR>> iteration count not provided"
RC=1
fi
if [ ! -d "${TARGETDIR}" ]
then echo "<<ERROR>> target directory not provided or does not exist"
RC=1
else touch ${TMPFILE} >/dev/null 2>&1
if [ $? -ne 0 ]
then echo "<<ERROR>> target directory not writeable"
RC=1
else dd if=/dev/zero of=/dev/null bs=1k count=1 2>${TMPFILE}
if [ $? -ne 0 ]
then echo "<<ERROR>> cannot execute dd command"
cat ${TMPFILE}
rm -f ${TMPFILE} >/dev/null 2>&1
RC=1
fi
fi
rm -f ${TMPFILE} >/dev/null 2>&1
fi
if [ -z "${BLOCKS}" ]
then echo "<<ERROR>> block count not provided"
RC=1
fi
if [ -z "${BLOCKSIZE}" ]
then echo "<<ERROR>> block size not provided"
RC=1
fi
}
# process dd based writes and reads
processwriteread ()
{
echo "<<STATUS>> executing write testing"
sync
((COUNT=0))
while [ ${COUNT} -lt ${ITERATIONS} ]
do
((COUNT=COUNT+1))
echo " launching iteration: ${COUNT}"
(/usr/bin/time -p dd if=/dev/zero of=${TARGETDIR}/${PROG}-dd-write.${COUNT} bs=${BLOCKSIZE}k count=${BLOCKS}) > ${PROG}-writetest.out.${COUNT} 2>&1 &
done
echo " waiting for all write tests to complete"
wait
((COUNT=1))
while [ ${COUNT} -lt ${ITERATIONS} ]
do
SIZE1=`ls -lt ${TARGETDIR}/${PROG}-dd-write.${COUNT} | awk '{ print $5 }'`
((COUNT=COUNT+1))
SIZE2=`ls -lt ${TARGETDIR}/${PROG}-dd-write.${COUNT} | awk '{ print $5 }'`
if [ ${SIZE1} -ne ${SIZE2} ]
then echo ""
echo "<<STATUS>> target filesystem [${TARGETDIR}] does not have an adequate amount of free disk space to complete the test"
echo ""
echo "<<STATUS>> results of df on target filesystem [${TARGETDIR}]"
echo ""
df ${TARGETDIR}
echo ""
echo ""
echo "<<STATUS>> file size of write attempts"
echo ""
ls -lt ${TARGETDIR}/${PROG}-dd-write.*
echo ""
echo ""
echo "<<STATUS>> cleaning up filesystem [${TARGETDIR}]"
echo ""
echo "<<ERROR>> ${PROG}: failed to execute write test successfully"
rm -f ${TARGETDIR}/${PROG}-dd-* >/dev/null 2>&1
RC=1
break
fi
done
if [ ${RC} -eq 0 ]
then echo " write test complete"
echo "<<STATUS>> executing read testing"
sync
((COUNT=0))
while [ ${COUNT} -lt ${ITERATIONS} ]
do
((COUNT=COUNT+1))
echo " launching iteration: ${COUNT}"
(/usr/bin/time -p dd if=${TARGETDIR}/${PROG}-dd-write.${COUNT} of=/dev/null bs=${BLOCKSIZE}k count=${BLOCKS}) \
> ${PROG}-readtest.out.${COUNT} 2>&1 &
done
echo " waiting for all read tests to complete"
wait
echo " read test complete"
fi
}
# create output file and cleanup
cleanup ()
{
echo "<<STATUS>> creating file: ${PROG}.results.${ITERATIONS} with results"
#egrep -i "real" ${PROG}-*out* > ${PROG}.results.${ITERATIONS}
grep -i "real" ${PROG}-*out* > ${PROG}.results.${ITERATIONS}
rm -f ${TARGETDIR}/${PROG}-dd-* >/dev/null 2>&1
TRR=0
TWR=0
while read RECORD
do
TYPE=`echo ${RECORD} | awk -F\: '{ print $1 }'`
case "${UNAME}" in
HP-UX) TIME=`echo ${RECORD} | awk '{ print $NF }'`
COUNT=`echo ${TIME} | awk -F\: '{ print NF }'`
if [ ${COUNT} -eq 3 ]
then H=`echo ${TIME} | awk -F\: '{ print $1*60*60 }'`
M=`echo ${TIME} | awk -F\: '{ print $2*60 }'`
S=`echo ${TIME} | awk -F\: '{ print $3 }'`
else H=0
if [ ${COUNT} -eq 2 ]
then M=`echo ${TIME} | awk -F\: '{ print $1*60 }'`
S=`echo ${TIME} | awk -F\: '{ print $2 }'`
else M=0
S=`echo ${TIME} | awk '{ print $1 }'`
fi
fi
S=`echo "scale=2;${S} + ${M} + ${H}" | bc -l`
;;
*) S=`echo ${RECORD} | awk '{ print $NF }'`
;;
esac
case ${TYPE} in
${PROG}-read*) TRR=`echo "scale=2;${TRR} + ${S}" | bc -l`
;;
${PROG}-write*) TWR=`echo "scale=2;${TWR} + ${S}" | bc -l`
;;
esac
done < ${PROG}.results.${ITERATIONS}
#((FS=BLOCKS*BLOCKSIZE*1024))
FS=`echo "${BLOCKS}*${BLOCKSIZE}*1024" | bc -l`
MB=`echo "scale=2;${FS} / 1024 / 1024" | bc -l`
ARR=`echo "scale=2;${TRR} / ${ITERATIONS}" | bc -l`
AWR=`echo "scale=2;${TWR} / ${ITERATIONS}" | bc -l`
ART=`echo "scale=2;${MB} / ${ARR}" | bc -l`
AGR=`echo "scale=2;${ART} * ${ITERATIONS}" | bc -l`
AWT=`echo "scale=2;${MB} / ${AWR}" | bc -l`
AGW=`echo "scale=2;${AWT} * ${ITERATIONS}" | bc -l`
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
#echo "RESULTS" | tee -a ${PROG}.results.${ITERATIONS}
echo "RESULTS" | tee ${PROG}.results.${ITERATIONS}
echo "-------" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "INVOCATION : ${PROG} ${OPTIONS}" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "TARGET DETAILS" | tee -a ${PROG}.results.${ITERATIONS}
echo " directory : ${TARGETDIR}" | tee -a ${PROG}.results.${ITERATIONS}
case "${UNAME}" in
AIX) echo " df : `df -P ${TARGETDIR} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
MP=`df -P ${TARGETDIR} 2>/dev/null | tail -1 | awk '{ print $NF }'`
echo " mount point :" `/usr/sbin/mount 2>/dev/null | grep ${MP} 2>/dev/null | tail -1 | sed "s; ;;g"` | tee -a ${PROG}.results.${ITERATIONS}
;;
HP-UX) MP=`df -P ${TARGETDIR} 2>/dev/null | awk '{ print $NF }' | tail -1`
echo " mount point : `/sbin/mount 2>/dev/null | grep ${MP} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
;;
SunOS) echo " df : `df ${TARGETDIR} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
MP=`df ${TARGETDIR} 2>/dev/null | tail -1 | awk '{ print $NF }'`
echo " mount point : `/usr/sbin/mount 2>/dev/null | grep ${MP} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
;;
Linux) echo " df : `df -P ${TARGETDIR} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
MP=`df -P ${TARGETDIR} 2>/dev/null | tail -1 | awk '{ print $NF }'`
echo " mount point : `/bin/mount 2>/dev/null | grep ${MP} 2>/dev/null | tail -1`" | tee -a ${PROG}.results.${ITERATIONS}
;;
esac
echo " filesize : ${FS} bytes or ${MB} megabytes" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "STATISTICS" | tee -a ${PROG}.results.${ITERATIONS}
printf '%36s %8s %20s\n' " average read time in seconds :" "${ARR}" " " | tee -a ${PROG}.results.${ITERATIONS}
printf '%36s %8s %20s\n' " average read throughput rate :" "${ART}" "megabytes per second" | tee -a ${PROG}.results.${ITERATIONS}
#printf '%36s %8s %20s\n' " aggregate read throughput rate :" "${AGR}" "megabytes per second" | tee -a ${PROG}.results.${ITERATIONS}
printf '%36s %8s %20s\n' " average write time in seconds :" "${AWR}" " " | tee -a ${PROG}.results.${ITERATIONS}
printf '%36s %8s %20s\n' " average write throughput rate :" "${AWT}" "megabytes per second" | tee -a ${PROG}.results.${ITERATIONS}
#printf '%36s %8s %20s\n' " aggregate write throughput rate :" "${AGW}" "megabytes per second" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "" | tee -a ${PROG}.results.${ITERATIONS}
echo "<<STATUS>> processing complete"
}
# main
RC=0
PROG=${0##*/}
export LANG=en_US
OPTIONS=${*}
while getopts ":i:t:b:s:" opt
do
case $opt in
i) ITERATIONS=${OPTARG}
;;
t) TARGETDIR=${OPTARG}
;;
b) BLOCKS=${OPTARG}
;;
s) BLOCKSIZE=${OPTARG}
;;
\?) echo "<<ERROR>> Invalid option: -$OPTARG" >&2
RC=1
;;
:) echo "<<ERROR>> Option -$OPTARG requires an argument." >&2
RC=1
;;
esac
done
validateinput
if [ ${RC} -eq 0 ]
then processwriteread
if [ ${RC} -eq 0 ]
then cleanup
fi
else usage
fi
exit ${RC}
./iotest.sh -i 3 -t /test/saswork/ -b 2064834 -s 64
Comments
Post a Comment