packages/net/speedtest-netperf/files/speedtest-netperf.sh
Tony Ambardar 463590e2bc speedtest-netperf: new package to measure network performance
The speedtest-netperf.sh script measures the network throughput while
monitoring latency under load and capturing key CPU usage and frequency
statistics. The script can emulate a web-based speed test by downloading
and then uploading from an internet server, or perform simultaneous
download and upload to mimic the stress of the FLENT test program.

It simplifies tasks such as validating ISP provisioned speeds or setting
up and fine-tuning SQM, directly on the router. The CPU usage details
can also help determine if the demands of SQM, routing and other tasks
such as the test itself are exhausting the device's CPUs.

This script leverages earlier scripts from the CeroWrt project used for
bufferbloat mitigation, betterspeedtest.sh and netperfrunner.sh. They are
used with the permission of the author, Rich Brown.

Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
2019-02-22 13:34:25 -08:00

408 lines
12 KiB
Bash

#!/bin/sh
# This speed testing script provides a convenient means of on-device network
# performance testing for OpenWrt routers, and subsumes functionality of the
# earlier CeroWrt scripts betterspeedtest.sh and netperfrunner.sh written by
# Rich Brown.
#
# When launched, the script uses netperf to run several upload and download
# streams to an Internet server. This places heavy load on the bottleneck link
# of your network (probably your Internet connection) while measuring the total
# bandwidth of the link during the transfers. Under this network load, the
# script simultaneously measures the latency of pings to see whether the file
# transfers affect the responsiveness of your network. Additionally, the script
# tracks the per-CPU processor usage, as well as the netperf CPU usage used for
# the test. On systems that report CPU frequency scaling, the script can also
# report per-CPU frequencies.
#
# The script operates in two modes of network loading: sequential and
# concurrent. The default sequential mode emulates a web-based speed test by
# first downloading and then uploading network streams, while concurrent mode
# provides a stress test by dowloading and uploading streams simultaneously.
#
# NOTE: The script uses servers and network bandwidth that are provided by
# generous volunteers (not some wealthy "big company"). Feel free to use the
# script to test your SQM configuration or troubleshoot network and latency
# problems. Continuous or high rate use of this script may result in denied
# access. Happy testing!
#
# For more information, consult the online README.md:
# https://github.com/openwrt/packages/blob/master/net/speedtest-netperf/files/README.md
# Usage: speedtest-netperf.sh [-4 | -6] [ -H netperf-server ] [ -t duration ] [ -p host-to-ping ] [ -n simultaneous-streams ] [ -s | -c ]
# Options: If options are present:
#
# -H | --host: netperf server name or IP (default netperf.bufferbloat.net)
# Alternate servers are netperf-east (east coast US),
# netperf-west (California), and netperf-eu (Denmark)
# -4 | -6: Enable ipv4 or ipv6 testing (ipv4 is the default)
# -t | --time: Duration of each direction's test - (default - 60 seconds)
# -p | --ping: Host to ping to measure latency (default - gstatic.com)
# -n | --number: Number of simultaneous sessions (default - 5 sessions)
# based on whether concurrent or sequential upload/downloads)
# -s | -c: Sequential or concurrent download/upload (default - sequential)
# Copyright (c) 2014 - Rich Brown <rich.brown@blueberryhillsoftware.com>
# Copyright (c) 2018 - Tony Ambardar <itugrok@yahoo.com>
# GPLv2
# Summarize contents of the ping's output file as min, avg, median, max, etc.
# input parameter ($1) file contains the output of the ping command
summarize_pings() {
# Process the ping times, and summarize the results
# grep to keep lines with "time=", and sed to isolate time stamps and sort them
# awk builds an array of those values, prints first & last (which are min, max)
# and computes average.
# If the number of samples is >= 10, also computes median, and 10th and 90th
# percentile readings.
sed 's/^.*time=\([^ ]*\) ms/\1 pingtime/' < $1 | grep -v "PING" | sort -n | awk '
BEGIN {numdrops=0; numrows=0;}
{
if ( $2 == "pingtime" ) {
numrows += 1;
arr[numrows]=$1; sum+=$1;
} else {
numdrops += 1;
}
}
END {
pc10="-"; pc90="-"; med="-";
if (numrows>=10) {
ix=int(numrows/10); pc10=arr[ix]; ix=int(numrows*9/10);pc90=arr[ix];
if (numrows%2==1) med=arr[(numrows+1)/2]; else med=(arr[numrows/2]);
}
pktloss = numdrops>0 ? numdrops/(numdrops+numrows) * 100 : 0;
printf(" Latency: [in msec, %d pings, %4.2f%% packet loss]\n",numdrops+numrows,pktloss)
if (numrows>0) {
fmt="%9s: %7.3f\n"
printf(fmt fmt fmt fmt fmt fmt, "Min",arr[1],"10pct",pc10,"Median",med,
"Avg",sum/numrows,"90pct",pc90,"Max",arr[numrows])
}
}'
}
# Summarize the contents of the load file, speedtest process stat file, cpuinfo
# file to show mean/stddev CPU utilization, CPU freq, netperf CPU usage.
# input parameter ($1) file contains CPU load/frequency samples
summarize_load() {
cat $1 /proc/$$/stat | awk -v SCRIPT_PID=$$ '
# track CPU frequencies
$1 == "cpufreq" {
sum_freq[$2]+=$3/1000
n_freq_samp[$2]++
}
# total CPU of speedtest processes
$1 == SCRIPT_PID {
tot=$16+$17
if (init_proc_cpu=="") init_proc_cpu=tot
proc_cpu=tot-init_proc_cpu
}
# track aggregate CPU stats
$1 == "cpu" {
tot=0; for (f=2;f<=8;f++) tot+=$f
if (init_cpu=="") init_cpu=tot
tot_cpu=tot-init_cpu
n_load_samp++
}
# track per-CPU stats
$1 ~ /cpu[0-9]+/ {
tot=0; for (f=2;f<=8;f++) tot+=$f
usg=tot-($5+$6)
if (init_tot[$1]=="") {
init_tot[$1]=tot
init_usg[$1]=usg
cpus[n_cpus++]=$1
}
if (last_tot[$1]>0) {
sum_usg_2[$1] += ((usg-last_usg[$1])/(tot-last_tot[$1]))^2
}
last_tot[$1]=tot
last_usg[$1]=usg
}
END {
printf(" CPU Load: [in %% busy (avg +/- std dev)")
for (i in sum_freq) if (sum_freq[i]>0) {printf(" @ avg frequency"); break}
if (n_load_samp>0) n_load_samp--
printf(", %d samples]\n", n_load_samp)
for (i=0;i<n_cpus;i++) {
c=cpus[i]
if (n_load_samp>0) {
avg_usg=(last_tot[c]-init_tot[c])
avg_usg=avg_usg>0 ? (last_usg[c]-init_usg[c])/avg_usg : 0
std_usg=sum_usg_2[c]/n_load_samp-avg_usg^2
std_usg=std_usg>0 ? sqrt(std_usg) : 0
printf("%9s: %5.1f +/- %4.1f", c, avg_usg*100, std_usg*100)
avg_freq=n_freq_samp[c]>0 ? sum_freq[c]/n_freq_samp[c] : 0
if (avg_freq>0) printf(" @ %4d MHz", avg_freq)
printf("\n")
}
}
printf(" Overhead: [in %% used of total CPU available]\n")
printf("%9s: %5.1f\n", "netperf", tot_cpu>0 ? proc_cpu/tot_cpu*100 : 0)
}'
}
# Summarize the contents of the speed file to show formatted transfer rate.
# input parameter ($1) indicates transfer direction
# input parameter ($2) file contains speed info from netperf
summarize_speed() {
printf "%9s: %6.2f Mbps\n" $1 $(awk '{s+=$1} END {print s}' $2)
}
# Capture process load, then per-CPU load/frequency info at 1-second intervals.
sample_load() {
local cpus="$(find /sys/devices/system/cpu -name 'cpu[0-9]*' 2>/dev/null)"
local f="cpufreq/scaling_cur_freq"
cat /proc/$$/stat
while : ; do
sleep 1s
egrep "^cpu[0-9]*" /proc/stat
for c in $cpus; do
[ -r $c/$f ] && echo "cpufreq $(basename $c) $(cat $c/$f)"
done
done
}
# Print a line of dots as a progress indicator.
print_dots() {
while : ; do
printf "."
sleep 1s
done
}
# Start $MAXSESSIONS datastreams between netperf client and server
# netperf writes the sole output value (in Mbps) to stdout when completed
start_netperf() {
for i in $( seq $MAXSESSIONS ); do
netperf $TESTPROTO -H $TESTHOST -t $1 -l $TESTDUR -v 0 -P 0 >> $2 &
# echo "Starting PID $! params: $TESTPROTO -H $TESTHOST -t $1 -l $TESTDUR -v 0 -P 0 >> $2"
done
}
# Wait until each of the background netperf processes completes
wait_netperf() {
# gets a list of PIDs for child processes named 'netperf'
# echo "Process is $$"
# echo $(pgrep -P $$ netperf)
local err=0
for i in $(pgrep -P $$ netperf); do
# echo "Waiting for $i"
wait $i || err=1
done
return $err
}
# Stop the background netperf processes
kill_netperf() {
# gets a list of PIDs for child processes named 'netperf'
# echo "Process is $$"
# echo $(pgrep -P $$ netperf)
for i in $(pgrep -P $$ netperf); do
# echo "Stopping $i"
kill -9 $i
wait $i 2>/dev/null
done
}
# Stop the current sample_load() process
kill_load() {
# echo "Load: $LOAD_PID"
kill -9 $LOAD_PID
wait $LOAD_PID 2>/dev/null
LOAD_PID=0
}
# Stop the current print_dots() process
kill_dots() {
# echo "Dots: $DOTS_PID"
kill -9 $DOTS_PID
wait $DOTS_PID 2>/dev/null
DOTS_PID=0
}
# Stop the current ping process
kill_pings() {
# echo "Pings: $PING_PID"
kill -9 $PING_PID
wait $PING_PID 2>/dev/null
PING_PID=0
}
# Stop the current load, pings and dots, and exit
# ping command catches and handles first Ctrl-C, so you have to hit it again...
kill_background_and_exit() {
kill_netperf
kill_load
kill_dots
rm -f $DLFILE
rm -f $ULFILE
rm -f $LOADFILE
rm -f $PINGFILE
echo; echo "Stopped"
exit 1
}
# Measure speed, ping latency and cpu usage of netperf data transfers
# Called with direction parameter: "Download", "Upload", or "Bidirectional"
# The function gets other info from globals and command-line arguments.
measure_direction() {
# Create temp files for netperf up/download results
ULFILE=$(mktemp /tmp/netperfUL.XXXXXX) || exit 1
DLFILE=$(mktemp /tmp/netperfDL.XXXXXX) || exit 1
PINGFILE=$(mktemp /tmp/measurepings.XXXXXX) || exit 1
LOADFILE=$(mktemp /tmp/measureload.XXXXXX) || exit 1
# echo $ULFILE $DLFILE $PINGFILE $LOADFILE
local dir=$1
local spd_test
# Start dots
print_dots &
DOTS_PID=$!
# echo "Dots PID: $DOTS_PID"
# Start Ping
if [ $TESTPROTO -eq "-4" ]; then
ping $PINGHOST > $PINGFILE &
else
ping6 $PINGHOST > $PINGFILE &
fi
PING_PID=$!
# echo "Ping PID: $PING_PID"
# Start CPU load sampling
sample_load > $LOADFILE &
LOAD_PID=$!
# echo "Load PID: $LOAD_PID"
# Start netperf datastreams between client and server
if [ $dir = "Bidirectional" ]; then
start_netperf TCP_STREAM $ULFILE
start_netperf TCP_MAERTS $DLFILE
else
# Start unidirectional netperf with the proper direction
case $dir in
Download) spd_test="TCP_MAERTS";;
Upload) spd_test="TCP_STREAM";;
esac
start_netperf $spd_test $DLFILE
fi
# Wait until background netperf processes complete, check errors
if ! wait_netperf; then
echo;echo "WARNING: netperf returned errors. Results may be inaccurate!"
fi
# When netperf completes, stop the CPU monitor, dots and pings
kill_load
kill_pings
kill_dots
echo
# Print TCP Download/Upload speed
if [ $dir = "Bidirectional" ]; then
summarize_speed Download $DLFILE
summarize_speed Upload $ULFILE
else
summarize_speed $dir $DLFILE
fi
# Summarize the ping data
summarize_pings $PINGFILE
# Summarize the load data
summarize_load $LOADFILE
# Clean up
rm -f $DLFILE
rm -f $ULFILE
rm -f $PINGFILE
rm -f $LOADFILE
}
# ------- Start of the main routine --------
# set an initial values for defaults
TESTHOST="netperf.bufferbloat.net"
TESTDUR="60"
PINGHOST="gstatic.com"
MAXSESSIONS=5
TESTPROTO="-4"
TESTSEQ=1
# read the options
# extract options and their arguments into variables.
while [ $# -gt 0 ]
do
case "$1" in
-s|--sequential) TESTSEQ=1 ; shift 1 ;;
-c|--concurrent) TESTSEQ=0 ; shift 1 ;;
-4|-6) TESTPROTO=$1 ; shift 1 ;;
-H|--host)
case "$2" in
"") echo "Missing hostname" ; exit 1 ;;
*) TESTHOST=$2 ; shift 2 ;;
esac ;;
-t|--time)
case "$2" in
"") echo "Missing duration" ; exit 1 ;;
*) TESTDUR=$2 ; shift 2 ;;
esac ;;
-p|--ping)
case "$2" in
"") echo "Missing ping host" ; exit 1 ;;
*) PINGHOST=$2 ; shift 2 ;;
esac ;;
-n|--number)
case "$2" in
"") echo "Missing number of simultaneous streams" ; exit 1 ;;
*) MAXSESSIONS=$2 ; shift 2 ;;
esac ;;
--) shift ; break ;;
*) echo "Usage: speedtest-netperf.sh [ -s | -c ] [-4 | -6] [ -H netperf-server ] [ -t duration ] [ -p host-to-ping ] [ -n simultaneous-sessions ]" ; exit 1 ;;
esac
done
# Check dependencies
if ! netperf -V >/dev/null 2>&1; then
echo "Missing netperf program, please install" ; exit 1
fi
# Start the main test
DATE=$(date "+%Y-%m-%d %H:%M:%S")
echo "$DATE Starting speedtest for $TESTDUR seconds per transfer session."
echo "Measure speed to $TESTHOST (IPv${TESTPROTO#-}) while pinging $PINGHOST."
echo -n "Download and upload sessions are "
[ "$TESTSEQ " -eq "1" ] && echo -n "sequential," || echo -n "concurrent,"
echo " each with $MAXSESSIONS simultaneous streams."
# Catch a Ctl-C and stop background netperf, CPU stats, pinging and print_dots
trap kill_background_and_exit HUP INT TERM
if [ $TESTSEQ -eq "1" ]; then
measure_direction "Download"
measure_direction "Upload"
else
measure_direction "Bidirectional"
fi