Skip to content
This repository was archived by the owner on Mar 28, 2018. It is now read-only.

Commit

Permalink
Merge pull request #655 from MarioCarrilloA/pnp
Browse files Browse the repository at this point in the history 
Add memory consumption tool
  • Loading branch information
Damien Lespiau authored Feb 8, 2017
2 parents 212f7bf + e456fb1 commit 001c9ff
Show file tree
Hide file tree
Showing 2 changed files with 328 additions and 0 deletions.
26 changes: 26 additions & 0 deletions tests/metrics/README.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -22,3 +22,29 @@ You can also run each tests script separately. e.g.
$ cd tests/metrics
$ bash workload_time/cor_create_time.sh <times_to_run>
```

### Metrics tool
The objective of `collect_mem_consmd.sh` tool is to measure the average memory consumption of any
Clear Containers components under a determined stress test. This test launches a user defined number
of Clear Containers (by default 5) simultaneously, and once all containers are running then start
a similar workload inside each CC instance. The number of Clear Containers launched and the workload
can be customized.

**Requirments:**
collectl

| Option | Description |
| ------ | ---------------------------------------------------- |
| -h | Help Page. |
| -c | Number of containers. |
| -m | Number of messages (cc workload). |
| -s | Size of the message. |
| -p | CC component name to be monitered (default cc-proxy) |
| -x | Get metrics just about exact match. |
| -v | Shows version |

**Usage example:**

```bash
# ./collect_mem_consmd.sh -c 50 -m 1000 -p cc-shim -x
```
302 changes: 302 additions & 0 deletions tests/metrics/collect_mem_consmd.sh
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,302 @@
#!/bin/bash

# This file is part of cc-oci-runtime.
#
# Copyright (C) 2016 Intel Corporation
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; either version 2
# of the License, or (at your option) any later version.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.

# Description of the test:
# The objective of this tool is to measure the average memory consumption
# of a Clear Containers component during a determined Clear Containers
# stress flow.
#
# Note: It is recommended to launch the same number of Clear Containers per
# execution in order to compare results.

# set -x
set -e

# General env
MYSELF=${0##*/}
CC_NUM=5
MSGS_NUM=10000
MSG_SIZE=80
SHARED_DIR="shared"
PROC="cc-proxy"

# Host/Container paths
HOST_SHARED_PATH="$(pwd)/$SHARED_DIR"
CONT_SHARED_PATH="/root/$SHARED_DIR"

# Aux files
HOST_GATE="$HOST_SHARED_PATH/gate"
CONT_GATE="$CONT_SHARED_PATH/gate"
HOST_RDM_FILE="$HOST_SHARED_PATH/out.data"
CONT_RDM_FILE="$CONT_SHARED_PATH/out.data"
CSV_FILE="mem.csv"

declare WORKLOAD
declare COLLECT_PID
declare F_OPTS
declare AVG

# Shows help about this script
function help()
{
usage=$(cat << EOF
Usage: $MYSELF [-h] [--help] [-v] [--version]
Description:
This tool takes memory consumption
information about a determined number
of containers.
Options:
-h, --help Help page
-c, --num-containers Set the number of containers
that will be launched.
-m, --num-messages The number of messages is the
workload of each container.
-s, --size-message The number of characters per
message.
-p, --process-filter This option restricts which
processes are selected for
collection/display tool.
-x --exact-match Get metrics just about exact
match.
-v, --version Show version.
EOF
)
echo "$usage";
}

# This workload will run inside of each Clear
# Container, it will monitor a shared path with
# host, this is in order to establish communication
# with host and activate the I/O read/write
# workload at the same time in all containers.
function implement_workload()
{
WORKLOAD=$(cat << EOF
#!/bin/bash
function print_rdmsg()
{
echo "$(cat /dev/urandom 2>/dev/null | \
LC_CTYPE=C tr -dc "[:alnum:]" 2>/dev/null | \
head -c $MSG_SIZE )"
}
while [ 1 ];do
if [ -f "$CONT_GATE" ];then
for (( i=0; i<$MSGS_NUM; i++ ));do
print_rdmsg
done
break;
else
sleep 0.2;
fi
done
exit 0;
EOF
)
}

# This function launch collectl tool
# in order to get metrics about a specific tool.
function start_collect()
{
collectl -oTm -sZm -i1:1 \
--procfilt f$PROC \
--procopts tm -f $PROC-mem &

COLLECT_PID=$!
echo "collectl pid: $COLLECT_PID"
}

# It kills the collectl process.
function stop_collect()
{
kill "$COLLECT_PID"
}

# Get the average about a "csv" file produced
# by collectl tool.
function get_average()
{
sum=0
count=0

while read line; do
cmd=$(echo "$line" | awk -F "," '{print $30}' | grep $PROC)
if [ ! -z "$cmd" ];then
binary="$(echo "$cmd" | cut -d " " -f1 | \
xargs basename -z | grep $F_OPTS $PROC)"
if [ ! -z "$binary" ];then
mem_data=$(echo "$line" | awk -F "," '{print $12}')
if (( $mem_data > 0 ));then
sum=$(( $sum + $mem_data ))
count=$(( $count + 1 ))
fi
fi
fi

done < "$CSV_FILE"

if (( $count == 0 ));then
echo -e "\e[31m$MYSELF: error: no collected samples\e[39m"
exit 1;
fi
AVG=$(echo "$sum / $count" | bc)
echo -e "\e[33mMemory consumption average: $AVG\e[39m"
}

# It creates a "csv" file using a raw information
# file created by collectl tool.
function create_csv_file()
{
raw_file=$(ls | grep *.gz | head -n1)
collectl -sZ -oT -P --sep ',' -p "$raw_file" > "$CSV_FILE"
}

# cc-proxy binary is launched by a systemd
# service (it is possible to launches it from cmd),
# if that service is not running, it is
# not possible to boot a Clear Container, this
# function verify that.
function check_proxy_service()
{
status=$(systemctl status cc-proxy | grep -ow active)

if [[ $status == "active" ]];then
echo "cc-proxy is running"
echo "restart cc-proxy"
systemctl restart cc-proxy.service
else
echo "cc-proxy is not running"
echo "start cc-proxy"
systemctl start cc-proxy.service
fi

if [ $? -ne 0 ];then
echo -e "\e[31merror: cc-proxy could not start\e[39m"
exit 1
fi

sleep 1
}

# This function will check the necessary requirements
# in the environment for running Clear Containers.
function set_environment()
{
echo "Cleaning environment"

# clean docker waste
waste="$(docker ps -qa)"
if [ ! -z "$waste" ];then
docker rm -f $waste
fi

# clean result files
rm -rf *.gz
rm -rf *.csv

if [ -d "$HOST_SHARED_PATH" ];then
echo "Cleaning shared directory"
rm -rf "$HOST_SHARED_PATH"
fi

mkdir -p $HOST_SHARED_PATH
}

# main function
function main()
{
while (( $# ));
do
case $1 in
-h|--help)
help
exit 0;
;;
-c|--num-containers)
shift
CC_NUM=$1
;;
-m|--num-messages)
shift
MSGS_NUM=$1
;;
-s|--size-message)
shift
MSG_SIZE=$1
;;
-p|--process-filter)
shift
PROC=$1
;;
-x|--exact-match)
F_OPTS="-wx"
;;
-v|--version)
echo "$MYSELF version 0.1";
exit 0;
esac
shift
done

set_environment
implement_workload

echo "Stop cc-proxy"
systemctl stop cc-proxy
start_collect
sleep 0.5
check_proxy_service

for (( i=0; i<$CC_NUM; i++ ))
do
docker run -v $HOST_SHARED_PATH:$CONT_SHARED_PATH \
-d -i -t ubuntu bash -c "$WORKLOAD"
done

echo "Active gate"
echo "True" > $HOST_GATE

while [ 1 ];
do
cont_num=$(docker ps | wc -l)
if (( $cont_num == 1 ));then
break
fi

sleep 0.1
done

echo "Workload finishes"
stop_collect
create_csv_file

echo "Making a total"
get_average

echo -e "\e[32mFinish\e[39m"

exit 0;
}

# call to main function
main $@

0 comments on commit 001c9ff

Please sign in to comment.