mirror of
				https://github.com/linuxkit/linuxkit.git
				synced 2025-11-04 05:46:41 +00:00 
			
		
		
		
	The previous version just created a network name space which does not allow us to also test additional namespaces, e.g. for unix domain sockets. This commit uses runc to create a fully namespaced container to run a test in. It creates a container, configures the network interfaces in the new network namespace before starting the container. A OCI config.json template is used and then customised for a given test based on command line arguments. Finally, instead of iperf, we use the socket stress test from https://github.com/linuxkit/virtsock as it provides finer-grained control over the traffic patterns (e.g. long lived vs lots of short lived connections). Signed-off-by: Rolf Neugebauer <rolf.neugebauer@docker.com>
		
			
				
	
	
		
			254 lines
		
	
	
		
			7.3 KiB
		
	
	
	
		
			Bash
		
	
	
		
			Executable File
		
	
	
	
	
			
		
		
	
	
			254 lines
		
	
	
		
			7.3 KiB
		
	
	
	
		
			Bash
		
	
	
		
			Executable File
		
	
	
	
	
#! /bin/sh
 | 
						|
 | 
						|
# This script creates a container with runc and then runs sock_stress
 | 
						|
# (a flexible stress program for UDP/TCP/virtio/Hyper-V/Unix Domain
 | 
						|
# sockets). It configures networking either for IPv4 or IPv6 and bind
 | 
						|
# mounts a directory for the Unix Domain Socket. After a randomised
 | 
						|
# amount of time (maximum set with -t) it kills the sockstress tests
 | 
						|
# or the container. This process is repeated several times (-i).
 | 
						|
#
 | 
						|
# For networking, currently only veth pairs are used. But we plan to
 | 
						|
# extent this in the future.
 | 
						|
#
 | 
						|
# sock_stress supports multiple concurrent connections and sends a
 | 
						|
# configurable amount of data over the socket from a client to a
 | 
						|
# server, which echoes the data back. By configuring the amount of
 | 
						|
# data sent per connection, sock_stress can be used to create a large
 | 
						|
# number of short-lived connections (-s). By default a random,
 | 
						|
# relatively large amount of data is transferred.
 | 
						|
#
 | 
						|
# By default the server is run in the container, with the client in
 | 
						|
# the parent namespace. the -r option reverses this.
 | 
						|
 | 
						|
# set -x
 | 
						|
 | 
						|
# Initialise the Random Number Seed with the PID
 | 
						|
RANDOM=$$
 | 
						|
 | 
						|
# defaults for arguments
 | 
						|
ARG_PROTO=tcp
 | 
						|
ARG_IP=4
 | 
						|
ARG_CONN=1
 | 
						|
ARG_SHORT=0
 | 
						|
ARG_ITER=20
 | 
						|
ARG_TIME=10
 | 
						|
ARG_REV=0
 | 
						|
 | 
						|
usage() {
 | 
						|
    echo "Stress test for network namespace using sock_stress"
 | 
						|
    echo " -p  tcp|udp|unix   Protocol to use [$ARG_PROTO]"
 | 
						|
    echo " -ip 4|6            IP version to use if tcp|udp [$ARG_IP]"
 | 
						|
    echo " -c  <n>            Number of concurrent connections [$ARG_CONN]"
 | 
						|
    echo " -s                 Use short lived connections (default long)"
 | 
						|
    echo " -i  <n>            Number of iterations [$ARG_ITER]"
 | 
						|
    echo " -l  <n>            Maximum time of test before killing in s [$ARG_LEN]"
 | 
						|
    echo " -r                 Reverse (client in container)"
 | 
						|
}
 | 
						|
 | 
						|
echo "$@"
 | 
						|
 | 
						|
# parse arguments
 | 
						|
while [[ $# -gt 0 ]]; do
 | 
						|
    key="$1"
 | 
						|
 | 
						|
    case $key in
 | 
						|
        -p)
 | 
						|
            ARG_PROTO="$2"
 | 
						|
            shift
 | 
						|
            ;;
 | 
						|
        -ip)
 | 
						|
            ARG_IP="$2"
 | 
						|
            shift
 | 
						|
            ;;
 | 
						|
        -c)
 | 
						|
            ARG_CONN="$2"
 | 
						|
            shift
 | 
						|
            ;;
 | 
						|
        -s)
 | 
						|
            ARG_SHORT=1
 | 
						|
            ;;
 | 
						|
        -i)
 | 
						|
            ARG_ITER="$2"
 | 
						|
            shift
 | 
						|
            ;;
 | 
						|
        -l)
 | 
						|
            ARG_TIME="$2"
 | 
						|
            shift
 | 
						|
            ;;
 | 
						|
        -r)
 | 
						|
            ARG_REV=1
 | 
						|
            ;;
 | 
						|
        *)
 | 
						|
            usage
 | 
						|
            exit 1
 | 
						|
            ;;
 | 
						|
    esac
 | 
						|
    shift
 | 
						|
done
 | 
						|
 | 
						|
# Work out the maximum length of a connection. For short connections
 | 
						|
# (ie many connections we transfer a max of 4k for long connections
 | 
						|
# (lots of data) we transfer up to 8G per connection)
 | 
						|
[ "$ARG_SHORT" = "1" ] && MAX_LEN=4096 || MAX_LEN=8388608
 | 
						|
[ "$ARG_PROTO" = "unix" ] && ARG_IP=
 | 
						|
 | 
						|
# Make sure we do IPv6 if asked for
 | 
						|
if [ "$ARG_IP" = "6" ]; then
 | 
						|
    echo 0 > /proc/sys/net/ipv6/conf/all/disable_ipv6
 | 
						|
    echo 0 > /proc/sys/net/ipv6/conf/default/disable_ipv6
 | 
						|
    echo 1 > /proc/sys/net/ipv6/conf/all/forwarding
 | 
						|
    echo 1 > /proc/sys/net/ipv6/conf/default/forwarding
 | 
						|
fi
 | 
						|
 | 
						|
veth_ipv4() {
 | 
						|
    nspid="$1"
 | 
						|
    h_ip="$2"
 | 
						|
    n_ip="$3"
 | 
						|
    mask=24
 | 
						|
 | 
						|
    # create veth pair and assign the peer to the namespace
 | 
						|
    h_dev="h_$nspid"
 | 
						|
    n_dev="n_$nspid"
 | 
						|
    ip link add "$h_dev" type veth peer name "$n_dev"
 | 
						|
    ip link set "$n_dev" netns "$nspid"
 | 
						|
 | 
						|
    # set up address and up the devices. Host first
 | 
						|
    ip addr add "$h_ip"/"$mask" dev "$h_dev"
 | 
						|
    ip link set "$h_dev" up
 | 
						|
 | 
						|
    ip netns exec "$nspid" ip addr add "$n_ip"/"$mask" dev "$n_dev" 
 | 
						|
    ip netns exec "$nspid" ip link set lo up
 | 
						|
    ip netns exec "$nspid" ip link set "$n_dev" up
 | 
						|
    sleep 2 # Wait for link to settle
 | 
						|
}
 | 
						|
 | 
						|
veth_ipv6() {
 | 
						|
    nspid="$1"
 | 
						|
    h_ip="$2"
 | 
						|
    n_ip="$3"
 | 
						|
    mask=64
 | 
						|
 | 
						|
    # create veth pair and assign the peer to the namespace
 | 
						|
    h_dev="h_$nspid"
 | 
						|
    n_dev="n_$nspid"
 | 
						|
    ip link add "$h_dev" type veth peer name "$n_dev"
 | 
						|
    ip link set "$n_dev" netns "$nspid"
 | 
						|
 | 
						|
    # set up address and up the devices. Host first
 | 
						|
    ip -6 addr add "$h_ip"/"$mask" dev "$h_dev"
 | 
						|
    ip link set "$h_dev" up
 | 
						|
 | 
						|
    ip netns exec "$nspid" ip -6 addr add "$n_ip"/"$mask" dev "$n_dev" 
 | 
						|
    ip netns exec "$nspid" ip link set lo up
 | 
						|
    ip netns exec "$nspid" ip link set "$n_dev" up
 | 
						|
    sleep 2 # Wait for link to settle
 | 
						|
}
 | 
						|
 | 
						|
PID="$$"
 | 
						|
echo "PID=$PID"
 | 
						|
D="/test/$PID"
 | 
						|
mkdir -p "$D"
 | 
						|
 | 
						|
##
 | 
						|
## Create a runc config.json file based on the template
 | 
						|
##
 | 
						|
cp /config.template.json "$D"/config.json
 | 
						|
 | 
						|
# Add a bind mount for unix domain sockets.
 | 
						|
BMOUNT="[{\"destination\": \"/data\", \"type\": \"bind\", \"source\": \"$D\", \"options\": [\"rw\", \"rbind\", \"rprivate\"]}]"
 | 
						|
jq --argjson args "$BMOUNT" '.mounts |= . + $args' \
 | 
						|
       "$D"/config.json  > "$D"/foo.json && mv "$D"/foo.json "$D"/config.json
 | 
						|
 | 
						|
##
 | 
						|
## Run the test $ARG_ITER time
 | 
						|
##
 | 
						|
for i in $(seq 1 "$ARG_ITER"); do
 | 
						|
    # Work out IP addresses
 | 
						|
    if [ "$ARG_IP" = "6" ]; then
 | 
						|
        sub0=$(($PID%65535)); sub1=$(($i%65535))
 | 
						|
        h_ip="2001:$sub0:$sub1::1"; h_ip_addr="[$h_ip]"
 | 
						|
        n_ip="2001:$sub0:$sub1::2"; n_ip_addr="[$n_ip]"
 | 
						|
    else
 | 
						|
        sub0=$(($PID%255)); sub1=$(($i%255))
 | 
						|
        h_ip="10.$sub0.$sub1.1"; h_ip_addr="$h_ip"
 | 
						|
        n_ip="10.$sub0.$sub1.2"; n_ip_addr="$n_ip"
 | 
						|
    fi
 | 
						|
 | 
						|
    if [ "$ARG_REV" = "0" ]; then
 | 
						|
        # Server in container
 | 
						|
        if [ "$ARG_PROTO" = "unix" ]; then
 | 
						|
            SADDR="/data/stress.sock"
 | 
						|
            CADDR="$D/stress.sock"
 | 
						|
        else
 | 
						|
            SADDR="$n_ip_addr"
 | 
						|
            CADDR="$n_ip_addr"
 | 
						|
        fi
 | 
						|
        C_CMD="[\"/sock_stress\", \"-s\", \"$ARG_PROTO$ARG_IP://$SADDR\"]"
 | 
						|
        H_CMD="/rootfs/sock_stress -c $ARG_PROTO$ARG_IP://$CADDR -i 10000000 -l $MAX_LEN -p $ARG_CONN -v 1"
 | 
						|
    else
 | 
						|
        # Client in container
 | 
						|
        if [ "$ARG_PROTO" = "unix" ]; then
 | 
						|
            CADDR="/data/stress.sock"
 | 
						|
            SADDR="$D/stress.sock"
 | 
						|
        else
 | 
						|
            SADDR="$h_ip_addr"
 | 
						|
            CADDR="$h_ip_addr"
 | 
						|
        fi
 | 
						|
        C_CMD="[\"/sock_stress\", \"-c\", \"$ARG_PROTO$ARG_IP://$CADDR\", \"-i\", \"10000000\", \"-l\", \"$MAX_LEN\", \"-p\", \"$ARG_CONN\", \"-v\", \"1\"]"
 | 
						|
        H_CMD="/rootfs/sock_stress -s $ARG_PROTO$ARG_IP://$SADDR"
 | 
						|
    fi
 | 
						|
 | 
						|
    # Splice container command into json
 | 
						|
    jq --argjson args "$C_CMD" '.process.args = $args' \
 | 
						|
       "$D"/config.json  > "$D"/foo.json && mv "$D"/foo.json "$D"/config.json
 | 
						|
 | 
						|
    # Create container, get the namespace ID, and set up symlink for ip utils
 | 
						|
    CNAME="c-$PID-$i"
 | 
						|
    runc create -b "$D" "$CNAME"
 | 
						|
    nspid=$(runc list -f json | jq --arg id "$CNAME" -r '.[] | select(.id==$id) | .pid')
 | 
						|
    mkdir -p /var/run/netns && \
 | 
						|
        ln -s /proc/"$nspid"/ns/net /var/run/netns/"$nspid"
 | 
						|
 | 
						|
    # Configure network
 | 
						|
    if [ "$ARG_IP" = "6" ]; then
 | 
						|
        veth_ipv6 "$nspid" "$h_ip" "$n_ip"
 | 
						|
    else
 | 
						|
        veth_ipv4 "$nspid" "$h_ip" "$n_ip"
 | 
						|
    fi
 | 
						|
 | 
						|
    # Run
 | 
						|
    if [ "$ARG_REV" = "0" ]; then
 | 
						|
        runc start "$CNAME"
 | 
						|
        sleep 2 # Wait for container to start
 | 
						|
        $H_CMD &
 | 
						|
        pid_host=$!
 | 
						|
    else
 | 
						|
        $H_CMD &
 | 
						|
        pid_host=$!
 | 
						|
        sleep 1 # Wait for server to start
 | 
						|
        runc start "$CNAME"
 | 
						|
    fi
 | 
						|
 | 
						|
    # wait for a while before killing processes
 | 
						|
    sleep $(((RANDOM % $ARG_TIME )+1))
 | 
						|
 | 
						|
    R=$(($RANDOM%2))
 | 
						|
    case $R in
 | 
						|
        0)
 | 
						|
            echo "Kill test first"
 | 
						|
            kill -9 "$pid_host"
 | 
						|
            runc kill "$CNAME"
 | 
						|
            runc delete "$CNAME"
 | 
						|
            ;;
 | 
						|
        1)
 | 
						|
            echo "Kill container first"
 | 
						|
            runc kill "$CNAME"
 | 
						|
            runc delete "$CNAME"
 | 
						|
            kill -9 "$pid_host"
 | 
						|
            ;;
 | 
						|
    esac
 | 
						|
 | 
						|
    rm /var/run/netns/"$nspid"
 | 
						|
done
 | 
						|
rm -rf "$D"
 |