open-nomad/scripts/example_weave.bash

104 lines
2.3 KiB
Bash
Executable file

#!/bin/bash
if [[ "$USER" != "vagrant" ]]; then
echo "WARNING: This script is intended to be run from Nomad's Vagrant"
read -rsp $'Press any key to continue anyway...\n' -n1
fi
set -e
if [[ ! -a /usr/local/bin/weave ]]; then
echo "Installing weave..."
sudo curl -L git.io/weave -o /usr/local/bin/weave
sudo chmod a+x /usr/local/bin/weave
fi
weave launch || echo "weave running"
eval "$(weave env)"
if curl -s localhost:8500 > /dev/null; then
echo "Consul running"
else
echo "Running Consul dev agent..."
consul agent -dev > consul.out &
fi
if curl -s localhost:4646 > /dev/null; then
echo "Nomad running"
else
echo "Running Nomad dev agent..."
nomad agent -dev > nomad.out &
fi
sleep 5
echo "Running Redis with Weave in Nomad..."
cat > redis-weave.nomad <<EOF
job "weave-example" {
datacenters = ["dc1"]
type = "service"
group "cache" {
count = 1
task "redis" {
driver = "docker"
config {
image = "redis:3.2"
port_map {
db = 6379
}
# Use Weave overlay network
network_mode = "weave"
}
resources {
cpu = 500 # 500 MHz
memory = 256 # 256MB
network {
mbits = 10
port "db" {}
}
}
# By default services will advertise the weave address
service {
name = "redis"
tags = ["redis", "weave-addr"]
port = "db"
# Since checks are done by Consul on the host system, they default to
# the host IP:Port.
check {
name = "host-alive"
type = "tcp"
interval = "10s"
timeout = "2s"
}
# Script checks are run from inside the container, so you can use
# environment vars to get the container ip:port.
check {
name = "container-script"
type = "script"
command = "/usr/local/bin/redis-cli"
args = ["-p", "\${NOMAD_PORT_db}", "QUIT"]
interval = "10s"
timeout = "2s"
}
}
# Setting address_mode = "host" will create a service entry with the
# host's address.
service {
name = "host-redis"
tags = ["redis", "host-addr"]
port = "db"
address_mode = "host"
}
}
}
}
EOF
nomad run redis-weave.nomad