open-nomad/command/init.go

175 lines
4.0 KiB
Go

package command
import (
"fmt"
"io/ioutil"
"os"
"strings"
)
const (
// DefaultInitName is the default name we use when
// initializing the example file
DefaultInitName = "example.nomad"
)
// InitCommand generates a new job template that you can customize to your
// liking, like vagrant init
type InitCommand struct {
Meta
}
func (c *InitCommand) Help() string {
helpText := `
Usage: nomad init
Creates an example job file that can be used as a starting
point to customize further.
`
return strings.TrimSpace(helpText)
}
func (c *InitCommand) Synopsis() string {
return "Create an example job file"
}
func (c *InitCommand) Run(args []string) int {
// Check for misuse
if len(args) != 0 {
c.Ui.Error(c.Help())
return 1
}
// Check if the file already exists
_, err := os.Stat(DefaultInitName)
if err != nil && !os.IsNotExist(err) {
c.Ui.Error(fmt.Sprintf("Failed to stat '%s': %v", DefaultInitName, err))
return 1
}
if !os.IsNotExist(err) {
c.Ui.Error(fmt.Sprintf("Job '%s' already exists", DefaultInitName))
return 1
}
// Write out the example
err = ioutil.WriteFile(DefaultInitName, []byte(defaultJob), 0660)
if err != nil {
c.Ui.Error(fmt.Sprintf("Failed to write '%s': %v", DefaultInitName, err))
return 1
}
// Success
c.Ui.Output(fmt.Sprintf("Example job file written to %s", DefaultInitName))
return 0
}
var defaultJob = strings.TrimSpace(`
# There can only be a single job definition per file.
# Create a job with ID and Name 'example'
job "example" {
# Run the job in the global region, which is the default.
# region = "global"
# Specify the datacenters within the region this job can run in.
datacenters = ["dc1"]
# Service type jobs optimize for long-lived services. This is
# the default but we can change to batch for short-lived tasks.
# type = "service"
# Priority controls our access to resources and scheduling priority.
# This can be 1 to 100, inclusively, and defaults to 50.
# priority = 50
# Restrict our job to only linux. We can specify multiple
# constraints as needed.
constraint {
attribute = "${attr.kernel.name}"
value = "linux"
}
# Configure the job to do rolling updates
update {
# Stagger updates every 10 seconds
stagger = "10s"
# Update a single task at a time
max_parallel = 1
}
# Create a 'cache' group. Each task in the group will be
# scheduled onto the same machine.
group "cache" {
# Control the number of instances of this groups.
# Defaults to 1
# count = 1
# Configure the restart policy for the task group. If not provided, a
# default is used based on the job type.
restart {
# The number of attempts to run the job within the specified interval.
attempts = 10
interval = "5m"
# A delay between a task failing and a restart occuring.
delay = "25s"
# Mode controls what happens when a task has restarted "attempts"
# times within the interval. "delay" mode delays the next restart
# till the next interval. "fail" mode does not restart the task if
# "attempts" has been hit within the interval.
mode = "delay"
}
# Define a task to run
task "redis" {
# Use Docker to run the task.
driver = "docker"
# Configure Docker driver with the image
config {
image = "redis:latest"
port_map {
db = 6379
}
}
service {
name = "${TASKGROUP}-redis"
tags = ["global", "cache"]
port = "db"
check {
name = "alive"
type = "tcp"
interval = "10s"
timeout = "2s"
}
}
# We must specify the resources required for
# this task to ensure it runs on a machine with
# enough capacity.
resources {
cpu = 500 # 500 Mhz
memory = 256 # 256MB
network {
mbits = 10
port "db" {
}
}
}
# Specify configuration related to log rotation
# logs {
# max_files = 10
# max_file_size = 15
# }
# Controls the timeout between signalling a task it will be killed
# and killing the task. If not set a default is used.
# kill_timeout = "20s"
}
}
}
`)