Feedback and high-quality pull requests are highly welcome!
This repository contains POSIX shell scripts that wait for a given time until resources (TCP ports, HTTP services) become available and then execute predefined commands.
These scripts where created to workaround the fact that Docker Swarm does not support the depends_on constraint of Docker Compose, which means there is no control over the startup order of multiple containers possible. For the interested reader, a lengthy discussion about this issue can be found here moby/moby#31333 with a lot of pros & cons.
In contrast to other third-party solutions, such as:
- https://github.com/betalo-sweden/await
- https://github.com/ufoscout/docker-compose-wait/
- https://github.com/Logimethods/docker-eureka and
- https://github.com/vishnubob/wait-for-it (a fork of https://github.com/jlordiales/wait-for-it)
the scripts provided here have minimal system requirements (a POSIX shell, timeout, nc, curl or wget), work with BusyBox, Alpine Linux images and can be used without modifying existing images.
The scripts are automatically tested using Bats and executed under ash, bash, busybox, dash, ksh and zsh.
Usage: await-cmd.sh [OPTION]... TIMEOUT TEST_COMMAND [ARG...] [-- COMMAND [ARG...]]
Executes TEST_COMMAND repeatedly until it's exit code is 0. Then executes COMMAND.
Parameters:
TIMEOUT - Duration in seconds within TEST_COMMAND must return exit code 0.
TEST_COMMAND - Command that will be executed to test if the waiting condition is met.
COMMAND - Command to be executed once the TEST_COMMAND succeeded (optional).
Options:
-f - Force execution of COMMAND even if timeout occurred.
-t SECS - Duration in seconds after which a TEST_COMMAND process is terminated (optional, default: 10 seconds).
-w SECS - Waiting period in seconds between each execution of TEST_COMMAND (optional, default: 5 seconds).
Examples:
await-cmd.sh 30 /opt/scripts/check_remote_services.sh -- /opt/server/start.sh --port 8080
await-cmd.sh -w 10 30 /opt/scripts/check_remote_services.sh -- /opt/server/start.sh --port 8080Usage: await-http.sh [OPTION]... TIMEOUT URL... [-- COMMAND [ARG...]]
Repeatedly performs HTTP GET requests until the URL returns a HTTP status code <= 399. Then executes COMMAND.
Parameters:
TIMEOUT - Number of seconds within the URL must be reachable.
URL - URL(s) to be checked using HTTP GET.
COMMAND - Command to be executed once the wait condition is satisfied.
Options:
-f - Force execution of COMMAND even if timeout occurred.
-t SECS - Duration in seconds after which a connection attempt is aborted (optional, default: 10 seconds).
-w SECS - Duration in seconds to wait between retries (optional, default: 5 seconds).
Examples:
await-http.sh 30 http://service1.local -- /opt/server/start.sh --port 8080
await-http.sh 30 http://service1.local https://service2.local -- /opt/server/start.sh --port 8080
await-http.sh -w 10 30 https://service1.local -- /opt/server/start.sh --port 8080Usage: await-tcp.sh [OPTION]... TIMEOUT HOSTNAME:PORT... [-- COMMAND [ARG...]]
Repeatedly attempts to connect to the given address until the TCP port is available. Then executes COMMAND.
Parameters:
TIMEOUT - Duration in seconds within the TCP port of the given host must be reachable.
HOSTNAME:PORT - Target TCP address(es) to connect to.
COMMAND - Command to be executed once a connection could be established (optional).
Options:
-f - Force execution of COMMAND even if timeout occurred.
-t SECS - Duration in seconds after which a connection attempt is aborted (optional, default: 10 seconds).
-w SECS - Duration in seconds to wait between retries (optional, default: 5 seconds).
Examples:
await-tcp.sh 30 service1.local:389 -- /opt/server/start.sh --port 8080
await-tcp.sh 30 service1.local:389 service2.local:5672 -- /opt/server/start.sh --port 8080
await-tcp.sh -w 10 30 service1.local:389 -- /opt/server/start.sh --port 8080Here is an example stack (compose file) where two nginx servers are started in an ordered fashion.
The idea is to upload the await script into the config store, then mount into a service and override the default command. This way no additional Dockerfile/image needs to be created containing the script to support horizontal scaling.
version: '3.7'
configs:
# stores the script into the swarm config service, which automatically distributes
# it to the nodes running the containers.
await_http_script:
file: /opt/await/await-http.sh # existing path on the swarm master node
services:
backend_service:
image: karthequian/helloworld
deploy:
mode: replicated
replicas: 1
restart_policy:
condition: on-failure
delay: 5s
frontend_service:
image: karthequian/helloworld
configs:
# mount the script from config service into the container
- source: await_http_script
target: /await-http.sh
mode: 0555
command:
# wait up to 30 seconds for backend_service to become available, then start nginx
/await-http.sh 30 http://backend_service -- nginx
ports:
- 80:80
deploy:
mode: replicated
replicas: 1
restart_policy:
condition: on-failure
delay: 5sOne way to use the scripts with Kubernetes is creating a configmap and mount it to the container. So you needn't change your containers.
kubectl create configmap await-config --from-file=await-cmd.sh --from-file=await-http.sh --from-file=await-tcp.shKubernetes multi-container pods have the same resources, so we can check the MySQL port on localhost within our helloworld container.
apiVersion: apps/v1
kind: Deployment
metadata:
name: myapp-deployment
labels:
app: deployment
spec:
replicas: 1
selector:
matchLabels:
app: deployment
template:
metadata:
labels:
app: deployment
spec:
containers:
- image: mysql:5.6
name: mysql
env:
# Use secret in real usage
- name: MYSQL_ROOT_PASSWORD
value: password
ports:
- containerPort: 3306
name: mysql
- name: frontendservice
image: karthequian/helloworld
imagePullPolicy: Always
command: ["/bin/sh","-c", "echo 'start await' && cp /opt/scripts/*.sh /tmp && sh /tmp/await-tcp.sh 30 localhost:3306 -- nginx"]
ports:
- containerPort: 80
volumeMounts:
- name: await-volume
mountPath: /opt/scripts
volumes:
- name: await-volume
configMap:
name: await-config➜ kubectl logs myapp-deployment-67d6946f86-8qxwc frontendservice
start await
Waiting up to 30 seconds for [localhost:3306] to get ready...
=> executing [perl -e 'use IO::Socket;
my $socket=IO::Socket::INET->new(PeerAddr => "localhost", PeerPort => 3306, Timeout => 10);
if (defined $socket) {sleep 1; (defined $socket->connected?exit(0):exit(1))} else {exit(1)}']...ERROR
=> executing [perl -e 'use IO::Socket;
my $socket=IO::Socket::INET->new(PeerAddr => "localhost", PeerPort => 3306, Timeout => 10);
if (defined $socket) {sleep 1; (defined $socket->connected?exit(0):exit(1))} else {exit(1)}']...ERROR
=> executing [perl -e 'use IO::Socket;
my $socket=IO::Socket::INET->new(PeerAddr => "localhost", PeerPort => 3306, Timeout => 10);
if (defined $socket) {sleep 1; (defined $socket->connected?exit(0):exit(1))} else {exit(1)}']...OK
SUCCESS: Waiting condition is met.
Executing [nginx]...All files are released under the Apache License 2.0.
Individual files contain the following tag instead of the full license text:
SPDX-License-Identifier: Apache-2.0
This enables machine processing of license information based on the SPDX License Identifiers that are available here: https://spdx.org/licenses/.