Merge pull request #1058 from arianacosta/integrating_docker_sync

Integrating docker sync
This commit is contained in:
Mahmoud Zalt 2017-07-16 22:26:08 +03:00 committed by GitHub
commit f258b38a18
6 changed files with 196 additions and 9 deletions

1
.gitignore vendored
View File

@ -3,3 +3,4 @@
/data /data
.env .env
/.project /.project
.docker-sync

View File

@ -156,21 +156,82 @@ You might use the `--no-cache` option if you want full rebuilding (`docker-compo
<br> <br>
<a name="Docker-Sync"></a> <a name="Docker-Sync"></a>
## Docker-Sync ## Speed up with docker-sync
Docker on the Mac [is slow](https://github.com/docker/for-mac/issues/77), at the time of writing. Especially for larger projects, this can be a problem. The problem is [older than March 2016](https://forums.docker.com/t/file-access-in-mounted-volumes-extremely-slow-cpu-bound/8076) - as it's a such a long-running issue, we're including it in the docs here. Docker for Mac is [slow](https://github.com/docker/for-mac/issues/77) due to poor performance when the application accesses files shared with the host machine.
One solution is to use [docker-sync](https://github.com/EugenMayer/docker-sync).
The problem originates in bind-mount performance on MacOS. Docker for Mac uses osxfs by default. This is not without reason, it has [a lot of advantages](https://docs.docker.com/docker-for-mac/osxfs/). In simple terms, docker-sync creates a docker container with a copy of all the application files that can be accessed very quickly from the other containers.
On the other hand, docker-sync runs a process on the host machine that continuously tracks and updates files changes from the host to this intermediate container.
Solutions to resolve this issue are easily installed however, we're hoping it'll be fixed by Docker themselves over time. They are currently [adding "cached and delegated" options](https://github.com/docker/for-mac/issues/77#issuecomment-283996750), which is partly available for Docker Edge. Out of the box, it comes pre-configured for OS X, but using it on Windows is very easy to set-up by modifying the `DOCKER_SYNC_STRATEGY` on the `.env`
Options are [to switch over to NFS](https://github.com/IFSight/d4m-nfs) which is the simplest. The fastest option is [Docker-Sync "native"](https://github.com/EugenMayer/docker-sync) which is still quite easy to install. #### Usage
Clone [this repo](https://github.com/EugenMayer/docker-sync-boilerplate) to your machine, copy `default/docker-sync.yml` to your Laradock directory and run `docker-sync-stack start`. Be sure to use `docker-sync-stack clean` to stop and `docker-compose build` to rebuild. More information can be found [in the Docker-sync docs](https://github.com/EugenMayer/docker-sync). Laradock comes with `sync.sh`, an optional bash script, that automates installing, running and stopping docker-sync. Note that to run the bash script you may need to change the permissions `chmod 755 sync.sh`
1) Configure your Laradock environment as you would normally do and test your application to make sure that your sites are running correctly.
2) Make sure to set `DOCKER_SYNC_STRATEGY` on the `.env`. Read the [syncing strategies](https://github.com/EugenMayer/docker-sync/wiki/8.-Strategies) for details.
```
# osx: 'native_osx' (default)
# windows: 'unison'
# linux: docker-sync not required
DOCKER_SYNC_STRATEGY=native_osx
```
2) Install the docker-sync gem on the host-machine:
```bash
./sync.sh install
```
3) Start docker-sync and the Laradock environment.
Specify the services you want to run, as you would normally do with `docker-compose up`
```bash
./sync.sh up nginx mysql
```
Please note that the first time docker-sync runs, it will copy all the files to the intermediate container and that may take a very long time (15min+).
4) To stop the environment and docker-sync do:
```bash
./sync.sh down
```
#### Setting up Aliases (optional)
You may create bash profile aliases to avoid having to remember and type these commands for everyday development.
Add the following lines to your `~/.bash_profile`:
```bash
alias devup="cd /PATH_TO_LARADOCK/laradock; ./sync.sh up nginx mysql" #add your services
alias devbash="cd /PATH_TO_LARADOCK/laradock; ./sync.sh bash"
alias devdown="cd /PATH_TO_LARADOCK/laradock; ./sync.sh down"
```
Now from any location on your machine, you can simply run `devup`, `devbash` and `devdown`.
#### Additional Commands
Opening bash on the workspace container (to run artisan for example):
```bash
./sync.sh bash
```
Manually triggering the synchronization of the files:
```bash
./sync.sh sync
```
Removing and cleaning up the files and the docker-sync container. Use only if you want to rebuild or remove docker-sync completely. The files on the host will be kept untouched.
```bash
./sync.sh clean
```
**Additional Notes:**
- You may run laradock with or without docker-sync at any time using with the same `.env` and `docker-compose.yml`, because the configuration is overridden automatically when docker-sync is used.
- You may inspect the `sync.sh` script to learn each of the commands and even add custom ones.
- If a container cannot access the files on docker-sync, you may need to set a user on the Dockerfile of that container with an id of 1000 (this is the UID that nginx and php-fpm have configured on laradock). Alternatively, you may change the permissions to 777, but this is **not** recommended.
Visit the [docker-sync documentation](https://github.com/EugenMayer/docker-sync/wiki) for more details.
<br> <br>
<a name="Add-Docker-Images"></a> <a name="Add-Docker-Images"></a>

17
docker-compose.sync.yml Normal file
View File

@ -0,0 +1,17 @@
version: '2'
services:
### Applications Code Container #############################
applications:
image: tianon/true
volumes:
- applications-sync:/var/www:nocopy # nocopy is required
### Volumes Setup #############################################
volumes:
applications-sync:
external:
name: "applications-docker-sync"

13
docker-sync.yml Normal file
View File

@ -0,0 +1,13 @@
version: "2"
options:
verbose: true
syncs:
applications-docker-sync: # name of the intermediary sync volume
compose-dev-file-path: 'docker-compose.sync.yml' # docker-compose override file
src: '${APPLICATION}' # host source directory
sync_userid: 1000 # giving permissions to www-data user (as defined in nginx and php-fpm Dockerfiles)
sync_strategy: '${DOCKER_SYNC_STRATEGY}' # for osx use 'native_osx', for windows use 'unison'
sync_excludes: ['laradock', 'ignored_folder_example'] # ignored directories

View File

@ -255,6 +255,13 @@ CADDY_CUSTOM_CADDYFILE=./caddy/Caddyfile
LARAVEL_ECHO_SERVER_PORT=6001 LARAVEL_ECHO_SERVER_PORT=6001
### DOCKER-SYNC ################################################################################################
# osx: 'native_osx' (default)
# windows: 'unison'
# linux: docker-sync not required
DOCKER_SYNC_STRATEGY=native_osx
##### TO BE CONTINUE ................................. ##### TO BE CONTINUE .................................
@ -262,8 +269,6 @@ LARAVEL_ECHO_SERVER_PORT=6001
# ......... # .........
# ......... # .........
############################ ############################
# Miscellaneous # Miscellaneous
############################ ############################
@ -281,3 +286,4 @@ PHP_IDE_CONFIG=serverName=laradock
# Fix for windows users to make sure the application path works. # Fix for windows users to make sure the application path works.
COMPOSE_CONVERT_WINDOWS_PATHS=1 COMPOSE_CONVERT_WINDOWS_PATHS=1

89
sync.sh Executable file
View File

@ -0,0 +1,89 @@
#!/bin/bash
# This shell script is an optional tool to simplify
# the installation and usage of laradock with docker-sync.
# Make sure that the DOCKER_SYNC_STRATEGY is set in the .env
# DOCKER_SYNC_STRATEGY=native_osx # osx
# DOCKER_SYNC_STRATEGY=unison # windows
# To run, make sure to add permissions to this file:
# chmod 755 sync.sh
# USAGE EXAMPLE:
# Install docker-sync: ./sync.sh install
# Start sync and services with nginx and mysql: ./sync.sh up nginx mysql
# Stop containers and sync: ./sync.sh down
# prints colored text
print_style () {
if [ "$2" == "info" ] ; then
COLOR="96m"
elif [ "$2" == "success" ] ; then
COLOR="92m"
elif [ "$2" == "warning" ] ; then
COLOR="93m"
elif [ "$2" == "danger" ] ; then
COLOR="91m"
else #default color
COLOR="0m"
fi
STARTCOLOR="\e[$COLOR"
ENDCOLOR="\e[0m"
printf "$STARTCOLOR%b$ENDCOLOR" "$1"
}
display_options () {
printf "Available options:\n";
print_style " install" "info"; printf "\t\t Installs docker-sync gem on the host machine.\n"
print_style " up [services]" "success"; printf "\t Starts docker-sync and runs docker compose.\n"
print_style " down" "success"; printf "\t\t\t Stops containers and docker-sync.\n"
print_style " bash" "success"; printf "\t\t\t Opens bash on the workspace.\n"
print_style " sync" "info"; printf "\t\t\t Manually triggers the synchronization of files.\n"
print_style " clean" "danger"; printf "\t\t Removes all files from docker-sync.\n"
}
if [[ $# -eq 0 ]] ; then
print_style "Missing arguments.\n" "danger"
display_options
exit 1
fi
if [ "$1" == "up" ] ; then
print_style "Initializing Docker Sync\n" "info"
print_style "May take a long time (15min+) on the first run\n" "info"
docker-sync start;
print_style "Initializing Docker Compose\n" "info"
shift # removing first argument
docker-compose -f docker-compose.yml -f docker-compose.sync.yml up -d ${@}
elif [ "$1" == "down" ]; then
print_style "Stopping Docker Compose\n" "info"
docker-compose down
print_style "Stopping Docker Sync\n" "info"
docker-sync stop
elif [ "$1" == "bash" ]; then
docker-compose exec workspace bash
elif [ "$1" == "install" ]; then
print_style "Installing docker-sync\n" "info"
gem install docker-sync
elif [ "$1" == "sync" ]; then
print_style "Manually triggering sync between host and docker-sync container.\n" "info"
docker-sync sync;
elif [ "$1" == "clean" ]; then
print_style "Removing and cleaning up files from the docker-sync container.\n" "warning"
docker-sync clean
else
print_style "Invalid arguments.\n" "danger"
display_options
exit 1
fi