SyncIt

Dave Redfern

Published: 21 Feb 01:20 in Tools

Introduction

Sync-It for Mutagen

SyncIt is a helper (phar archive) to make it easier to work with sessions when running mutagen.

SyncIt runs from a config file typically located in your project root named: .mutagen_sync_it.yaml. This file should be committed to your version control.

Inside the config file you can define any number of sync tasks. Each task should be a unique combination of a source (the alpha) and a target (beta).

SyncIt requires Mutagen >=0.10.0.

There is support for labels including project prefixes.

SyncIt has only been tested on macOS Mojave.

Features

simple yaml configuration
common options and ignore rules that are shared by tasks
all mutagen create flags are supported
labels
multiple tasks per project file
custom file location via a MUTAGEN_SYNC_IT_CONFIG env param
phar archive
support for .env files (including overrides)
docker container resolution from a specified name
groups to start/stop multiple tasks at the same time

Setup

Grab the phar archive and copy it to /usr/local/bin or add it to your path. Symlink the phar to syncit or a.n.other name.

Or install via brew:

brew install somnambulist-tech/somnambulist/syncit

In your project root e.g. ~/Projects/my-web-project run: syncit init to create a default, commented config file named .mutagen_sync_it.yaml. Tweak the settings.

Caution: before using mutagen ensure you have read and understood the docs. This is provided as-is without warranty of any kind. Use at your own risk!

Caution: mis-configuring a mutagen session can cause serious data-loss!

Removing SyncIt

Remove any symlinks you have and delete the phar file. No other files are created except for any config yaml files. If you installed using brew, then brew remove syncit. Again: use the following script at your own risk!

unlink /usr/local/bin/syncit && rm -v /usr/local/bin/mutagen-sync-it.phar

Protecting Yourself From File Overwrites

By way of some safe-guards you can configure a global mutagen config by adding a .mutagen.yml file to your home folder cd ~. This file defines global defaults that will be applied to all sessions (see mutagen.io for more details):

sync:
    defaults:
        mode: one-way-replica
        ignore:
            vcs: true
            paths:
                # System files
                - ".DS_Store"
                - "._*"
                  
                # Vim files
                - "*~"
                - "*.sw[a-p]"

                # Common folders and files
                - ".idea"
        symlink:
            mode: ignore
        permissions:
            defaultFileMode: 0644
            defaultDirectoryMode: 0755

The above ensures that all syncs are created using one-way-replica. This ensures that no changes are written back to the source but the target will be overwritten.

Note: that this will happily override your local in the source is set as a remote and the target is a local folder. Be very careful!

.env Files

SyncIt will read a .env file if it finds one in the project root. This is processed using Symfony DotEnv and will support .env.local and other overrides.

Building the phar archive

To build the phar archive, first checkout / clone the sync-it project, run composer install and ensure that phar.readonly is set to 0 in your php.ini.

You can then run: bin/compile which will create a mutagen-sync-it.phar file in the project root. The compile will output the SHA384 hash together with the file location / name.

Configuration

The Config File

The config file is split into 2 sections:

common
tasks

Sample Config

mutagen:
    common:
        label_prefix: ~

        options:
            default-directory-mode: '0755'
            default-file-mode: '0644'
            ignore-vcs: ~
            symlink-mode: ignore

        ignore:
            - ".DS_Store"
            - "._*"
            - ".idea/"
            - "vendor/"
            - "var/"

    tasks:
        source_files:
            source: "${PROJECT_DIR}"
            target: "docker://container_1/app"
            groups:
                - group1
                - group2
            options:
                sync-mode: one-way-replica
            ignore:
                - "composer.*"

        composer_json:
            source: "${PROJECT_DIR}/composer.json"
            target: "docker://container_1/app/composer.json"
            use_common: false
            options:
                sync-mode: two-way-safe

        composer_lock:
            source: "${PROJECT_DIR}/composer.lock"
            target: "docker://container_1/app/composer.lock"
            use_common: false
            options:
                sync-mode: two-way-safe

The config file will expand any configured env args using ${ENV_NAME} notation. ${PROJECT_DIR} is an alias of ${PWD}. This expansion is done at run time only. For a full list of dedicated env vars and the key to use to access it run: syncit params

Note: no attempt is made to hide or mask sensitive env vars. SyncIt is intended for local dev usage only.

Common

The common area (not to be confused with global from the .mutagen.toml file); allows settings to be shared in the current project. Here you can add safe guards if you share this with other people e.g.: one-way-replica file permissions etc.

Options are the mutagen create flag names with the leading --. For flags with no value, they must have a ~ as the value. Flags that require a value where one is not set are not passed through to mutagen.

Read more: https://mutagen.io/documentation or use the --help option on the mutagen command line program e.g.: mutagen create --help.

Since 1.0.0-beta1 label_prefix has been added to the common section. This allows a common prefix e.g. a project name to be prefixed to all the task labels. Useful if you use common labels with multiple projects.

Tasks

The tasks define each sync task. These are either folder copies or single file copies. In the example above, the project source files are being copied to the /app folder in a docker container, excluding any composer files.

The remaining 2 tasks, ensure that changes to composer.json/lock are sync'd from the container back to the local dev. This ensures that composer can be run in the container and on the local machine.

In both composer cases the common options are ignored.

Each task must have a unique name and the source and target must be specified. The name is used as the label, and if running a version of mutagen >=0.9.0 this will be passed as the label when appropriate. Each task can optionally override any existing option defined in common, prevent common settings being used, and specify additional ignore rules.

Tasks will be matched to running sessions based on the source and target. This allows the task name to be used consistently regardless of mutagen version.

The target supports different transport mechanisms e.g. docker:// ssh:// etc. Be sure to read the format / rules at: https://mutagen.io/documentation/transports/

Multiple group names can be set on each tasks under the groups: entry. This will allow start/stop to work with all those tasks tagged with that group name. For example: a web app with a JavaScript build pipeline may have many tasks. These can now be grouped into app/build making it easier to start/stop in one go.

Docker Containers

When using docker; to make life easier you should use named containers with predictable / repeatable names via the --name flag on docker run or use docker compose and use a service name or, especially in dev, a specific name can be set using container_name:. If the latter is chosen, the container name will be exactly this value but there can only be one of them.

Alternatively: the target container name can be set to use a name that will attempt to be resolved from the running available containers as defined in the output from docker ps. To do this, change the container name to be: {docker:name=<some_keywords_to_match>} - don't include the <>. The name can contain any valid characters that a container name can have. The default that will be substituted is the matching container hex ID. Alternatively the matching container name can be used by adding :name to the string.

Doing this is the equivalent of performing:

$ docker ps --no-trunc --format="{{.ID}}" --filter=name="my-container"
$ docker ps --no-trunc --format="{{.Names}}" --filter=name="my-container"

The configuration using the name resolution would look like the following:

composer_lock:
    source: "${PROJECT_DIR}/composer.lock"
    target: "docker://{docker:name=my-container:name}/app/composer.lock"

The resulting output from "view" would then contain:

| Target (beta)        | docker://1fac35046452b0f6d7de0167399d6f4f7f68968ca3a844a385d |
|                      | b5ff361df4717/app                                            |

While using :name would give:

| Target (beta)        | docker://my-project_my-container_1/app                       |

Note: the chosen name must result in only 1 container. If none or more than one are found, SyncIt will raise an error. In the case of multiple matches, all matched names will be in the error (useful when used with :name).

Note: name resolution is performed after .env resolution. So you can use a ENV parameter for the container name and share it between the SyncIt config file and a docker-compose.yml file.

Managing Tasks

SyncIt acts as a wrapper over mutagen create|terminate|list in a pretty basic manner. Before continuing you should start the mutagen daemon: mutagen daemon start. Many commands will not run if the mutagen process is not running. If syncit detects mutagen is not running, it will attempt to start it automatically.

To start all configured tasks: syncit start. You will be prompted to start a specific task, or everything. Or: syncit start all

To start a specific task: syncit start --label=<label>

To stop all configured tasks: syncit stop. You will be prompted to stop a specific task, or everything. Or: syncit stop all

To stop a specific task: syncit stop --label=<label>

Extra information can be obtained by running: syncit view <label> or if no label is given, the available labels will be displayed.

To get an overview of running tasks: syncit status

Note: sessions are mapped to tasks by the combination of source and target; if these paths do not match the session will not be linked. Once labels are fully available, this limitation should be removed. In particular this means trailing slashes e.g.: source: ${PROJECT_DIR}/ should be removed as mutagen will remove it anyway.

To get detailed information on a task, run: syncit view You will be prompted for the task to investigate, and then you will see various status information about a given task; and how to get debug information directly from Mutagen.

Return to article