moonfire-nvr/guide/build.md

Building Moonfire NVR

This document has notes for software developers on building Moonfire NVR from source code for development. If you just want to install precompiled binaries, see the Docker installation instructions instead.

This document doesn't spell out as many details as the installation instructions. Please ask on Moonfire NVR's issue tracker or mailing list when stuck. Please also send pull requests to improve this doc.

Downloading

See the github page (in case you're not reading this text there already). You can download the bleeding-edge version from the commandline via git:

$ git clone https://github.com/scottlamb/moonfire-nvr.git

Docker builds

This command should prepare a deployment image for your local machine:

$ docker buildx build --load --tag=moonfire-nvr -f docker/Dockerfile .

If you want to iterate on code changes, doing a full Docker build from scratch every time will be painfully slow. You will likely find it more helpful to use the dev target. This is a self-contained developer environment which you can use from its shell via docker run or via something like Visual Studio Code's Docker plugin.

$ docker buildx build --load --tag=moonfire-dev --target=dev 
...
$ docker run \
        --rm --interactive=true --tty \
        --mount=type=bind,source=$(pwd),destination=/var/lib/moonfire-nvr/src \
        moonfire-dev

The development image overrides cargo's output directory to /var/lib/moonfire-nvr/target. (See ~moonfire-nvr/.buildrc.) This avoids using a bind filesystem for build products, which can be slow on macOS. It also means that if you sometimes compile directly on the host and sometimes within Docker, they don't trip over each other's target directories. directories.

You can also cross-compile to a different architecture. Adding a --platform=linux/arm64/v8,linux/arm/v7,linux/amd64 argument will compile Moonfire NVR for all supported platforms. For the dev target, this prepares a build which executes on your local architecture and is capable of building a binary for your desired target architecture.

On the author's macOS machine with Docker desktop 3.0.4, building for multiple platforms at once will initially fail with the following error:

$ docker buildx build ... --platform=linux/arm64/v8,linux/arm/v7,linux/amd64
[+] Building 0.0s (0/0)
error: multiple platforms feature is currently not supported for docker driver. Please switch to a different driver (eg. "docker buildx create --use")

Running docker buildx create --use once solves this problem, with a couple caveats:

• you'll need to specify an additional --load argument to make builds available to run locally.
• the --load argument only works for one platform at a time. With multiple platforms, it gives an error like the following:

  error: failed to solve: rpc error: code = Unknown desc = docker exporter does not currently support exporting manifest lists
  

  A comment on docker/buildx issue #59 suggests a workaround of building all three then using caching to quickly load the one of immediate interest:

  $ docker buildx build --platform=linux/arm64/v8,linux/arm/v7,linux/amd64 ...
  $ docker buildx build --load --platform=arm64/v8 ...
  

Major caveat: something appears to be making docker buildx build frequently redo builds rather than reusing cached results. The author is very annoyed by this and would welcome help in understanding this problem...

Non-Docker setup

You may prefer building without Docker on the host. Moonfire NVR should run natively on any Unix-like system. It's been tested on Linux and macOS. (In theory Windows Subsystem for Linux should also work. Please speak up if you try it.)

On macOS systems native builds may be noticeably faster than using Docker's Linux VM and filesystem overlay.

To build the server, you will need the following C libraries installed:

• ffmpeg version 2.x or 3.x, including libavutil, libavcodec (to inspect H.264 frames), and libavformat (to connect to RTSP servers and write .mp4 files).

  Note ffmpeg library versions older than 55.1.101, along with all versions of the competing project libav, don't support socket timeouts for RTSP. For reliable reconnections on error, it's strongly recommended to use ffmpeg library versions >= 55.1.101.

• SQLite3.

• ncursesw, the UTF-8 version of the ncurses library.

On recent Ubuntu or Raspbian Linux, the following command will install all non-Rust dependencies:

$ sudo apt-get install \
               build-essential \
               libavcodec-dev \
               libavformat-dev \
               libavutil-dev \
               libncurses-dev \
               libsqlite3-dev \
               pkgconf \
               sqlite3 \
               tzdata

On macOS with Homebrew and Xcode installed, try the following command:

$ brew install ffmpeg yarn

Next, you need Rust 1.45+ and Cargo. The easiest way to install them is by following the instructions at rustup.rs.

Finally, building the UI requires yarn. (On macOS, the brew command above installs it for you. On Linux, follow yarn's guide.)

Once prerequisites are installed, you can build the server and find it in target/release/moonfire-nvr:

$ cargo test
$ cargo build --release

You can build the UI via yarn and find it in the ui-dist directory:

$ yarn
$ yarn build

Running interactively straight from the working copy

The author finds it convenient for local development to set up symlinks so that the binaries in the working copy will run via just nvr:

$ sudo mkdir /usr/local/moonfire-nvr
$ sudo ln -s `pwd`/ui-dist /usr/local/moonfire-nvr/ui
$ sudo mkdir /var/lib/moonfire-nvr
$ sudo chown $USER:$USER /var/lib/moonfire-nvr
$ ln -s `pwd`/target/release/moonfire-nvr $HOME/bin/moonfire-nvr 
$ ln -s moonfire-nvr $HOME/bin/nvr
$ nvr init
$ nvr config
$ nvr run

(Alternatively, you could symlink to target/debug/moonfire-nvr and compile with cargo build rather than cargo build --release, for a faster build cycle and slower performance.)

Note this nvr is a little different than the nvr shell script you create when following the install instructions. With that shell wrapper, nvr run will create and run a detached Docker container with some extra arguments specified in the script. This nvr run will directly run from the terminal, with no extra arguments, until you abort with Ctrl-C. Likewise, some of the shell script's subcommands that wrap Docker (start, stop, and logs) have no parallel with this nvr.

Running as a systemd service

If you want to deploy a non-Docker build on Linux, you may want to use systemd. Create /etc/systemd/system/moonfire-nvr.service:

[Unit]
Description=Moonfire NVR
After=network-online.target

[Service]
ExecStart=/usr/local/bin/moonfire-nvr run \
    --db-dir=/var/lib/moonfire-nvr/db \
    --http-addr=0.0.0.0:8080 \
    --allow-unauthenticated-permissions='view_video: true'
Environment=TZ=:/etc/localtime
Environment=MOONFIRE_FORMAT=google-systemd
Environment=MOONFIRE_LOG=info
Environment=RUST_BACKTRACE=1
Type=simple
User=moonfire-nvr
Nice=-20
Restart=on-failure
CPUAccounting=true
MemoryAccounting=true
BlockIOAccounting=true

[Install]
WantedBy=multi-user.target

Note that the arguments used here are insecure. You can change that via replacing the --allow-unauthenticated-permissions argument here as described in Securing Moonfire NVR and exposing it to the Internet.

Some handy commands:

$ sudo systemctl daemon-reload                                  # reload configuration files
$ sudo systemctl start moonfire-nvr                             # start the service now
$ sudo systemctl stop moonfire-nvr                              # stop the service now (but don't wait for it finish stopping)
$ sudo systemctl status moonfire-nvr                            # show if the service is running and the last few log lines
$ sudo systemctl enable moonfire-nvr                            # start the service on boot
$ sudo systemctl disable moonfire-nvr                           # don't start the service on boot
$ sudo journalctl --unit=moonfire-nvr --since='-5 min' --follow # look at recent logs and await more

See the systemd documentation for more information. The manual pages for systemd.service and systemctl may be of particular interest.