diff --git a/.agents/skills/test-release-canary/SKILL.md b/.agents/skills/test-release-canary/SKILL.md index 4bf7d38ae3..a7234368b2 100644 --- a/.agents/skills/test-release-canary/SKILL.md +++ b/.agents/skills/test-release-canary/SKILL.md @@ -12,7 +12,9 @@ The Release Canary (`.github/workflows/release-canary.yml`) smoke-tests the arti | Job | Runner | Verifies | |---|---|---| | `macos` | `macos-latest-xlarge` | `install.sh` resolves the Homebrew formula, brew installs the cask, and `openshell status` reaches the brew-services–backed local gateway with the VM driver. | -| `ubuntu` | `ubuntu-latest` | `install.sh` installs the Debian package, the post-install systemd user service starts, and `openshell status` reaches the local gateway with the Docker driver. | +| `ubuntu-deb-native-docker` | `ubuntu-latest` | `install.sh` auto-detects the Debian package when both snapd and native Docker are present, installs the deb, the post-install systemd user service starts, and `openshell status` reaches the local gateway with the Docker driver. Once snapd 2.77 is released to `latest/stable`, `install.sh` should install the `openshell` snap here instead, even with native Docker present. This job should then be renamed to `ubuntu-snap-native-docker`, and its purpose will be to verify that the openshell snap installs and works when a native (non-snap) Docker is already present on the system. | +| `ubuntu-deb` | `ubuntu-latest` | `install.sh` installs the Debian package after snapd is fully removed (forcing the deb path), starts the systemd user service, and `openshell status` reaches the local gateway with the Docker driver. | +| `ubuntu-snap-without-docker` | `ubuntu-latest` | `install.sh` auto-detects and installs both the `openshell` and `docker` snaps from the Snap Store when snapd is present and there is no Docker daemon on the system, the gateway daemon starts with auto-connected interfaces, and `openshell status` reaches the local gateway. | | `fedora` | `fedora:latest` container | `install.sh` installs the RPM packages, the local gateway starts under Podman, and `openshell status` succeeds. | | `kubernetes` | `ubuntu-latest` + kind | `helm install oci://ghcr.io/nvidia/openshell/helm-chart --version 0.0.0-dev` succeeds in a kind cluster, the gateway pod becomes Ready, port-forward exposes 8080, and the released CLI registers the in-cluster gateway and runs `openshell status` against it. | @@ -107,8 +109,9 @@ Loopback registration auto-derives the gateway name to `openshell` if `--name` i | Symptom | Likely cause | Where to look | |---|---|---| -| `macos`/`ubuntu`/`fedora` job fails on `install.sh` | Latest tagged release missing an asset, checksum mismatch, or `install.sh` regression on this branch. | Job log around the `curl … install.sh \| sh` step. | -| `macos`/`ubuntu`/`fedora` job fails on `openshell status` | Local gateway service did not start (systemd/brew/podman). Often a driver issue. | Service logs in the job log; `OPENSHELL_DRIVERS` env in the "Ensure …" step. | +| `macos`/`ubuntu-deb-native-docker`/`ubuntu-deb`/`fedora` job fails on `install.sh` | Latest tagged release missing an asset, checksum mismatch, or `install.sh` regression on this branch. | Job log around the `curl … install.sh \| sh` step. | +| `macos`/`ubuntu-deb-native-docker`/`ubuntu-deb`/`fedora` job fails on `openshell status` | Local gateway service did not start (systemd/brew/podman). Often a driver issue. | Service logs in the job log; `OPENSHELL_DRIVERS` env in the "Ensure …" step. | +| `ubuntu-snap-without-docker` job fails on `openshell status` | Snap gateway daemon did not start. | `snap services openshell` and `snap logs openshell.gateway` in the job log. | | `kubernetes` job fails on `helm install --wait` | Chart did not deploy in 5 min — usually image pull failure or readiness probe failing. | "Diagnostics on failure" step dumps `helm status`, manifest, pod describe, pod logs. | | `kubernetes` job fails on `kubectl wait` | Gateway pod stuck `CrashLoopBackOff` or `ImagePullBackOff`. | Diagnostics dump; check `:dev` image existence at `ghcr.io/nvidia/openshell/gateway`. | | `kubernetes` job fails on `openshell gateway add` or `status` | Port-forward not reachable, or CLI/gateway proto mismatch. | `port-forward.log` and `openshell gateway list` in the diagnostics dump. | diff --git a/.github/workflows/release-canary.yml b/.github/workflows/release-canary.yml index 896d12d190..f7fbec3c84 100644 --- a/.github/workflows/release-canary.yml +++ b/.github/workflows/release-canary.yml @@ -30,11 +30,17 @@ jobs: curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/${{ github.event.workflow_run.head_sha || github.sha }}/install.sh | sh openshell status - ubuntu: - name: Ubuntu Docker + ubuntu-deb-native-docker: + name: Ubuntu (native Docker) if: ${{ github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success' }} runs-on: ubuntu-latest timeout-minutes: 20 + # TODO: Once snapd 2.77 is released to latest/stable, the openshell snap + # will be able to connect to system Docker via the docker interface, + # removing the requirement for the docker snap. At that point, + # install.sh should select snap here even with native Docker present. + # Rename this job to `ubuntu-snap-native-docker`, replace the dpkg + # assertion below with `snap list openshell`. steps: - name: Ensure Docker run: | @@ -50,6 +56,82 @@ jobs: - name: Install and check status run: | curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/${{ github.event.workflow_run.head_sha || github.sha }}/install.sh | sh + # Expected to fail once snapd 2.77 is in latest/stable, at which + # point install.sh should select snap instead of deb here. + dpkg -l openshell | grep '^ii' + openshell --version + systemctl --user is-active openshell-gateway + openshell status + + ubuntu-deb: + name: Ubuntu Debian package + if: ${{ github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success' }} + runs-on: ubuntu-latest + timeout-minutes: 20 + steps: + - name: Remove snapd + run: | + set -euo pipefail + # Remove all snaps in a single transaction. Core bases and content + # providers cannot be removed before snaps that consume them, so + # removing one at a time fails; snap handles ordering within a + # single `snap remove` invocation. + snaps=$(snap list 2>/dev/null | awk 'NR>1 {print $1}') + if [ -n "$snaps" ]; then + sudo snap remove $snaps + fi + sudo apt-get -y --purge remove snapd + ! command -v snap >/dev/null 2>&1 + + - name: Ensure Docker + run: | + if ! command -v docker >/dev/null 2>&1; then + sudo apt-get update + sudo apt-get install -y docker.io + fi + sudo systemctl start docker || sudo service docker start + mkdir -p "${HOME}/.config/openshell" + printf 'OPENSHELL_DRIVERS=docker\n' > "${HOME}/.config/openshell/gateway.env" + docker info + + - name: Install and check status + run: | + curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/${{ github.event.workflow_run.head_sha || github.sha }}/install.sh | sh + dpkg -l openshell | grep '^ii' + openshell --version + systemctl --user is-active openshell-gateway + openshell status + + ubuntu-snap-without-docker: + name: Ubuntu without Docker preinstalled + if: ${{ github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success' }} + runs-on: ubuntu-latest + timeout-minutes: 20 + steps: + - name: Ensure snapd + run: | + set -euo pipefail + sudo apt-get update + sudo apt-get install -y snapd + sudo systemctl enable --now snapd.socket + sudo systemctl start snapd + sudo snap wait system seed.loaded + + - name: Remove native Docker + run: | + set -euo pipefail + sudo systemctl stop docker.service docker.socket 2>/dev/null || true + sudo apt-get -y --purge remove docker-ce docker-ce-cli containerd.io \ + docker-buildx-plugin docker-compose-plugin docker.io 2>/dev/null || true + ! command -v docker >/dev/null 2>&1 + + - name: Install and check status + run: | + set -euo pipefail + curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/${{ github.event.workflow_run.head_sha || github.sha }}/install.sh | sh + snap list openshell + openshell --version + sudo snap services openshell openshell status fedora: @@ -145,56 +227,6 @@ jobs: run: | docker rm -f "${FEDORA_CANARY_CONTAINER}" >/dev/null 2>&1 || true - ubuntu-snap: - name: Ubuntu Snap - if: ${{ github.event.workflow_run.conclusion == 'success' }} - runs-on: ubuntu-latest - timeout-minutes: 20 - steps: - - name: Install snapd - run: | - set -euo pipefail - sudo apt-get update - sudo apt-get install -y snapd - sudo systemctl enable --now snapd.socket - sudo systemctl start snapd - sudo snap wait system seed.loaded - - - name: Install Docker snap - run: | - set -euo pipefail - sudo snap install docker - - - name: Download snap from release-dev artifacts - uses: actions/download-artifact@3e5f45b2cfb9172054b4087a40e8e0b5a5461e7c # v8.0.1 - with: - github-token: ${{ github.token }} - run-id: ${{ github.event.workflow_run.id }} - pattern: snap-linux-amd64 - path: release/ - merge-multiple: true - - - name: Install snap (dangerous — from release, not store) - run: | - set -euo pipefail - sudo snap install ./release/*.snap --dangerous - - - name: Connect interfaces - run: | - set -euo pipefail - sudo snap connect openshell:docker docker:docker-daemon - sudo snap connect openshell:log-observe - sudo snap connect openshell:system-observe - - - name: Register snap gateway and check status - run: | - set -euo pipefail - openshell --version - sudo snap services openshell - openshell gateway add http://127.0.0.1:17670 --local --name snap-docker - openshell gateway select snap-docker - openshell status - kubernetes: name: Kubernetes Helm (kind) if: ${{ github.event_name == 'workflow_dispatch' || github.event.workflow_run.conclusion == 'success' }} diff --git a/docs/about/installation.mdx b/docs/about/installation.mdx index 2ac077e7b9..4288f52ac1 100644 --- a/docs/about/installation.mdx +++ b/docs/about/installation.mdx @@ -16,7 +16,10 @@ Install OpenShell with a single command: curl -LsSf https://raw.githubusercontent.com/NVIDIA/OpenShell/main/install.sh | sh ``` -The script detects your operating system and installs the OpenShell CLI and gateway with your native package manager. It then starts the local gateway server so you can begin creating sandboxes. +The script detects your operating system and installs the OpenShell CLI and +gateway. On Linux, the Snap path is preferred when `snapd` is available; +otherwise the script uses the native DEB or RPM package. The gateway then +starts automatically so you can begin creating sandboxes. You can also download release artifacts directly from the [OpenShell GitHub Releases](https://github.com/NVIDIA/OpenShell/releases) page. @@ -51,9 +54,15 @@ brew services restart openshell ## Linux +On distributions that ship with `snapd`, the install script uses the Snap path +described below, unless one of the following is true: + +- `snapd` is not present +- a native (non-snap) Docker package is present (the `openshell` snap requires the `docker` snap, until snapd 2.77 is released to the `latest/stable` channel) + On Fedora and RHEL, the install script uses RPM packages. The RPM installs the `openshell` CLI, the `openshell-gateway` daemon, and a systemd user service. -On Debian and Ubuntu, the install script uses a Debian package. The Debian package installs the `openshell` CLI, the `openshell-gateway` daemon, VM sandbox support, and a systemd user service. +On Debian and Ubuntu (where the snap is not applicable), the install script uses a Debian package. The Debian package installs the `openshell` CLI, the `openshell-gateway` daemon, VM sandbox support, and a systemd user service. Linux packages require glibc 2.28 or newer. The installer checks libc before downloading packages and exits with an error on older glibc versions, Alpine, musl-based distributions, or unknown libc environments. @@ -77,7 +86,12 @@ sudo loginctl enable-linger $USER ## Snap -Install the OpenShell snap from the Snap Store: +On Linux distributions that ship with `snapd`, the install script installs +OpenShell from the Snap Store. The OpenShell snap bundles the CLI, the terminal +UI, and a managed gateway daemon. `snapd` handles upgrades and rollback; the +gateway runs as a system service inside the snap. + +You can also install the snap directly: ```shell sudo snap install openshell @@ -94,19 +108,24 @@ typically `~/snap/openshell/common`. Gateway registrations live under `$SNAP_USER_COMMON/.config/openshell/gateways/` instead of `~/.config/openshell/gateways/`. -### Snap store installs - -When installing from the Snap Store, snapd automatically connects the `home`, -`network`, and `network-bind` plugs. The `docker` plug still -requires manual connection: +The OpenShell snap requires the Docker snap, which you can install with: ```shell -sudo snap connect openshell:docker docker:docker-daemon +sudo snap install docker ``` -The snap declares `default-provider: docker` on the Docker plug so snapd will -offer to install the Docker snap, but the connection itself must be made -manually. +### Snap store installs + +When installing from the Snap Store, snapd automatically connects the `docker`, +`home`, `log-observe`, `network`, `network-bind`, and `system-observe plugs. + +The `docker` interface is used to connect the `openshell` snap's `docker` plug +to the `docker` snap's `docker-daemon` slot; the `openshell` snap does not yet +work with a host-installed Docker Engine. The installer runs this and the other +interface connections for you after installing the `openshell` snap; run them +by hand if you install the snap manually. The installer is best-effort: if a +connect fails (for example because the `docker` snap is not yet running), the +snap still installs and the installer prints a warning. ### Locally built snap packages @@ -127,6 +146,26 @@ to read logs and inspect system processes. The `docker` plug requires the `docker:docker-daemon` slot from the Docker snap and does not work with system-installed Docker. +### Verify the gateway + +The snap-managed gateway service is `openshell.gateway`. Inspect it with: + +```shell +snap services openshell +snap logs -n 100 openshell.gateway +``` + +Register the gateway with the CLI: + +```shell +openshell gateway add http://127.0.0.1:17670 --local --name openshell +openshell status +``` + +The gateway listens on `http://127.0.0.1:17670` and stores its state under +`/var/snap/openshell/common/`. Override gateway settings by creating +`/var/snap/openshell/common/gateway.toml`. + ### Gateway service The gateway runs as a snap daemon with `refresh-mode: endure`, meaning snapd @@ -138,6 +177,17 @@ a snap refresh when you need the updated binary: sudo systemctl restart snap.openshell.gateway ``` +### When to choose Snap + +Use the `openshell` snap when `snapd` is available and no native Docker Engine +is installed, and you want atomic upgrades and rollback, a single self-contained +and sandboxed install, or a desktop launcher that surfaces the OpenShell +terminal UI in the application menu. + +Use the `openshell` `.deb` or `.rpm` file when `snapd` is unavailable or when +you already run Docker from a non-snap source. The installer falls back to +these methods on hosts with native Docker. + ## Kubernetes Kubernetes deployments use the OpenShell Helm chart. For step-by-step installation, refer to [Kubernetes Setup](/kubernetes/setup). For chart values and packaging details, refer to the [Helm chart README](https://github.com/NVIDIA/OpenShell/blob/main/deploy/helm/openshell/README.md). diff --git a/install.sh b/install.sh index 6cda59f8bc..028af8de22 100755 --- a/install.sh +++ b/install.sh @@ -60,9 +60,10 @@ NOTES: from ${GITHUB_URL}/releases/latest. Linux installs the Debian package on amd64/arm64 or the RPM packages on - x86_64/aarch64, depending on the host package manager. - macOS installs the release Homebrew formula on Apple Silicon and starts a - brew services-backed local gateway. + x86_64/aarch64, depending on the host package manager. When snapd is + available and no native Docker Engine is installed, the installer prefers + the Snap path. macOS installs the release Homebrew formula on Apple Silicon + and starts a brew services-backed local gateway. EOF } @@ -76,6 +77,14 @@ require_cmd() { fi } +has_snapd() { + has_cmd snap && systemctl is-active --quiet snapd.socket 2>/dev/null +} + +has_native_docker() { + has_cmd docker && ! snap list docker >/dev/null 2>&1 +} + download() { _url="$1" _output="$2" @@ -468,12 +477,14 @@ detect_platform() { } linux_package_method() { - if has_cmd dpkg; then + if has_snapd && ! has_native_docker; then + echo "snap" + elif has_cmd dpkg; then echo "deb" elif has_cmd rpm; then echo "rpm" else - error "Linux installs require either dpkg or rpm" + error "Linux installs require either snapd, dpkg, or rpm" fi } @@ -953,6 +964,75 @@ install_linux_rpm() { start_user_gateway } +install_linux_snap() { + require_cmd snap + set_linux_target_runtime_dir + + # Docker snap must be installed before openshell so that the + # openshell:docker -> docker:docker-daemon interface connection + # can succeed. The docker-daemon slot doesn't exist until the + # Docker snap itself is present. + if ! snap list docker >/dev/null 2>&1; then + info "docker snap not found, installing..." + as_root snap install docker + else + info "docker snap already present" + fi + + info "installing openshell snap..." + as_root snap install openshell + + info "installed openshell snap from Snap Store" + info "registering local gateway as ${TARGET_USER}..." + register_local_gateway_snap + wait_for_local_gateway_listener_snap + wait_for_local_gateway_status +} + +register_local_gateway_snap() { + _register_bin="${OPENSHELL_REGISTER_BIN:-openshell}" + + if _add_output="$($_register_bin gateway add "http://127.0.0.1:${LOCAL_GATEWAY_PORT}" --local --name openshell 2>&1)"; then + [ -z "$_add_output" ] || print_gateway_add_output "$_add_output" + return 0 + fi + + _add_status=$? + + case "$_add_output" in + *"already exists"*) + info "local gateway already exists; removing and re-adding it..." + remove_local_gateway_registration + "$_register_bin" gateway add "http://127.0.0.1:${LOCAL_GATEWAY_PORT}" --local --name openshell + ;; + *) + printf '%s\n' "$_add_output" >&2 + return "$_add_status" + ;; + esac +} + +wait_for_local_gateway_listener_snap() { + _timeout="${OPENSHELL_INSTALL_GATEWAY_TIMEOUT_SNAP:-90}" + _elapsed=0 + _last_output="" + _probe_url="http://127.0.0.1:${LOCAL_GATEWAY_PORT}/" + + info "waiting for local gateway listener to become reachable..." + while [ "$_elapsed" -lt "$_timeout" ]; do + if _last_output="$(curl -sS --max-time 2 "$_probe_url" 2>&1)"; then + info "local gateway listener is reachable" + return 0 + fi + sleep 1 + _elapsed=$((_elapsed + 1)) + done + + [ -z "$_last_output" ] || printf '%s +' "$_last_output" >&2 + error "local gateway listener did not become reachable at ${_probe_url} within ${_timeout}s" +} + install_macos_homebrew() { check_macos_platform @@ -1033,12 +1113,16 @@ main() { case "$PLATFORM" in linux) - require_linux_package_glibc case "$(linux_package_method)" in + snap) + install_linux_snap + ;; deb) + require_linux_package_glibc install_linux_deb ;; rpm) + require_linux_package_glibc install_linux_rpm ;; *) diff --git a/snapcraft.yaml b/snapcraft.yaml index d221c51e7e..1802a8c03a 100644 --- a/snapcraft.yaml +++ b/snapcraft.yaml @@ -20,13 +20,10 @@ description: | sudo snap install docker - Support for system-installed Docker is coming in snapd 2.76. + Support for system-installed Docker is coming in snapd 2.77. - 2. Connect the required interfaces and start the gateway: + 2. Start the gateway: - sudo snap connect openshell:docker docker:docker-daemon - sudo snap connect openshell:log-observe - sudo snap connect openshell:system-observe sudo snap start openshell.gateway 3. Verify the gateway and register it locally: diff --git a/tasks/scripts/test-install-sh.sh b/tasks/scripts/test-install-sh.sh index a1259cf0bc..5ab28e731a 100755 --- a/tasks/scripts/test-install-sh.sh +++ b/tasks/scripts/test-install-sh.sh @@ -100,4 +100,72 @@ assert_glibc_preflight_fails \ "OpenShell Linux packages require glibc >= 2.28; detected musl or unsupported libc." \ setup_ldd_musl +# Package format is detected based on the host environment. Shim the has_* +# helpers so the auto-detect path is deterministic regardless of the host +# running the tests. +assert_package_format_detection() { + local name=$1 + local snapd=$2 + local native_docker=$3 + local dpkg=$4 + local rpm=$5 + local expected=$6 + + local result + if ! result="$( + has_snapd() { [ "$snapd" = "1" ]; } + has_native_docker() { [ "$native_docker" = "1" ]; } + has_cmd() { + case "$1" in + dpkg) [ "$dpkg" = "1" ] ;; + rpm) [ "$rpm" = "1" ] ;; + *) return 1 ;; + esac + } + linux_package_method + )" 2>"$err"; then + echo "FAIL: ${name}: call failed" >&2 + cat "$err" >&2 || true + exit 1 + fi + + if [ "$result" != "$expected" ]; then + echo "FAIL: ${name}: expected '${expected}', got '${result}'" >&2 + exit 1 + fi +} + +assert_package_format_detection \ + "prefers snap when snapd and no native docker" \ + 1 0 1 1 "snap" + +assert_package_format_detection \ + "skips snap when native docker present" \ + 1 1 1 0 "deb" + +assert_package_format_detection \ + "skips snap when snapd absent" \ + 0 0 1 0 "deb" + +assert_package_format_detection \ + "falls back to rpm when no dpkg" \ + 0 0 0 1 "rpm" + +# Host with no snapd, no dpkg, no rpm must error. +if ( + has_snapd() { return 1; } + has_native_docker() { return 1; } + has_cmd() { return 1; } + linux_package_method +) >"$out" 2>"$err"; then + echo "FAIL: host with no package managers should error" >&2 + exit 1 +fi + +if ! grep -Fq "Linux installs require either snapd, dpkg, or rpm" "$err"; then + echo "FAIL: missing no-package-manager error" >&2 + cat "$err" >&2 || true + exit 1 +fi + echo "install.sh libc preflight tests passed"