Merge branch 'fix/tunnel-execstartpre' into 'main'

fix(install.sh): clear ExecStartPre in the run-host drop-in; config/grant before aborting; 0.1.4

See merge request tikali/applications/monky/monky-deployd!6

.gitlab-ci.yml

@@ -10,8 +10,9 @@
 #   GITEA_TOKEN   — Gitea API token (write:repository) for `release:gitea`; without it the job is manual
 #
 # Release assets are published to BOTH the GitLab generic package registry (the installer's primary
-# source — the project is public; gitea.cbs.tikali.net is split-horizon inside the estate, cbs/iac#102)
-# and the Gitea mirror release (`install.sh --source gitea`, off-estate).
+# source — the project is PRIVATE (parent groups are private), so install.sh sends a read-only deploy
+# token (read_package_registry) as DEPLOY-TOKEN; gitea.cbs.tikali.net is split-horizon inside the
+# estate, cbs/iac#102) and the Gitea mirror release (`install.sh --source gitea`, off-estate).
 #   GITLAB_TRANSLATE_TOKEN — used by the translate-docs component (group level)
 
 include:

CHANGELOG.md

@@ -1,6 +1,40 @@
 <!-- xlate:verbatim-fences -->
 # Changelog
 
+## 0.1.4 — 2026-09-07
+
+- **install.sh: `ziti-edge-tunnel.service` failed to start after enrolment** (`203/EXEC`: the package's
+  `ExecStartPre` script is not executable by user `ziti` on ziti-edge-tunnel 1.18.x) and, because the
+  unit was enabled with `--now` under `set -e`, the script aborted BEFORE writing config/ACLs/grant/timer —
+  a second kit was needed. The run-host drop-in now clears `ExecStartPre=`; a tunneler start failure is a
+  WARNING until everything else is in place, then a hard error at the end.
+
+## 0.1.3 — 2026-09-06
+
+- **install.sh: `--version` was clobbered by `/etc/os-release`** (it defines `VERSION`), so the kit tried to download
+  `monky-deployd_26.04 LTS (Resolute Raccoon)_amd64.deb` (env-qa-02 pilot, first real kit run). Script variables are now
+  `DEPLOYD_VERSION`; os-release is read in a subshell.
+- **install.sh installs Docker Engine when absent** (Docker's apt suite for the codename → fallback `docker.io`), with
+  `--docker-data-root <dir>` to place the data-root before first start. One script for the end user, per the operator's rule.
+
+## v0.1.2 — 2026-09-05
+
+- **The GitLab project is private** (its parent groups are private, so it cannot be made public;
+  found when the v0.1.1 one-liner returned 401 anonymously). `install.sh` gains `--token <deploy-token>`
+  / `MONKY_DEPLOYD_TOKEN`: every download from the generic package registry — the script itself
+  included — sends `DEPLOY-TOKEN: <token>`, a read-only GitLab **deploy token** (scope
+  `read_package_registry` only, revocable), seeded in OpenBao at `monky/monky-tenancy/deployd-download`
+  (key `token`) and handed to the box by the monky-tenancy install kit. The token never reaches the
+  command line, the log or an xtrace (curl `-K` config file, 0600, deleted after the download;
+  `set +x` forced). The one-liner now fetches `install.sh` from the registry
+  (`…/monky-deployd/<ver>/install.sh`) instead of `-/raw/main`, and takes the bootstrap grant via
+  `--bootstrap-file` — with `curl … | bash -s --` stdin IS the script, so it is no longer read for
+  the grant in that mode. `--source gitea` (no token) stays the off-estate alternative.
+- Ansible role: `monky_deployd_download_token` (vaulted) → `DEPLOY-TOKEN` header on both fetches,
+  `no_log: true`.
+- README / OPERATIONS: why (private project; split-horizon Gitea, cbs/iac#102), token scope, revocation.
+  No agent behaviour change.
+
 ## v0.1.1 — 2026-09-05
 
 - `install.sh` / ansible role / README: the **primary download is the public GitLab project's generic

CLAUDE.md

@@ -75,8 +75,9 @@ systemd-analyze verify packaging/systemd/*.service   # when available
 
 Bump `monky_deployd.__version__` + `pyproject.toml` + `packaging/install.sh` `DEFAULT_VERSION` +
 `CHANGELOG.md`, merge to `main`, tag `vX.Y.Z` (protected `v*`). The tag pipeline builds the `.deb`,
-publishes the **GitLab generic package registry + release** the installer downloads from (public
-project; the Gitea name is split-horizon inside the estate, cbs/iac#102) and — with `GITEA_TOKEN` —
+publishes the **GitLab generic package registry + release** the installer downloads from (PRIVATE
+project — the installer sends a read-only deploy token as `DEPLOY-TOKEN`; the Gitea name is
+split-horizon inside the estate, cbs/iac#102) and — with `GITEA_TOKEN` —
 the **Gitea release** (`install.sh --source gitea`, off-estate; `docs/OPERATIONS.md` has the manual recipe). `lint` refuses a tag whose version
 differs from `__version__`.
 

README.md

@@ -18,12 +18,14 @@ Verified on **Ubuntu 26.04**.
 ## Install (one-liner, from the enrolment kit)
 
 An admin reveals the kit once in the console (`GET /v1/backends/{id}/agent/install`); it hands
-you the enrolment JWT and a one-time **bootstrap deploy grant**. On the box:
+you the enrolment JWT, a one-time **bootstrap deploy grant** and the read-only **download token**
+(the kit runs the line below for you). By hand, on the box:
 
 ```sh
-curl -fsSL https://scm.tikali.ai/tikali/applications/monky/monky-deployd/-/raw/main/packaging/install.sh \
-  | sudo bash -s -- --env env-qa-02 --site cbs --enrol-jwt ./monky-host.env-qa-02.jwt < bootstrap.jwt
-#   [--transport sdk|proxy|system] [--version 0.1.1] [--laptop] [--bao-ca openbao-ca.pem] [--source gitlab|gitea]
+T=<deploy token>   # read-only GitLab deploy token (read_package_registry); the kit carries it
+curl -sSf -H "DEPLOY-TOKEN: $T" https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd/0.1.4/install.sh \
+  | sudo bash -s -- --env env-qa-02 --site cbs --token "$T" --bootstrap-file bootstrap.jwt --enrol-jwt ./monky-host.env-qa-02.jwt
+#   [--transport sdk|proxy|system] [--version 0.1.4] [--docker-data-root /home/docker-data] [--laptop] [--bao-ca openbao-ca.pem] [--source gitlab|gitea]
 ```
 
 `install.sh` installs `ziti-edge-tunnel` (OpenZiti `jammy` suite) and `docker-compose-plugin` if
@@ -32,11 +34,21 @@ absent, downloads the pinned `.deb` + `.sha256` from the [scm.tikali.ai generic
 enrols `monky-host.<env>` if the identity is missing, switches the tunneler to `run-host`,
 writes `/etc/monky-deployd/config.yaml`, grants the agent read access to the identity (ACL),
 stages the bootstrap grant (0600), enables `monky-deployd.timer`, runs one tick and deletes the
-JWT. The GitLab project is **public**, so anonymous downloads work from scm.tikali.ai. Why not the
-Gitea mirror: inside the estate `gitea.cbs.tikali.net` is split-horizon to jump1's RED EIP
-(`10.10.0.175`), which has no HTTP ingress, so backend boxes cannot reach it (cbs/iac#102);
-`--source gitea` (or `MONKY_DEPLOYD_SOURCE=gitea`) keeps the
-[Gitea release](https://gitea.cbs.tikali.net/mdella/monky-deployd/releases) as the off-estate alternative.
+JWT.
+
+**Why a token.** The GitLab project is **private** and cannot be made public (its parent groups
+are private), so anonymous downloads answer 401. Every registry fetch — the script itself
+included — therefore sends `DEPLOY-TOKEN: <token>`: a GitLab **deploy token** with the single
+scope `read_package_registry` (it can download packages and nothing else: no code, no API, no
+write; revoke it in the project's *Settings → Repository → Deploy tokens* and issue a new one).
+The operator seeds it in OpenBao at `monky/monky-tenancy/deployd-download` (key `token`); the
+monky-tenancy install kit reads it from there and passes `--token` (`MONKY_DEPLOYD_TOKEN` also
+works). The script never prints it (curl `-K` config file, 0600, deleted after the download;
+`set -x` is switched off). Why not the Gitea mirror: inside the estate `gitea.cbs.tikali.net` is
+split-horizon to jump1's RED EIP (`10.10.0.175`), which has no HTTP ingress, so backend boxes
+cannot reach it (cbs/iac#102); `--source gitea` (or `MONKY_DEPLOYD_SOURCE=gitea`, no token) keeps
+the [Gitea release](https://gitea.cbs.tikali.net/mdella/monky-deployd/releases) as the
+off-estate alternative.
 
 ## Transports
 
@@ -145,8 +157,8 @@ systemd-analyze verify packaging/systemd/*.service   # where systemd is availabl
 egress to github.com and pypi.org) and `package` builds the `.deb` with `nfpm` (binary from GitHub
 releases, goreleaser apt repo as fallback). Both were proven on the v0.1.0 tag pipeline and are
 blocking on `main`/tags (manual on MRs); the `.deb` always ships the SDK wheel. On a `v*` tag
-`release` uploads to the GitLab generic package registry + release (**the primary download**, public
-project), and `release:gitea` publishes the same assets on the Gitea mirror (the `--source gitea`
+`release` uploads to the GitLab generic package registry + release (**the primary download**; the
+project is private, so the installer sends the read-only deploy token), and `release:gitea` publishes the same assets on the Gitea mirror (the `--source gitea`
 alternative; automatic when `GITEA_TOKEN` is set, manual otherwise — see `docs/OPERATIONS.md` for
 the by-hand recipe). Both locations keep being published.
 
@@ -155,3 +167,10 @@ the by-hand recipe). Both locations keep being published.
 - `docs/PROTOCOL.md`, `docs/OPERATIONS.md`, `CHANGELOG.md`
 - monky-tenancy `docs/usage.md` (agent protocol), `app/api/agent.py`, `app/schemas_backends.py`
 - monky-deploy (the renderer whose `render_files` produces the bundle)
+
+### Docker Engine (since 0.1.3)
+
+`install.sh` installs Docker Engine when it is absent (Docker's apt suite for the host's Ubuntu codename, falling back to
+Ubuntu's `docker.io`), so an end user runs exactly one script. `--docker-data-root <dir>` writes `/etc/docker/daemon.json`
+before the daemon first starts (pilot VMs keep the data-root on a dedicated disk, e.g. `/home/docker-data`). A pre-existing
+Docker is left untouched. macOS/Windows laptops still need Docker Desktop from the user.

ansible/roles/monky_deployd/README.md

@@ -4,7 +4,9 @@ Installs and configures [monky-deployd](https://scm.tikali.ai/tikali/application
 (the Monky backend pull agent, MONKY-ADR-0028) on a docker host that already carries an enrolled
 host identity (`roles/ziti_tunneler`, run-host mode). Skeleton for **osg1-07**; copy it there.
 
-What it does: pin + download the `.deb` from the scm.tikali.ai package registry (sha256 verified; `monky_deployd_base_url`/`_deb_url` switch to the Gitea release off-estate) → ACL
+What it does: pin + download the `.deb` from the scm.tikali.ai package registry (sha256 verified; the
+project is **private**, so the fetches send the read-only deploy token `monky_deployd_download_token`
+as `DEPLOY-TOKEN`, `no_log`; `monky_deployd_base_url`/`_deb_url` switch to the Gitea release off-estate) → ACL
 `u:monky-deployd:r` on the identity (`rx` on the dir) → `/etc/monky-deployd/config.yaml` from the
 template → optional openbao-ca PEM → optional `monky-deployd-proxy.service` (`transport: proxy`)
 → the **one-time bootstrap deploy grant** from a vault var (0600, `no_log`) → `monky-deployd.timer`
@@ -14,7 +16,8 @@ template → optional openbao-ca PEM → optional `monky-deployd-proxy.service`
 
 | var | note |
 |---|---|
-| `monky_deployd_version` | pinned release, e.g. `0.1.1` |
+| `monky_deployd_version` | pinned release, e.g. `0.1.4` |
+| `monky_deployd_download_token` | **vaulted**: GitLab deploy token, scope `read_package_registry` only (revocable) — the registry is private; seeded in OpenBao at `monky/monky-tenancy/deployd-download` key `token` (path/key are the operator's choice). Empty = no header (only works with the Gitea `base_url`) |
 | `monky_deployd_env_id` / `_site` / `_transport` | per host (`env-dev-06`, `cbs`, `sdk`) |
 | `monky_deployd_bootstrap_grant` | tenancy-minted deploy grant (1 h) — `ansible-vault` or a lookup at play time; empty keeps the existing token |
 | `monky_deployd_bao_ca_pem` | the `openbao-ca` certificate (PEM) |
@@ -30,6 +33,7 @@ template → optional openbao-ca PEM → optional `monky-deployd-proxy.service`
     - role: monky_deployd
       vars:
         monky_deployd_bootstrap_grant: "{{ lookup('pipe', 'tenancy-mint-grant ' ~ inventory_hostname) }}"
+        monky_deployd_download_token: "{{ vault_monky_deployd_download_token }}"   # ansible-vault / OpenBao lookup
 ```
 
 Rollout order (plan §E): pilot env-qa-02 → env-dev-06..09 → env-dev-01 last (after its live key moves into Bao).

ansible/roles/monky_deployd/defaults/main.yml

@@ -1,16 +1,24 @@
 ---
 # monky_deployd — install and configure the Monky backend pull agent (MONKY-ADR-0028 §D).
 # Copy this role into osg1-07 (roles/monky_deployd) and roll to env-dev-06..09 after the pilot.
-monky_deployd_version: "0.1.1"
+monky_deployd_version: "0.1.4"
 monky_deployd_deb: "monky-deployd_{{ monky_deployd_version }}_amd64.deb"
-# PRIMARY download = the public GitLab project's generic package registry on scm.tikali.ai. Inside the
-# estate gitea.cbs.tikali.net is split-horizon to jump1's RED EIP (10.10.0.175, no HTTP ingress), so
-# backend boxes cannot fetch from the Gitea mirror (cbs/iac#102). Off-estate alternative (Gitea release):
+# PRIMARY download = the GitLab project's generic package registry on scm.tikali.ai. Inside the estate
+# gitea.cbs.tikali.net is split-horizon to jump1's RED EIP (10.10.0.175, no HTTP ingress), so backend
+# boxes cannot fetch from the Gitea mirror (cbs/iac#102). The project is PRIVATE (its parent groups are
+# private), so the registry needs a read-only GitLab deploy token (scope read_package_registry only,
+# revocable) sent as the `DEPLOY-TOKEN` header — `monky_deployd_download_token`, a VAULTED var
+# (ansible-vault or an OpenBao lookup: KV `monky/monky-tenancy/deployd-download`, key `token`).
+# Off-estate alternative (Gitea release, no token):
 #   monky_deployd_base_url: "https://gitea.cbs.tikali.net/mdella/monky-deployd"
 #   monky_deployd_deb_url: "{{ monky_deployd_base_url }}/releases/download/v{{ monky_deployd_version }}/{{ monky_deployd_deb }}"
+#   monky_deployd_download_token: ""
 monky_deployd_base_url: "https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd"
 monky_deployd_deb_url: "{{ monky_deployd_base_url }}/{{ monky_deployd_version }}/{{ monky_deployd_deb }}"
 monky_deployd_deb_sha256_url: "{{ monky_deployd_deb_url }}.sha256"
+monky_deployd_download_token: ""
+# the header the download tasks send (empty map when no token)
+monky_deployd_download_headers: "{{ {'DEPLOY-TOKEN': monky_deployd_download_token} if monky_deployd_download_token | length > 0 else {} }}"
 
 # per host (inventory / host_vars)
 monky_deployd_env_id: "{{ inventory_hostname }}"        # env-dev-06 ...

ansible/roles/monky_deployd/tasks/main.yml

@@ -22,20 +22,26 @@
   changed_when: false
   failed_when: false
 
+# The registry is PRIVATE: both fetches send `DEPLOY-TOKEN: {{ monky_deployd_download_token }}`
+# (read_package_registry only) and run no_log so the header never reaches the play output.
 - name: monky_deployd | download .deb + sha256 from the scm.tikali.ai package registry
   when: monky_deployd_installed.stdout != monky_deployd_version
   block:
     - name: monky_deployd | fetch sha256
       ansible.builtin.uri:
         url: "{{ monky_deployd_deb_sha256_url }}"
+        headers: "{{ monky_deployd_download_headers }}"
         return_content: true
       register: monky_deployd_sha
+      no_log: true
     - name: monky_deployd | fetch .deb (checksum verified)
       ansible.builtin.get_url:
         url: "{{ monky_deployd_deb_url }}"
+        headers: "{{ monky_deployd_download_headers }}"
         dest: "/var/cache/apt/archives/{{ monky_deployd_deb }}"
         checksum: "sha256:{{ monky_deployd_sha.content.split()[0] }}"
         mode: "0644"
+      no_log: true
     - name: monky_deployd | install .deb
       ansible.builtin.apt:
         deb: "/var/cache/apt/archives/{{ monky_deployd_deb }}"

config.example.yaml

@@ -1,4 +1,4 @@
-# /etc/monky-deployd/config.yaml — monky-deployd v0.1.1 (MONKY-ADR-0028 §D)
+# /etc/monky-deployd/config.yaml — monky-deployd v0.1.4 (MONKY-ADR-0028 §D)
 # Written by packaging/install.sh (or the ansible role monky_deployd). YAML *subset*: maps, scalars,
 # simple lists, comments. Keys not listed here are a config error.
 

docs/OPERATIONS.md

@@ -88,14 +88,27 @@ the picker, nothing is retired automatically.
 
 ## Where the installer downloads from
 
-The primary source is the **scm.tikali.ai generic package registry** of this (public) project:
+The primary source is the **scm.tikali.ai generic package registry** of this project:
 `https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd/<ver>/<file>` for
-`monky-deployd_<ver>_amd64.deb`, `.sha256` and `install.sh`; the one-liner fetches the script from
-`https://scm.tikali.ai/tikali/applications/monky/monky-deployd/-/raw/main/packaging/install.sh`.
-Inside the estate `gitea.cbs.tikali.net` is split-horizon to jump1's RED EIP (`10.10.0.175`), which
-has no HTTP ingress, so backend boxes cannot fetch from the Gitea mirror (cbs/iac#102). Off-estate,
-`install.sh --source gitea` (ansible: `monky_deployd_base_url`/`_deb_url`) uses the Gitea release
-instead. The tag pipeline publishes to both (`release`, `release:gitea`).
+`monky-deployd_<ver>_amd64.deb`, `.sha256` and `install.sh`. The project is **private** (its parent
+groups are private, so it cannot be made public): every fetch, the script itself included, sends the
+`DEPLOY-TOKEN` header with a read-only GitLab **deploy token** — scope `read_package_registry` only,
+nothing else (no repository, no API, no write); revocable at any time in the project's *Settings →
+Repository → Deploy tokens*. It lives in OpenBao at `monky/monky-tenancy/deployd-download` (key
+`token`); the monky-tenancy install kit carries it and passes `--token`, the ansible role sends it
+from the vaulted `monky_deployd_download_token`. The one-liner:
+
+```sh
+curl -sSf -H "DEPLOY-TOKEN: $T" https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd/<ver>/install.sh \
+  | sudo bash -s -- --env <id> --site <site> --token "$T" --bootstrap-file bootstrap.jwt
+```
+
+`install.sh` never prints the token (it goes through a 0600 curl `-K` file that is deleted after the
+download; xtrace is switched off). A 401 on the download means the token is missing, revoked or
+lacks the scope. Inside the estate `gitea.cbs.tikali.net` is split-horizon to jump1's RED EIP
+(`10.10.0.175`), which has no HTTP ingress, so backend boxes cannot fetch from the Gitea mirror
+(cbs/iac#102). Off-estate, `install.sh --source gitea` (ansible: `monky_deployd_base_url`/`_deb_url`,
+no token) uses the Gitea release instead. The tag pipeline publishes to both (`release`, `release:gitea`).
 
 ## Publishing a release to Gitea by hand
 

monky_deployd/__init__.py

@@ -4,4 +4,4 @@ Dials monky-tenancy over the mesh with the box's host identity, fetches the rend
 leases a deploy grant, logs in to OpenBao, reads its own secrets, runs `docker compose`,
 reports. Stdlib only; the optional `openziti` SDK is the `sdk` transport."""
 
-__version__ = "0.1.1"
+__version__ = "0.1.4"

packaging/install.sh

@@ -1,18 +1,26 @@
 #!/usr/bin/env bash
 # monky-deployd installer — Ubuntu 26.04 (verified target).
 #
-#   curl -fsSL https://scm.tikali.ai/tikali/applications/monky/monky-deployd/-/raw/main/packaging/install.sh \
-#     | sudo bash -s -- --env env-qa-02 --site cbs [--transport sdk|proxy|system] [--version 0.1.1] \
-#     [--enrol-jwt /path/monky-host.env-qa-02.jwt] [--laptop] [--source gitlab|gitea] < bootstrap.jwt
+#   T=<read-only GitLab deploy token, scope read_package_registry>   # from the install kit / OpenBao
+#   curl -sSf -H "DEPLOY-TOKEN: $T" \
+#     https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd/0.1.4/install.sh \
+#     | sudo bash -s -- --env env-qa-02 --site cbs --token "$T" --bootstrap-file bootstrap.jwt \
+#     [--transport sdk|proxy|system] [--version 0.1.4] [--enrol-jwt /path/monky-host.env-qa-02.jwt] \
+#     [--laptop] [--source gitlab|gitea] [--docker-data-root /home/docker-data]
 #
-# stdin (or --bootstrap-file): the ONE-TIME bootstrap deploy grant from the install kit
-# (GET /v1/backends/{id}/agent/install). The enrolment JWT is read from --enrol-jwt or
-# /etc/monky-deployd/enrol.jwt and is only needed when the host identity is not enrolled yet.
+# --token / MONKY_DEPLOYD_TOKEN: the GitLab project is PRIVATE (its parent groups are private, so it
+# cannot be made public); every download from the generic package registry — this script included —
+# sends `DEPLOY-TOKEN: <token>`. The token is read-only (read_package_registry), revocable, and is
+# never printed by this script. Not needed with --source gitea.
+# --bootstrap-file (or stdin when the script runs from a FILE): the ONE-TIME bootstrap deploy grant
+# from the install kit (GET /v1/backends/{id}/agent/install). The enrolment JWT is read from
+# --enrol-jwt or /etc/monky-deployd/enrol.jwt and is only needed when the host identity is not
+# enrolled yet.
 #
 # What it does (idempotent):
 #   1. apt: ziti-edge-tunnel (OpenZiti `jammy` suite) if absent, docker-compose-plugin, acl
 #   2. downloads the pinned monky-deployd_<ver>_amd64.deb + .sha256 from the scm.tikali.ai package registry
-#      (--source gitea: the Gitea release, off-estate), verifies, installs
+#      (DEPLOY-TOKEN header; --source gitea: the Gitea release, off-estate), verifies, installs
 #   3. enrols /opt/openziti/etc/identities/monky-host.<env>.json if absent (ziti-edge-tunnel enroll),
 #      chown ziti:ziti 0600, switches ziti-edge-tunnel.service to `run-host` via a drop-in
 #   4. writes /etc/monky-deployd/config.yaml, ACLs so user monky-deployd can read the identity,
@@ -21,22 +29,25 @@
 set -euo pipefail
 umask 077
 
-DEFAULT_VERSION="0.1.1"
-# Download source. PRIMARY is the public GitLab project's generic package registry on scm.tikali.ai:
-# inside the estate gitea.cbs.tikali.net is split-horizon to jump1's RED EIP (10.10.0.175), which has
-# no HTTP ingress, so backend boxes cannot reach the Gitea mirror (cbs/iac#102); scm.tikali.ai they can.
-# `--source gitea` keeps the Gitea release as the off-estate alternative. --base-url / MONKY_DEPLOYD_BASE_URL
-# override the base for whichever layout is selected.
+DEFAULT_VERSION="0.1.4"
+# Download source. PRIMARY is the GitLab project's generic package registry on scm.tikali.ai: inside
+# the estate gitea.cbs.tikali.net is split-horizon to jump1's RED EIP (10.10.0.175), which has no HTTP
+# ingress, so backend boxes cannot reach the Gitea mirror (cbs/iac#102); scm.tikali.ai they can. The
+# project is PRIVATE (parent groups are private), so the registry needs the read-only deploy token
+# (--token / MONKY_DEPLOYD_TOKEN, sent as the DEPLOY-TOKEN header). `--source gitea` keeps the Gitea
+# release as the off-estate alternative (no token). --base-url / MONKY_DEPLOYD_BASE_URL override the
+# base for whichever layout is selected.
 GITLAB_BASE_URL="https://scm.tikali.ai/api/v4/projects/69/packages/generic/monky-deployd"
 GITEA_BASE_URL="https://gitea.cbs.tikali.net/mdella/monky-deployd"
 SOURCE="${MONKY_DEPLOYD_SOURCE:-gitlab}"
 BASE_URL="${MONKY_DEPLOYD_BASE_URL:-}"
+TOKEN="${MONKY_DEPLOYD_TOKEN:-}"
 OPENZITI_SUITE="${OPENZITI_SUITE:-jammy}"
 IDENTITY_DIR="/opt/openziti/etc/identities"
 ETC="/etc/monky-deployd"
-ENV_ID="" SITE="" TRANSPORT="sdk" VERSION="$DEFAULT_VERSION" ENROL_JWT="" BOOTSTRAP_FILE="" NO_RUN="" FORCE_CONFIG="" BAO_CA="" LAPTOP="false"
+ENV_ID="" SITE="" TRANSPORT="sdk" DOCKER_DATA_ROOT="" TUNNEL_FAILED="" DEPLOYD_VERSION="$DEFAULT_VERSION" ENROL_JWT="" BOOTSTRAP_FILE="" NO_RUN="" FORCE_CONFIG="" BAO_CA="" LAPTOP="false"
 
-usage() { sed -n '2,20p' "$0"; exit "${1:-0}"; }
+usage() { if [ -f "$0" ]; then sed -n '2,28p' "$0"; else echo "monky-deployd install.sh — see README.md (Install)"; fi; exit "${1:-0}"; }
 die() { echo "install.sh: $*" >&2; exit 1; }
 log() { echo "==> $*"; }
 
@@ -45,12 +56,14 @@ while [ $# -gt 0 ]; do
     --env) ENV_ID="$2"; shift 2 ;;
     --site) SITE="$2"; shift 2 ;;
     --transport) TRANSPORT="$2"; shift 2 ;;
-    --version) VERSION="$2"; shift 2 ;;
+    --version) DEPLOYD_VERSION="$2"; shift 2 ;;
+    --docker-data-root) DOCKER_DATA_ROOT="$2"; shift 2 ;;
     --enrol-jwt) ENROL_JWT="$2"; shift 2 ;;
     --bootstrap-file) BOOTSTRAP_FILE="$2"; shift 2 ;;
     --bao-ca) BAO_CA="$2"; shift 2 ;;
     --base-url) BASE_URL="$2"; shift 2 ;;
     --source) SOURCE="$2"; shift 2 ;;
+    --token) TOKEN="$2"; shift 2 ;;
     --laptop) LAPTOP="true"; shift ;;
     --no-run) NO_RUN=1; shift ;;
     --force-config) FORCE_CONFIG=1; shift ;;
@@ -67,13 +80,20 @@ SITE="${SITE,,}"
 [[ "$SITE" =~ ^(cbs|pdx)$ ]] || die "site must be cbs|pdx"
 [[ "$TRANSPORT" =~ ^(sdk|proxy|system)$ ]] || die "transport must be sdk|proxy|system"
 [[ "$SOURCE" =~ ^(gitlab|gitea)$ ]] || die "source must be gitlab|gitea"
+# never let the token leak through xtrace / the environment of children
+{ set +x; } 2>/dev/null
+export -n MONKY_DEPLOYD_TOKEN 2>/dev/null || true
 IDENTITY="$IDENTITY_DIR/monky-host.$ENV_ID.json"
 export DEBIAN_FRONTEND=noninteractive
 
 if [ -r /etc/os-release ]; then
-  . /etc/os-release
-  if [ "${ID:-}" != "ubuntu" ] || [ "${VERSION_ID:-}" != "26.04" ]; then
-    echo "WARNING: verified on Ubuntu 26.04; this is ${PRETTY_NAME:-unknown}. Continuing." >&2
+  # read os-release in a SUBSHELL: it defines VERSION/NAME/ID/... and sourcing it inline clobbered
+  # this script's own variables (0.1.2 → "26.04 LTS (Resolute Raccoon)" on the env-qa-02 pilot)
+  OS_ID=$(. /etc/os-release; printf '%s' "${ID:-}")
+  OS_VERSION_ID=$(. /etc/os-release; printf '%s' "${VERSION_ID:-}")
+  OS_CODENAME=$(. /etc/os-release; printf '%s' "${VERSION_CODENAME:-${UBUNTU_CODENAME:-}}")
+  if [ "$OS_ID" != "ubuntu" ] || [ "$OS_VERSION_ID" != "26.04" ]; then
+    echo "WARNING: verified on Ubuntu 26.04 only (this is ${OS_ID:-?} ${OS_VERSION_ID:-?}); continuing" >&2
   fi
 fi
 
@@ -93,7 +113,34 @@ fi
 if [ "$TRANSPORT" = "proxy" ] && ! command -v ziti >/dev/null 2>&1; then
   need_pkgs+=(openziti)   # the `ziti` CLI (ziti tunnel proxy) from the same repo
 fi
-command -v docker >/dev/null 2>&1 || die "docker is not installed; install Docker Engine first (https://docs.docker.com/engine/install/ubuntu/)"
+if ! command -v docker >/dev/null 2>&1; then
+  # Docker Engine from Docker's apt repo (the suite for this Ubuntu codename); falls back to Ubuntu's
+  # docker.io if Docker has no suite for the codename yet. --docker-data-root puts the data-root on
+  # a dedicated disk BEFORE the daemon first starts (pilot VMs mount one at /home/docker-data).
+  log "docker is not installed — installing Docker Engine"
+  export DEBIAN_FRONTEND=noninteractive
+  install -d -m 0755 /etc/apt/keyrings
+  if curl -fsSL --max-time 20 "https://download.docker.com/linux/ubuntu/dists/${OS_CODENAME:-noble}/Release" -o /dev/null 2>/dev/null; then
+    curl -fsSL https://download.docker.com/linux/ubuntu/gpg | gpg --dearmor -o /etc/apt/keyrings/docker.gpg
+    chmod a+r /etc/apt/keyrings/docker.gpg
+    echo "deb [arch=$(dpkg --print-architecture) signed-by=/etc/apt/keyrings/docker.gpg] https://download.docker.com/linux/ubuntu ${OS_CODENAME:-noble} stable" > /etc/apt/sources.list.d/docker.list
+    apt-get update -qq
+    docker_pkgs="docker-ce docker-ce-cli containerd.io docker-buildx-plugin docker-compose-plugin"
+  else
+    log "Docker has no apt suite for '${OS_CODENAME:-?}' yet — using Ubuntu's docker.io"
+    apt-get update -qq
+    docker_pkgs="docker.io docker-compose-v2"
+  fi
+  if [ -n "$DOCKER_DATA_ROOT" ]; then
+    install -d -m 0710 "$DOCKER_DATA_ROOT"
+    install -d -m 0755 /etc/docker
+    [ -f /etc/docker/daemon.json ] || printf '{ "data-root": "%s" }\n' "$DOCKER_DATA_ROOT" > /etc/docker/daemon.json
+  fi
+  # shellcheck disable=SC2086
+  apt-get install -y -qq --no-install-recommends $docker_pkgs
+  systemctl enable --now docker
+  log "installed $(docker --version) (data-root $(docker info -f '{{.DockerRootDir}}' 2>/dev/null || echo default))"
+fi
 docker compose version >/dev/null 2>&1 || need_pkgs+=(docker-compose-plugin)
 python3 -c 'import sys; sys.exit(0 if sys.version_info >= (3, 12) else 1)' 2>/dev/null || die "python3 >= 3.12 is required"
 if [ "${#need_pkgs[@]}" -gt 0 ]; then
@@ -104,19 +151,28 @@ fi
 
 # --- 2. the pinned .deb -----------------------------------------------------------------------------
 installed="$(dpkg-query -W -f='${Version}' monky-deployd 2>/dev/null || true)"
-if [ "$installed" = "$VERSION" ]; then
-  log "monky-deployd $VERSION already installed"
+if [ "$installed" = "$DEPLOYD_VERSION" ]; then
+  log "monky-deployd $DEPLOYD_VERSION already installed"
 else
   tmp="$(mktemp -d)"; trap 'rm -rf "$tmp"' EXIT
-  deb="monky-deployd_${VERSION}_amd64.deb"
+  deb="monky-deployd_${DEPLOYD_VERSION}_amd64.deb"
+  # the deploy token travels in a 0600 curl config file (-K), never on the command line, so neither
+  # `ps` nor an xtrace shows it; the file dies with $tmp
+  curlrc="$tmp/curlrc"; : > "$curlrc"; chmod 0600 "$curlrc"
   if [ "$SOURCE" = gitlab ]; then
-    url="${BASE_URL:-$GITLAB_BASE_URL}/${VERSION}"
+    url="${BASE_URL:-$GITLAB_BASE_URL}/${DEPLOYD_VERSION}"
+    if [ -n "$TOKEN" ]; then
+      printf 'header = "DEPLOY-TOKEN: %s"\n' "$TOKEN" > "$curlrc"
+    else
+      echo "WARNING: no --token/MONKY_DEPLOYD_TOKEN: the scm.tikali.ai project is private, the download will 401 (use --source gitea off-estate)" >&2
+    fi
   else
-    url="${BASE_URL:-$GITEA_BASE_URL}/releases/download/v${VERSION}"
+    url="${BASE_URL:-$GITEA_BASE_URL}/releases/download/v${DEPLOYD_VERSION}"
   fi
-  log "downloading $deb from $url"
-  curl -fsSL -o "$tmp/$deb" "$url/$deb"
-  curl -fsSL -o "$tmp/$deb.sha256" "$url/$deb.sha256"
+  log "downloading $deb from $url${TOKEN:+ (DEPLOY-TOKEN)}"
+  curl -fsSL -K "$curlrc" -o "$tmp/$deb" "$url/$deb" || die "download of $deb failed (private registry: is the deploy token set and valid?)"
+  curl -fsSL -K "$curlrc" -o "$tmp/$deb.sha256" "$url/$deb.sha256" || die "download of $deb.sha256 failed"
+  rm -f "$curlrc"
   (cd "$tmp" && sha256sum -c "$deb.sha256") || die "sha256 mismatch on $deb"
   apt_update_once || true
   apt-get install -y -qq "$tmp/$deb"
@@ -138,14 +194,21 @@ chmod 0600 "$IDENTITY"
 # ziti-edge-tunnel as a HOST (bind side, no tun/DNS) — the agent dials with the SDK or the proxy
 install -d /etc/systemd/system/ziti-edge-tunnel.service.d
 cat > /etc/systemd/system/ziti-edge-tunnel.service.d/run-host.conf <<'DROPIN'
-# monky-deployd: run-host mode (no tun, no DNS); identities from the standard directory
+# monky-deployd: run-host mode (no tun, no DNS); identities from the standard directory.
+# ExecStartPre is cleared: the package's ziti-edge-tunnel.sh (auto-enrol *.jwt in the identity dir)
+# is not executable by user ziti on 1.18.x (203/EXEC, env-qa-02 pilot) and run-host does not need it.
 [Service]
+ExecStartPre=
 ExecStart=
 ExecStart=/opt/openziti/bin/ziti-edge-tunnel run-host --identity-dir=/opt/openziti/etc/identities
 DROPIN
 systemctl daemon-reload
-systemctl enable --now ziti-edge-tunnel.service
-systemctl restart ziti-edge-tunnel.service || true
+systemctl enable ziti-edge-tunnel.service >/dev/null 2>&1 || true
+if ! systemctl restart ziti-edge-tunnel.service; then
+  # do NOT abort here: config, ACLs, grant and timer below must land so a re-run needs no new kit
+  echo "WARNING: ziti-edge-tunnel.service failed to start — see: journalctl -u ziti-edge-tunnel; continuing" >&2
+  TUNNEL_FAILED=1
+fi
 
 # --- 4. config, ACLs, bootstrap grant -----------------------------------------------------------------
 install -d -m 0750 -o root -g monky-deployd "$ETC"
@@ -162,7 +225,7 @@ if [ -s "$ETC/config.yaml" ] && [ -z "$FORCE_CONFIG" ]; then
   log "keeping existing $ETC/config.yaml (use --force-config to rewrite)"
 else
   cat > "$ETC/config.yaml" <<CFG
-# written by install.sh $(date -u +%FT%TZ) — monky-deployd $VERSION
+# written by install.sh $(date -u +%FT%TZ) — monky-deployd $DEPLOYD_VERSION
 env_id: $ENV_ID
 site: $SITE
 transport: $TRANSPORT
@@ -194,10 +257,11 @@ if [ "$TRANSPORT" = "proxy" ]; then
   printf 'ZITI_IDENTITY=%s\n' "$IDENTITY" > "$ETC/proxy.env"; chmod 0644 "$ETC/proxy.env"
   systemctl enable --now monky-deployd-proxy.service
 fi
-# bootstrap grant: stdin (the kit pipes it) or --bootstrap-file; 0600, owned by the agent so it can consume it
+# bootstrap grant: --bootstrap-file, or stdin — but ONLY when this script runs from a file: with
+# `curl … | bash -s --` stdin IS the script, so reading it here would swallow the rest of it
 grant=""
 if [ -n "$BOOTSTRAP_FILE" ]; then grant="$(tr -d '[:space:]' < "$BOOTSTRAP_FILE")"
-elif [ ! -t 0 ]; then grant="$(tr -d '[:space:]' </dev/stdin || true)"; fi
+elif [ ! -t 0 ] && [ -f "$0" ]; then grant="$(tr -d '[:space:]' </dev/stdin || true)"; fi
 if [ -n "$grant" ]; then
   printf '%s\n' "$grant" > "$ETC/bootstrap.jwt"
   chown monky-deployd:monky-deployd "$ETC/bootstrap.jwt"; chmod 0600 "$ETC/bootstrap.jwt"
@@ -207,7 +271,7 @@ elif [ -s /var/lib/monky-deployd/bao.token ]; then
 else
   echo "WARNING: no bootstrap grant on stdin and no bao.token: the agent cannot check in until you provide one" >&2
 fi
-unset grant
+unset grant TOKEN
 
 # --- 5. timer + first tick ----------------------------------------------------------------------------------
 systemctl daemon-reload
@@ -221,7 +285,7 @@ rm -f "$ETC/enrol.jwt"
 
 cat <<CHECK
 
-monky-deployd $VERSION installed for $ENV_ID ($SITE, transport $TRANSPORT). Verification checklist (Ubuntu 26.04):
+monky-deployd $DEPLOYD_VERSION installed for $ENV_ID ($SITE, transport $TRANSPORT). Verification checklist (Ubuntu 26.04):
   python3 --version                              # >= 3.12
   docker compose version                         # compose plugin present
   systemctl is-active ziti-edge-tunnel           # run-host mode (drop-in run-host.conf)
@@ -235,3 +299,7 @@ monky-deployd $VERSION installed for $ENV_ID ($SITE, transport $TRANSPORT). Veri
 CHECK
 CHECK_STATUS="$(monky-deployd status 2>&1 || true)"
 echo "$CHECK_STATUS" | sed 's/^/  | /'
+if [ -n "$TUNNEL_FAILED" ] && ! systemctl is-active --quiet ziti-edge-tunnel.service; then
+  echo "ERROR: ziti-edge-tunnel.service is not running; the agent cannot reach the mesh until it is. Everything else is installed — fix the unit and 'systemctl restart ziti-edge-tunnel monky-deployd'." >&2
+  exit 1
+fi

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "monky-deployd"
-version = "0.1.1"
+version = "0.1.4"
 description = "Monky backend pull agent: checkin -> bundle -> lease -> OpenBao -> docker compose -> report, over the ziti mesh (MONKY-ADR-0028)"
 readme = "README.md"
 requires-python = ">=3.12"