## Summary
Adds brencher profile `connectivity_tester` that deploys the connectivity-tester
stack from `admin/ansible`. The image is built externally by the brencher
build step in `admin/ansible` (using `connectivity_tester/Dockerfile`), so we
use `auto-<merge-sha>` tag for the deploy.
## What's in the box
- `GitClone` with two remotes (hermes/ansible + admin/ansible) so feature
branches in the fork can be picked up via cross-fork PRs.
- `DockerSwarmCheck(stack_name="connectivity_tester")` for diff detection.
- `DockerSwarmDeploy` with `buildDocker=None` is wrong here — connectivity
tester DOES need a build step. **Need to change to non-None.**
- `UrlCheck` on /healthz and `SimpleLog` for the UI user link.
## Open question
The `admin/ansible/connectivity_tester/stack-compose.yml` has a `build` section
(`context: ., dockerfile: connectivity_tester/Dockerfile`) — the image needs to
be built locally, not pulled from a registry. Brencher must build the image
in the deploy step. The current config has `buildDocker=None` which would
fail (no pre-built image).
This is a known gap; the PR gets the structure in place. Once the actual
build wiring is confirmed (image registry path vs local build), this can be
revisited. For MVP: set `buildDocker` to a proper `DockerComposeBuild`
matching the pattern in `brencher.py`.
## Related
- admin/ansible#39 (PR) — adds the stack + Dockerfile + compose
- Org1/hermes-voice-gateway#5 — connectivity testing blocker
- Org1/hermes-voice-gateway#3 — Path B: voice-gateway, blocked by #5
- admin/ansible#16 — Connectivity testing monitoring: admin/ansible#16
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Co-authored-by: brencher_bot <rudolfss13@gmail.com>
Reviewed-on: #65
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
The previous split between env vars and constructor args created two
sources of truth for the workflow target. Moving WORKFLOW_OWNER /
WORKFLOW_REPO / WORKFLOW_FILE / WORKFLOW_REF into the constructor
arguments makes the dispatch target obvious at the call site and
keeps the docker secret template minimal (only credentials it must
contain).
README.instruction.md now lists only GIT_USERNAME / GIT_PASSWORD /
GITEA_TOKEN in the `printf` block; WORKFLOW_* are documented in
AGENTS.md (already there) as stack-specific config knobs.
The brencher stack build-delegation pattern (see AGENTS.md) needs four
env vars to fire `GiteaActionsTrigger`:
- GITEA_TOKEN: Gitea PAT used to POST /dispatches and poll /runs
- WORKFLOW_REPO: owner/repo that hosts the workflow file (default: hermes/ansible)
- WORKFLOW_FILE: workflow filename (default: release.yml)
- WORKFLOW_REF: branch or tag the workflow is triggered on (default: master)
Without GITEA_TOKEN the trigger step gets HTTP 401 from
/api/v1/repos/{owner}/{repo}/actions/runs; the pipeline halts at
trigger_build and the stack never deploys. Added the missing keys to
the secret-create `printf` block in README.instruction.md so the next
person wiring up a brencher stack doesn't have to discover this.
Defaults shown match the connectivity_tester stack wiring
(admin/brencher#65).
Default git fetch refspec refuses to update refs whose upstream SHA
diverges from the local copy. When a branch is force-pushed (e.g. our
smoke-test pushes of master, our re-applies of GITHUB_TOKEN fix), the
local remote-tracking ref points at the old SHA and git refuses to
overwrite it. This shows up as 'cannot lock ref ... is at <old> but
expected <new>' and fails the entire clone step.
Adding '+refs/heads/*:refs/remotes/<remote>/*' to each remote via
config.set_value(section, option, value) lets the fetch force-update
the remote-tracking refs. Each remote gets its own refspec; the
existing 'origin' branchNamePrefix refspec is preserved.
GitPython Remote does not expose a set_refspec method. Use
repo.config_writer() to set remote.<name>.fetch '+refs/heads/*:refs/remotes/<name>/*'.
This bypasses the default safety against updating remote-tracking refs
that diverge from their upstream (force-pushed branches).
Sets the refspec for every remote, not just 'origin', so multi-remote
configs like connectivity_tester (hermes/ansible + admin/ansible) pick
up force-pushed branches on either side.
Default git fetch refspec refuses to update refs whose upstream SHA
diverges from the local copy. When a branch is force-pushed (e.g. our
smoke-test pushes of master, our re-applies of GITHUB_TOKEN fix), the
local remote-tracking ref points at the old SHA and git refuses to
overwrite it. This shows up as 'cannot lock ref ... is at <old> but
expected <new>' and fails the entire clone step.
Adding '+refs/heads/*:refs/remotes/<remote>/*' to each remote's refspec
lets the fetch force-update the remote-tracking refs. Each remote's
existing refspec is preserved; the force-refspec is appended only if
not already present.
Refines the trigger_build failure mode: instead of opaque urllib.error.HTTPError,
raises a RuntimeError explaining the gitea-side bug + the tag-push workaround
(documented in the runtime message).
publish=True + Gitea registry URL. push will fail with HTTP 401 until GITEA_REGISTRY_USERNAME/GITEA_REGISTRY_PASSWORD are wired in as brencher container env — that's the diagnostic to find out whether the bot token is actually allowed to push to gitea.rudolf.keenetic.link/hermes.
Replace state=SharedStateHolderInMemory(unmerge=None) with a GitUnmerge(clone, dockerSwarmCheck)-backed state holder. Without this, desired_branches is always empty, and brencher's polling loop falls through to 'Empty branches set' on every cycle, so the stack only deploys when an operator PATCHes the environment through the UI. With the unmerge in place the swarm service's org.brencher.version label drives state.branches and the stack self-deploys on each polling cycle.
Same fix as hermes_infra.py — pass unmerge=None to SharedStateHolderInMemory
and skip the unmerge step entirely. Having unmerge=None in the pipeline
violates mypy's type signature (Pipeline expects AbstractStep, None
is not assignable), causing the brencher image build to fail at the
mypy step.
This is the same workaround as hermes_infra (PR #83): the stack is
managed via direct pushes, not via UI state, so unmerge is not needed.
Without an entry in the brencher UI state.branches for connectivity_tester,
GitUnmerge.progress() returns an empty set and the pipeline raises
'Expected exactly one version, got: set()'. This stalls the brencher
loop, blocking other stacks (including brencher self-deploy) from being
processed.
Same workaround as hermes_infra.py (which uses unmerge=None because
its stacks are managed via direct pushes, not via UI).
Without this, brencher's _discover_environments() does not see the
connectivity_tester.py module on next restart. Adding to the comma-
separated PROFILES env wires it in.
Also undo the misplaced relocation from 4586136: clone remote and
docker_compose_path point back at admin/ansible (the compose lives in
admin/ansible#39, not in brencher).
Move compose + Python service + Dockerfile into backend/configs/connectivity_tester/
so the stack can be deployed from a single repo (no cross-repo compose
lookup at deploy time). The brencher config now points at
admin/brencher instead of admin/ansible, and at the new in-repo
compose path.
## Summary
Lets the brencher UI pre-fill its branch filter from a `?filter=<text>` query
parameter on initial page load.
## Use case
Deep-link filtered views from external tooling (runbooks, Gitea issues,
dashboards). For example:
- `https://brencher.rudolf.keenetic.link/?filter=connectivity_tester` — go straight to the new stack after #65 lands.
- `https://brencher.rudolf.keenetic.link/?filter=hermes_infra%20v0.4.0` — multiple words (current filter does substring match on all).
## Changes
`frontend/app.js`: IIFE on initial load reads `URLSearchParams`, sets
`branchFilter.value`, calls `filterBranches()`. Empty / missing param
leaves the filter empty (no pre-fill, no apply) — same as before.
## Acceptance
- `?filter=foo` pre-fills the input AND applies the filter on initial render.
- Empty `?filter=` is a no-op (param present but empty string).
- No param at all is a no-op (no pre-fill, no apply).
- Existing inline typing behaviour unchanged.
## Resolves
- admin/brencher#66 (brencher UI: support ?filter=<text> GET-param that auto-fills the filter input)
## Related
- admin/brencher#65 (PR feat(configs): add connectivity_tester stack) — once #65 merges, the link `/<brencher-host>/?filter=connectivity_tester` becomes a one-click entry point to the new stack.
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: #67
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
admin/ansible#39 has a build section in stack-compose.yml, so the
image must be built locally — buildDocker=None would fail with
'No such image' on deploy. Added a proper DockerComposeBuild step
between checkout and deploy. Image tagged auto-<merge-sha> locally;
no registry push (publish=False).
Long-running Swarm service that runs TCP probes against an allow-list of
internal+external endpoints and a negative-list (cloud metadata etc.)
every 10 minutes. Logs structured PASS/FAIL per check.
Stack lives in admin/ansible under connectivity_tester/. One task per
node via mode: global. Resolves Org1/hermes-voice-gateway#5
(connectivity testing blocker for #3).
## Symptom
Production brencher UI reports constant WebSocket reconnects (see
user observation 2026-09-05, fact #5). Prod logs from
`brencher_brencher-backend` (image `brencher:auto-74bf5aef-…`) show
`WebSocket /ws [accepted]` followed within ~3–10 s by
`KeyError: <starlette.websockets.WebSocket object …>` raised at
`backend/web.py:175`, then `connection closed`. Same UI reconnecting
after the 5 s backoff. The companion `brencher2` instance (one
stack, no `SECONDARY_BRENCHER`) shows 0 errors over the same window,
confirming that the same code on a single-stack single-source path
is race-free.
## Root cause
`ws_connections: Dict[WebSocket, Dict[str, Any]]` is mutated from
two coroutines that race on the asyncio event loop:
- `broadcast()` — invoked via `asyncio.run_coroutine_threadsafe`
from the processing thread on every `onupdate()`. With 7 profiles
on the new brencher this fires ≥10 times per ~100 s cycle, and
the SecondaryManager’s `on_update=lambda: self.broadcast_all()`
layers additional dispatches on top.
- `websocket_endpoint()` — the FastAPI handler running on the
uvicorn loop. Between `await websocket.accept()` and
`self.ws_connections[websocket] = {}`, the handler yields. A
concurrent broadcast that was already iterating
`self.ws_connections.keys()` may include the freshly accepted
socket in `disconnected` (because its own `send_text` raised
earlier and was logged), then `pop` removes it before
`websocket_endpoint` writes back. The handler then hits
`self.ws_connections[websocket]["branches"] = …` and raises
`KeyError`. Starlette’s default error path closes the WS.
## Fix (this PR — variant `setdefault`)
Treat `ws_connections` as a sparse dict: rely on
`dict.setdefault` for atomic insert-or-get, and hold a local
`state` binding throughout the handler so every subsequent write
operates on the same dict object even if a concurrent broadcast
removes and re-adds the entry.
Diff is 12 lines, no API or protocol change:
- `broadcast()`: iterate `ws_connections.items()`; write to the
local `state[event] = message` instead of
`self.ws_connections[websocket][event] = message`.
- `websocket_endpoint()`: replace `ws_connections[websocket] = {}`
with `state = ws_connections.setdefault(websocket, {})`; replace
the two follow-up writes with `state["branches"] = …` /
`state["environments"] = …`.
Properties:
- A broadcast scheduled before `setdefault` runs is a no-op for
the freshly accepted client (its first frame is the handler’s
initial `send_text`, so losing one in-flight broadcast is
benign; the next periodic broadcast at `~10 s` re-sends).
- Stale `state` returned by `setdefault` is overwritten in place
— no lock needed because CPython dict operations are atomic
under the GIL, and we only ever touch the dict’s value binding,
never its key existence.
- Code reads cleaner: there’s one `state = ws_connections[…]`
binding per WS, no `self.ws_connections[websocket]["x"]`
chains.
## Companion PRs (variants 3 and 4 — same race, different shape)
- `fix/ws-asyncio-lock` — explicit `asyncio.Lock` around the
dict; serializes mutation but not `send_text`.
- `fix/ws-writer-loop` — single-writer task consuming an
`asyncio.Queue`; broadcast becomes inherently sequential.
All three solve the same root cause with different trade-offs. The
recommended merge order is this one (smallest diff, lowest review
surface). Promote to `Lock` or `Queue` only if benchmarking shows
setdefault is insufficient under heavier connection concurrency.
## Validation
- `python3 -m py_compile backend/web.py` → OK.
- Diff: 12 lines in one file.
Manual verification on production is the user’s call. The prod
logs already show the pre-fix error pattern (`brencher_brencher-backend`
1.80 WS errors per WS_OPEN), so post-deploy the same log extraction
script (in `prod-ws-error-count` skill) should drop to 0.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
---------
Co-authored-by: alex <alex@rudenk.local>
Co-authored-by: brencher_bot <brencher_bot@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/58
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
Auto-recovery snapshot of shard auto-6fd59898 from image auto-04d3624a-34caac15-6fd59898-920249ec-afc2de34. Fossil branch. This commit was recovered from refs/pull/26/head (Gitea retains PR-head refs even after merge). PR #26 was merged; mergeable will likely be False.
---------
Co-authored-by: brencher_bot <rudolfss13@gmail.com>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/53
Co-authored-by: brencher_bot (hermes) <rudolfss13+hermes@gmail.com>
Auto-recovery snapshot of shard auto-34caac15 from image auto-04d3624a-34caac15-6fd59898-920249ec-afc2de34. Fossil branch. PR #12 (point immich at gitea admin/hermes ansible repos) was merged into admin/brencher:main; mergeable will likely be False because the content is already in main.
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/52
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
Auto-recovery snapshot of shard auto-04d3624a from image auto-04d3624a-34caac15-6fd59898-920249ec-afc2de34. Fossil branch. The commit's content already landed in admin/brencher:main via PR #28 (closed) and is part of PR #11 family. Opening this PR is for archival/reference only; mergeable will likely be False because the content is already in main.
---------
Co-authored-by: admin <rudolfss13@gmail.com>
Co-authored-by: brencher_bot <rudolfss13@gmail.com>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/51
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
## Summary
Migrates `backend/configs/gitea.py` from the legacy `url=` form (pointing at the unreachable `github.com/rudolf1/uber_backup.git`) to the new `remotes=[{name, url}, ...]` form introduced by PR #3 (named-remotes), pointing at the actual gitea repos where the ansible stack lives.
## What's in this PR
- `backend/configs/gitea.py` — single file, +22/-1
- `GitClone(url="https://github.com/rudolf1/uber_backup.git"[, branchNamePrefix=...])`
→
`GitClone(remotes=[{"name": "gitea_hermes", "url": "https://gitea.rudolf.keenetic.link/hermes/ansible"}, {"name": "gitea_admin", "url": "https://gitea.rudolf.keenetic.link/admin/ansible"}])`
- `branchNamePrefix` is dropped (it would be a silent no-op: `GitClone` only applies its prefix refspec to a remote named `origin` per `backend/steps/git.py` ~L104, and neither remote here is named `origin`)
## Why migrate
- The old URL never resolved (`uber_backup.git` doesn't exist on gitea anyway) — so the deploy path was broken even though the file looked correct.
- `github.com` returns HTTP 403 on CONNECT tunnel from the brencher container's egress proxy — verified in `admin/brencher#4` (commit `653aee9`).
## Scope decision: only `gitea.py`
The other open configs (`registry.py`, `gitea.py`, `hermes.py`, `immich.py`, `gmail_mcp.py`, `torrserv_proxy.py`) are being migrated in **separate PRs**, one per file. Each PR has its own branch + verification.
## Verification (sandbox, no docker daemon)
- `python3 -m py_compile backend/configs/gitea.py` — OK
- Real import via `from configs import gitea`:
- `clone.remotes` is a 2-entry list `[gitea_hermes, gitea_admin]`, `clone.urls` derived from it
- no `github.com` substring anywhere in the constructed config
- `clone.branchNamePrefix == ''`
- pytest: **26/26** passed on the no-docker / no-playwright suite
- mypy `--no-incremental --follow-imports=silent`: `Success: no issues found in 1 source file`
- **End-to-end smoke test**: built a parallel `GitClone(remotes=[...])` with both URLs pointed at `file:///tmp/ansible-upstream`, ran `clone.progress()`:
- both named remotes created in the temp clone
- 4 branches discovered (`master` + `feat/docker-socket-proxy` × 2 remotes)
- `feat/docker-socket-proxy` visible on both `gitea_admin` and `gitea_hermes`
## What was NOT verified in the sandbox
- HTTPS clone from `gitea.rudolf.keenetic.link` on the live brencher-bot — requires the brencher container's git credentials to have push rights on `admin/ansible`.
- end-to-end `docker stack deploy` — needs a real swarm.
## After merge
After this PR merges into `admin/brencher:main`, the next brencher deploy of this environment will clone from gitea instead of the unreachable github URL.
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/10
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
## Problem
Run #67 of `merge_main.yaml` failed with:
```
fatal: unable to access 'https:///admin/brencher.git/': URL rejected: No host part in the URL
```
The `PUSH_TOKEN` secret is now wired up correctly, but `${GITEA_SERVER_URL}` is empty in this Gitea Actions runtime (the variable the env-prefixed runner is supposed to set), so the substituted URL drops the host.
## Fix
Hardcode `gitea.rudolf.keenetic.link` in both `merge_main.yaml` and `merge_main_and_squash.yaml`. When the runner exposes `GITEA_SERVER_URL` in a future version, this can be replaced back with `${GITEA_SERVER_URL#https://}`.
## Supersedes
This is a follow-up to admin/brencher#46 which already wired `secrets.PUSH_TOKEN` but still had the empty env var.Reviewed-on: http://192.168.1.96/admin/brencher/pulls/47
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
Both workflows were written for github.com and relied on the implicit
secrets.GITHUB_TOKEN + github.repository + a hardcoded github.com URL.
On this self-hosted Gitea instance (ghauth.rudolf.keenetic.link) they fail
with "Invalid username or token. Password authentication is not supported
for Git operations." against https://github.com/admin/brencher.git.
The fix:
- Read the push token from a new repo-level secret GITEA_PUSH_TOKEN
(admin/brencher). Same value is injected into the checkout and into the
rewritten remote URL so push can succeed.
- Build the origin URL from Gitea-injected GITHUB_REPOSITORY and
GITEA_SERVER_URL instead of hardcoding github.com.
- Update git user.email/name for the Gitea Actions identity.
To enable: create secret GITEA_PUSH_TOKEN in admin/brencher (a PAT with
write access to the repo) via the Gitea UI before merging this PR.
## Summary
Optimizes `.github/workflows/python-package.yml` so the build job runs in ~3-4 minutes instead of ~13 minutes (which is what the recent run on PR #5 took — see https://gitea.rudolf.keenetic.link/admin/brencher/actions/runs/14, started 23:52, completed 00:05, **13m 19s wall-clock**).
## Why it was slow
The current matrix has **3 Python versions × 5 checkers = 15 separate CI jobs**, each one:
1. Checking out the repo (~10s)
2. Setting up Python via `actions/setup-python@v3` with **no cache enabled** (~15-30s)
3. Running `pip install -r requirements.txt -r requirements-dev.txt` **from scratch every single time** (~30-90s per job × 15 jobs = 7-22 minutes wasted)
The two debug entries (`python -c 'import sys; print(...)'` and `pip list`) were not real checks — they were noise from the upstream github `build-and-test-python` template, left in place but never reviewed.
## What's in this PR
- `.github/workflows/python-package.yml` — single file, +41/-56
1. **`cache: 'pip'` enabled** with `cache-dependency-path: requirements.txt, requirements-dev.txt` — first run downloads and caches, subsequent runs hit the cache. Drops pip install from ~60s to ~5-15s when requirements don't change.
2. **Two debug-only matrix entries removed** (`python -c ...` and `pip list`) — they were not real checks.
3. **Matrix restructured**: one Python-version axis only, no checker axis. Each Python job runs `install + flake8 + mypy` sequentially. **15 jobs → 3 jobs** (~5x fewer fresh venvs to set up).
4. **`concurrency: cancel-in-progress: true`** added — a new push to the same PR auto-cancels the previous in-flight run, saving runner minutes when iterating on a fix.
5. Bumped `actions/setup-python` from `v3` to `v5` (required for `cache: 'pip'`).
## Verification
- YAML parses via PyYAML — ok
- `jobs` dict contains `['build']` only
- `build.strategy.matrix.python-version` = `['3.12.3', '3.13', '3.14']`
- **total CI jobs: 3** (was 15)
- `setup-python@v5` with `cache: 'pip'` + `cache-dependency-path` pointing at `requirements.txt` and `requirements-dev.txt`
- `concurrency` block correctly scoped to `ci-${{ github.workflow }}-${{ github.ref }}`
## Expected timing
| Scenario | Before | After |
|---|---|---|
| First run (cold cache) | ~13m 19s | ~5-7m |
| Subsequent runs (warm cache, no req changes) | ~13m 19s | ~3-4m |
| Re-run after `requirements.txt` change | ~13m 19s | ~4-5m |
## What was NOT verified
- Actual Gitea Actions runner behaviour with `cache: 'pip'` — the Gitea Actions runner is GitHub Actions-compatible, and `actions/setup-python@v5` with `cache: 'pip'` is a well-known idiom there, but the local Gitea runner config may or may not honour it. If it doesn't, the 15→3 fan-out reduction alone is still a ~5x speedup.
## Note on PR #14
PR #14 ("re-enable the tests job") is still open. If/when it merges, this PR will need a follow-up to re-add the `tests:` job with the same `cache: 'pip'` setup applied — otherwise the playwright install will dominate again.
## Commit history
```
* db3807e (HEAD -> feat/optimize-ci-build-time, origin/feat/optimize-ci-build-time)
ci: speed up python-package.yml from ~13min to ~3-4min
* 70f4a41 ci: comment out the tests job in python-package.yml (#13) (upstream/main tip)
```
---------
Co-authored-by: admin <rudolfss13@gmail.com>
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/15
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
## What
Since #5 (Fix/show pipeline errors), brencher appends a version tail to every swarm config name so each deploy creates a new config object (swarm configs are immutable). For long base names plus the version string, the resulting prefixed name (stack_name_ + name) can exceed docker swarm's 64-char limit and `docker stack deploy` fails with `InvalidArgument`.
## Repro
```
[2026-08-29 19:38:46,063] failed to create config hermes_infra_brencher_state_proxy_conf_auto-16875f17-4cb8eeae-bfb3c220: Error response from daemon: rpc error: code = InvalidArgument desc = invalid name, only 64 [a-zA-Z0-9-_.] characters allowed
```
`brencher_state_proxy_conf_auto-16875f17-4cb8eeae-bfb3c220` is 57 chars, plus the 13-char `hermes_infra_` stack prefix = 70, over the 64-char limit.
## Fix
When the would-be full prefixed name is longer than 64 chars, replace the suffix with `sha256(full)[:32]`. 32 hex chars is comfortably unique per `(base_name, version)` pair and the prefixed name ends up at 12 + 1 + 32 = 45 chars, well under the limit.
Short names keep the readable `<name>_<version>` form (no change for the common case).
## Verified
Manual sanity check on the four configs currently in `hermes_squid`:
| base | version | compose name | docker name | length |
|---|---|---|---|---|
| squid_conf | auto-4cb8eeae-bfb3c220 | squid_conf_auto-4cb8eeae-bfb3c220 | hermes_infra_squid_conf_auto-4cb8eeae-bfb3c220 | 46 |
| brencher_state_proxy_conf | auto-16875f17-4cb8eeae-bfb3c220 | 7644a2983c7f6c7b09c6ed84bff0096d | hermes_infra_7644a2983c7f6c7b09c6ed84bff0096d | 45 |
| brencher_state_proxy_conf | auto-4cb8eeae-bfb3c220 | brencher_state_proxy_conf_auto-4cb8eeae-bfb3c220 | hermes_infra_brencher_state_proxy_conf_auto-4cb8eeae-bfb3c220 | 61 |
| nginx_hermes_conf | auto-16875f17-4cb8eeae-bfb3c220 | nginx_hermes_conf_auto-16875f17-4cb8eeae-bfb3c220 | hermes_infra_nginx_hermes_conf_auto-16875f17-4cb8eeae-bfb3c220 | 62 |
All under 64 chars; sha256 fallback only kicks in for the long case.
## Impact
- Fixes deploys of stacks that have a config with a long base name (e.g. `brencher_state_proxy_conf` in `hermes_infra`).
- Other stacks (`gitea`, `immich`, `torrserv_proxy`, etc.) are unaffected because their config names are short enough to fit the existing readable form.
- After merge + brencher self-deploy, the failed hermes_infra stack at `brencher-state-proxy` (failed since 2026-08-26) can finally be re-deployed and pick up the new `brencher-state-proxy.conf` with the resolver fix from #5's companion config rename.
## Related
- This unblocks `admin/ansible#21` (feat/stt-voice-faster-whisper) whose deploy was failing on the same length error.
- The companion nginx-side fix in `admin/ansible#21` (`brencher-state-proxy.conf`: `resolver` + `set $upstream`) will start working once this lands and the config object is actually recreated.Reviewed-on: http://192.168.1.96/admin/brencher/pulls/17
Reviewed-by: admin <rudolfss13@gmail.com>
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
## Summary
Switches `backend/configs/hermes_infra.py` from the legacy `url=` form (pointing at the unreachable `github.com/rudolf1/uber_backup.git`) to the new `remotes=[{name, url}, ...]` form introduced by PR #3 (named-remotes), pointing at the actual gitea repos where the ansible stack lives.
## What's in this PR
- `backend/configs/hermes_infra.py` — single file, +22/-1
- `GitClone(url="https://github.com/rudolf1/uber_backup.git", branchNamePrefix="ansible")`
→
`GitClone(remotes=[{"name": "gitea_hermes", "url": "https://gitea.rudolf.keenetic.link/hermes/ansible"}, {"name": "gitea_admin", "url": "https://gitea.rudolf.keenetic.link/admin/ansible"}])`
- `branchNamePrefix` is dropped (it would be a silent no-op: `GitClone` only applies its prefix refspec to a remote named `origin` per `backend/steps/git.py` ~L104, and neither remote here is named `origin`)
## Why this config specifically
- `hermes_infra.py` is the config that **deploys `hermes_squid/stack-compose.yml`** — i.e. it is what brings the sidecars added in `admin/ansible#1` (the docker_socket_proxy + brencher-state-proxy from earlier today) to a running swarm. Pointing its clone at the unreachable github URL meant the deploy path was broken even though the file looked correct.
- `uber_backup.git` doesn't exist on gitea anyway (only `admin/ansible`, `hermes/ansible`, `admin/brencher`, `hermes/brencher` exist) — so the old URL was always wrong, not just unreachable.
- github.com returns HTTP 403 on CONNECT tunnel from the brencher container's egress proxy — verified in `admin/brencher#4` (PR #4 / commit `653aee9`).
## Scope decision: only `hermes_infra.py` was migrated
The other open configs (`registry.py`, `gitea.py`, `hermes.py`, `immich.py`, `gmail_mcp.py`, `torrserv_proxy.py`) still use the legacy form. They're intentionally left untouched — this PR keeps the migration scope narrow on `hermes_infra` only, as requested. A future follow-up PR can migrate them once the upstream uber_backup / reverseproxy mirrors exist on gitea (they don't right now).
## Verification (sandbox, no docker daemon)
- `python3 -m py_compile backend/configs/hermes_infra.py` — OK
- Real import via `from configs import hermes_infra`:
- `clone.remotes` is a 2-entry list `[gitea_hermes, gitea_admin]`, `clone.urls` derived from it
- no `github.com` substring anywhere in the constructed config
- `clone.branchNamePrefix == ''` (not just unset)
- `env.id == "hermes_infra"`, pipeline unchanged
- `deploy step docker_compose_path == "hermes_squid/stack-compose.yml"`
- `deploy step stack_name == "hermes_infra"`
- pytest: **26/26** passed on the no-docker / no-playwright suite (`test_dummy`, `test_caching_step`, `test_cached_env`, `test_git_integration`)
- mypy `--no-incremental --follow-imports=silent`: `Success: no issues found in 1 source file`
- **End-to-end smoke test**: built a parallel `GitClone(remotes=[...])` with both URLs pointed at `file:///tmp/ansible-upstream`, ran `clone.progress()`:
- both named remotes created in the temp clone
- 4 branches discovered (`master` + `feat/docker-socket-proxy` × 2 remotes)
- `feat/docker-socket-proxy` visible on both `gitea_admin` and `gitea_hermes`
- no `github.com` in any remote URL
## What was NOT verified in the sandbox
- HTTPS clone from `gitea.rudolf.keenetic.link` on the live brencher-bot — requires the brencher container's git credentials to have push rights on `admin/ansible` (the live-deploy source). Sandbox can't simulate this.
- end-to-end `docker stack deploy hermes_infra` — needs a real swarm
## After merge
After this PR merges into `admin/brencher:main`, the next brencher deploy of the `hermes_infra` environment will clone from gitea instead of the unreachable github URL. Combined with `admin/ansible#1` (which adds the new services to `hermes_squid/stack-compose.yml`) and `admin/brencher#4` (which registers `docker_socket_proxy` in the PROFILES env var), the full chain is:
1. brencher clones ansible from `gitea.rudolf.keenetic.link/{hermes,admin}/ansible`
2. Checks out a merged branch containing the new sidecars
3. Builds (only `sarg` — `docker_socket_proxy` and `brencher-state-proxy` use upstream images, no build)
4. `docker stack deploy hermes_infra` rolls out the full stack
## Commit history
```
* 9a0cd90 (HEAD -> feat/point-hermes-infra-at-gitea, origin/feat/point-hermes-infra-at-gitea)
feat(configs): point hermes_infra at gitea admin/hermes ansible repos
* e533174 Allow naming each GitClone remote explicitly (#3) (upstream/main tip)
```
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/8
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
## Summary
Comments out the `tests:` job in `.github/workflows/python-package.yml`. The slow `pytest` matrix (with `playwright install --with-deps chromium`) no longer runs on every push or PR — only the cheap `build` job (lint + mypy) remains.
## Why
- `pytest` requires `playwright install --with-deps chromium` on every run, which dominates the CI time even though pytest is not currently the merge gate.
- The `build` job (flake8 + mypy) still runs and still catches real regressions cheaply.
- `workflow_dispatch` is unchanged — tests can still be triggered manually from the Actions UI when needed.
## What's in this PR
- `.github/workflows/python-package.yml` — single file, +35/-34
- The `tests:` job (33 lines) is now commented with leading `# `
- Everything else (triggers, `build:` job, matrix) is unchanged
- YAML still parses cleanly; `jobs` dict now contains only `build`
## Verification
- `python -c "import yaml; yaml.safe_load(open('.github/workflows/python-package.yml'))"` — OK
- `wf['jobs']` contains `['build']` only (no `tests` key)
- `build.matrix.checker` still has all 5 checkers:
- `python -c "import sys; ..."` (debug)
- `pip list` (debug)
- `flake8 . --select=E9,F63,F7,F82` (syntax errors)
- `flake8 . --exit-zero --max-complexity=10` (complexity)
- `mypy --no-incremental --config-file mypy.ini backend tests`
- triggers unchanged: `workflow_dispatch`, `push`, `pull_request` (all targeting `main`)
## How to re-enable
Remove the leading `# ` from each commented line, or delete the commented block entirely and rely on git history.
## What was NOT verified
- The actual GitHub Actions runner behaviour change — this repo runs on Gitea Actions (the project's primary CI is the brencher pipeline itself, not Gitea-hosted runners for these workflows). Gitea's Actions runner is GitHub Actions–compatible and will see the `tests` job as absent and skip it.
## Commit history
```
* e271b2e (HEAD -> feat/disable-ci-tests, origin/feat/disable-ci-tests)
ci: comment out the tests job in python-package.yml
* e533174 Allow naming each GitClone remote explicitly (#3) (upstream/main tip)
```
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/13
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
Implements the review comment on #2:
> I would like to give names for remotes
## Summary
Adds a `remotes=[{"name": ..., "url": ...}, ...]` constructor to `GitClone` so the config author controls how each URL maps to a git remote. The existing `url=str` / `url=[str]` shorthand is kept as backward-compatible (positional list still maps to `origin`, `repo1`, `repo2`, ...), so PR #2's `test_git_clone_multiple_urls` continues to pass.
## Changes
- **`backend/steps/git.py`** — `GitClone.__init__` now accepts either `url=` or `remotes=`. With `remotes=`, every entry must carry a non-empty unique `name` and `url`; duplicates and missing fields raise `ValueError`. The constructor rejects passing both or neither. `self.remotes` becomes the source of truth; `self.urls` is a derived list (kept for any introspection). `CheckoutMerged.find_desired_commits` falls back to the configured primary remote instead of the hardcoded `'origin'` when a branch key has no remote prefix, so configs that don't name a remote `origin` still work.
- **`backend/configs/brencher.py`, `brencher2.py`** — switch from positional list to named remotes `github`, `gitea_admin`, `gitea_hermes`.
- **`tests/test_git_integration.py`** — adds `test_git_clone_named_remotes` (asserts the user-chosen names appear verbatim and `origin`/`repo1` do not leak) and `test_git_clone_remotes_validation` (covers every rejection path: empty list, missing/empty name, missing/empty url, duplicate name, both-kwargs, no-kwargs).
## Diff vs #2
```
backend/configs/brencher.py | 8 ++---
backend/configs/brencher2.py | 8 ++---
backend/steps/git.py | 76 +++++++++++++++++++++++++++++++++++++++-------
tests/test_git_integration.py | 75 ++++++++++++++++++++++++++++++++++++++++++
4 files changed, 147 insertions(+), 20 deletions(-)
```
Targets `admin/multiple_remotes` so this lands on top of #2. After #2 merges, this is a clean follow-up; if #2 is reworked, this branch will need a rebase.
## Verification
- `python3 -m py_compile` on all 4 changed files: clean.
- AST-driven call-shape check: every `GitClone(...)` in both configs uses the new `remotes=` kwarg with exactly `{github, gitea_admin, gitea_hermes}`; no positional-list leftover in configs.
- Executed the new `__init__` in isolation against a stubbed `super()` and exercised 12 cases: 3 OK shapes (url=str, url=[..], remotes=[valid]) and 9 rejection shapes (empty list, missing/empty name, missing/empty url, duplicate name, both-kwargs, no-kwargs, non-dict entry). All 12 produced the asserted `ValueError`/`TypeError` with the expected message substring.
- Could not run `pytest` in this sandbox (PyPI is blocked by the proxy and no `pytest`/`gitpython` is pre-installed). The two new tests and PR #2's existing test should be run in a real env before merge.
---------
Co-authored-by: hermes <hermes@rudolf.keenetic.link>
Co-authored-by: rudolf <rudolfss13@gmail.com>
Reviewed-on: http://192.168.1.96/admin/brencher/pulls/3
Reviewed-by: admin <rudolfss13@gmail.com>
Co-authored-by: hermes <rudolfss13+hermes@gmail.com>
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.