e2eaabca06
Full read-through of the project. Grouped by what actually breaks:
Browser concurrency (the invariant the design rests on)
- Every Playwright session now goes through scheduler.submit() onto the
single worker queue. "Refresh Rules", cookie import and the boot sync
were calling automation.py directly from Flask request threads, so
with gunicorn's threads=4 a refresh during a scheduled run opened a
second Chromium and both wrote /data/auth_state.json — which
Playwright writes itself, non-atomically. A torn write there reads
back as "session expired" and costs a pointless cookie re-import.
- Request handlers that need a result use run_browser_task(), which
waits REQUEST_WAIT_SECONDS (90) then raises BrowserBusy -> 503
without cancelling the work. Previously a slow in-request scrape could
exceed gunicorn's 120s timeout; killing that worker takes the
APScheduler jobstore and run queue with it, silently dropping every
pending verification poll.
Verification correctness (the feature that exists to catch silent failures)
- run_rule() no longer starts a poll loop when the watermark capture
itself failed. watermark_id was None both for "capture failed" and
"rule has no prior activity", and check_completion() skips its
comparison on None — so a failed capture matched the *previous* run's
Complete entry and reported a false "verified", the exact class of bug
the verification step was added to catch.
- A refresh now reconciles schedules.json against the live rule list.
Rule ids are stable, names are not, and the cached name is what the
Activity-page match keys on: a renamed rule kept running while
reporting "Not Confirmed" forever, training you to ignore the signal.
A deleted rule was worse — its card disappears while its cron jobs
keep firing, with no UI left to disable them. Skipped on an empty
scrape so a MassPrice redesign can't delete every saved schedule.
- The dashboard's session banner reads a session_state.json flag set by
_ensure_authenticated() instead of guessing from run_history.json.
list_rules() writes no history entry, so a failed refresh — often the
first symptom — never raised the banner, despite the docs saying it
covered "run/refresh".
Scheduler robustness
- rebuild_jobs() parses and adds each entry under its own guard. It runs
*after* clearing existing jobs, so one malformed entry escaping as an
exception left the scheduler holding no jobs at all — every schedule
silently dead until the next restart. _parse_entry() also accepts
HH:MM:SS (some browsers emit seconds, which the old two-value unpack
crashed on) and range-checks hour/minute.
- The API validates entries at save time rather than storing something
every rebuild skips, and the frontend surfaces save failures instead
of discarding them.
- An unknown TZ logs and falls back to UTC instead of crash-looping the
container with a traceback that doesn't obviously point at .env.
Auth and web hygiene
- Empty-but-set LOGIN_PASSWORD/SECRET_KEY now fall back to their
defaults. docker-compose always sets these, so a blank .env value
arrived as "" — and compare_digest("", "") is True, i.e. no login at
all. An empty submitted password is rejected outright too.
- Password compared as bytes: compare_digest raises TypeError on
non-ASCII str, turning any accented character into a 500.
- CSRF token per session, in a meta tag, echoed as X-CSRF-Token on all
four POST endpoints; explicit SameSite=Lax/HttpOnly.
- /api/* answers 401 instead of redirecting. fetch() follows a 302 to
/login, gets 200 HTML, and reads it as success — "Run Now" reported
"Queued" for runs that were never queued.
- ?next= is restricted to local paths (open redirect).
Smaller correctness fixes
- Timestamps render in TZ via a localtime filter; they were shown as raw
UTC while schedules run in America/Phoenix.
- Activity's 24h filter excludes unparseable timestamps rather than
including them, and treats naive values as UTC instead of raising.
- SameSite=None cookies are forced Secure — Chromium drops them
otherwise, silently, possibly losing the one cookie carrying the login.
- The Activity selector escapes the rule name, so a name containing a
quote no longer produces an unparseable selector.
- pytz pinned explicitly; it only ever resolved as an apscheduler 3.x
transitive dep, which 4.x drops.
Docs corrected where they disagreed with the code: auth_state.json is
written by Playwright, not through storage.py's atomic helper; the
banner's real trigger; and the deploy skill's claim that a rename causes
"rule id not found" (it can't — schedules are keyed by the stable id).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01H6TgffxTvYhqxykAbVQg2v
6 KiB
| name | description |
|---|---|
| deploy-massprice | Use when deploying MassPrice Scheduler, restarting the massprice container on Unraid, confirming a deploy landed, logging into MassPrice via the Session page, or troubleshooting its scheduled-rule automation. |
Deployment
Target: Unraid server, container path /mnt/user/appdata/massprice,
built from Dockerfile/docker-compose.yml in this repo. Container name
massprice, port 5057 (overridable via HOST_PORT — 5056 was
already taken by arr-summary, despite its own docs saying 5055) →
container port 5000.
Deploy mechanism: Forgejo push-to-deploy, the same pattern as
arr-summary, card-grader, lunch-notifier, n64-tracker — see
arr-summary's deploy-arr-summary skill or card-grader's CLAUDE.md
for the full runner mechanics. Push to main on the forgejo remote
(https://git.hippofam.com/ninja_hippo/massprice.git) triggers
.forgejo/workflows/deploy.yml, which syncs into
/mnt/user/appdata/massprice (excluding .git, .env, data/) and
runs docker compose up -d --build.
Restarting the container manually: use Compose, not the Unraid Docker UI — same reasoning as the other projects (every deploy recreates the container with a new Docker ID; Unraid's UI caches the old one). From the Unraid terminal:
cd /mnt/user/appdata/massprice && docker compose restart
docker compose up -d --force-recreate for a guaranteed-clean recreate;
docker ps --filter name=massprice to check real state.
Confirming a deploy landed: the footer renders
v<APP_VERSION> · <APP_VERSION_DATE> from constants near the top of
app.py. Bump them every commit.
First deploy checklist
- Create
/mnt/user/appdata/masspriceon Unraid and a real.envthere (copy.env.example, fill inLOGIN_PASSWORD,SECRET_KEY,HOST_PORT,TZ— and optionallyGMAIL_USER/GMAIL_APP_PASSWORD/ALERT_EMAILfor the session-expiry email alert, same Gmail-app- password pattern aslunch-notifier). - First
docker compose up -d --build— confirm the container starts and/loginresponds onHOST_PORT. - Log into the webui, go to the Session page, and follow its
instructions: log into massprice.tcgplayer.com in your own regular
Chrome, export its cookies with the Cookie-Editor extension, paste
them into the box, and click "Import Session". MassPrice's login is
behind hCaptcha, which blocks any form of automated login (confirmed
by testing — see
CLAUDE.md), so this manual export is the only path in, and needs repeating whenever the session eventually expires. - Click "Refresh Rules" on the Dashboard — confirms the saved session actually works and pulls the real rule list.
- Set up at least one schedule and use "Run Now" once to confirm a
click actually lands — cross-check by looking at MassPrice's own
Activity Log (
https://massprice.tcgplayer.com/Activity/List) for a fresh timestamp on that rule. - Only after step 5 succeeds, go to each existing rule on massprice.tcgplayer.com (three-dot menu → Edit Schedule) and set Status to Inactive, so native scheduling stops double-triggering runs and this tool becomes the sole source of scheduled runs.
Persistent state, never touched by deploys
.env holds LOGIN_PASSWORD, SECRET_KEY, and (optionally)
GMAIL_USER/GMAIL_APP_PASSWORD/ALERT_EMAIL. data/ (bind-mounted to
/data) holds schedules.json (the schedules set from the webui),
auth_state.json (the saved MassPrice session imported via the Session
page — a live bearer credential for the user's MassPrice account while
valid, see CLAUDE.md; losing this just costs one re-import, not data),
run_history.json, rules_cache.json, session_state.json (the
"session is known-bad" flag behind the dashboard banner), and
notify_state.json (the session-expiry alert's send-once dedupe flag).
Both are excluded from the deploy rsync (which runs --delete) and from
git.
If scheduled runs stop working
Check, in order:
0. Local Activity says success but nothing shows up in MassPrice's own
Activity Log — this exact symptom happened in production once
already: run_rule() was only opening MassPrice's "Are you sure?"
confirmation modal, never clicking its #Confirmation_Run button, so
zero requests ever fired despite no error being thrown. Fixed, but if
it recurs (e.g. a MassPrice redesign changes the modal/selectors),
this is the first thing to suspect — see CLAUDE.md.
- Dashboard banner / Activity page — a "session expired or missing"
message means go log in again via the Session page. "Rule id … not
found on /Rule/List" means that rule was deleted on MassPrice
(a rename can't cause it — schedules are keyed by the stable rule
id): click "Refresh Rules", which re-syncs
schedules.jsonagainst the live list, drops schedules for deleted rules, and updates cached names for renamed ones. Renames matter because run verification matches the Activity page by name — a stale one shows every run as "Not Confirmed" while it's actually succeeding, so refresh after renaming anything. - Is the container still running —
docker ps --filter name=massprice— a Chromium crash inside the container (rare, but possible ifshm_sizeever gets dropped fromdocker-compose.yml) would show as the container restarting. - Timezone —
TZin.envcontrols what "6:00 AM" means for every schedule; confirm it matches what the user actually expects before assuming a schedule itself is wrong. - Session import fails with "didn't authenticate" — the export was
probably taken from the wrong tab or before login finished. Re-export
only after you're actually looking at massprice.tcgplayer.com's own
Dashboard/Rules pages, not the login page or
store.tcgplayer.com. - Session import fails with "none of the pasted cookies are for tcgplayer.com" — the extension exported the wrong site's cookies; make sure the active tab was massprice.tcgplayer.com when you clicked Export.