mirror of
https://github.com/maziggy/bambuddy.git
synced 2026-10-18 11:56:01 +02:00
dev
7
Commits
| Author | SHA1 | Message | Date | |
|---|---|---|---|---|
 maziggy
|
ebc73e46ba |
Say when AMS drying was running and a print never started (#2758)
Dispatching to an X2D with two AMS units mid-drying failed silently: the file uploaded, the printer accepted it and stayed idle. The watchdog waits for an active state or HMS_MQTT_VERIFY_FAILED, and a drying refusal is neither, so it timed out, re-uploaded the whole 3MF twice more, and closed with advice about the printer screen and the SD card. Studio, asked directly, said it could not start the job because of the drying. Latch the AMS dry_time telemetry across both watchdog phases and name the units in the give-up message, plus an INFO log on every failed window so the correlation reaches a support bundle from the first attempt. Detection only, no gate. These models support drying CONTINUING through a print (supports_drying_while_printing covers X2D from 01.01.00.00), so drying is not incompatible with printing and stopping it before every dispatch would tear down cycles the hardware is happy to run. One of the two units was also drying without its external PSU, which would make this a power budget problem at start-of-print calibration rather than a drying one -- dry_sf_reason 1/8 exist for exactly that. The message names both possibilities rather than asserting one. Also correct _sync_drying_state's docstring, which claimed to adopt drying it did not start; it only prunes. Behaviour unchanged -- populating it would let the scheduler stop a cycle the user started by hand. |
||
|
maziggy
|
5e2b7b53e6 |
fix(printers): surface the printer's own "command verification failed"
A P1S on firmware 01.10.00.00 rejected every control command and said so: HMS 0500-0500-0001-0007, "MQTT command verification failed". Bambuddy received that, dropped it, and reported a healthy printer instead. The frontend filtered it out. This code's meaning lives in attr's low half (0500) and code's high half (0001), both of which the MMMM_EEEE short form discards, so it collapsed to "0500_0007" — no catalog entry, no firmware actions, and filterKnownHMSErrors drops uncatalogued action-less errors. Catalog lookups now try full_code first, in both the description and the filter, and errors matched that way display the four-group code the printer's own screen shows. The remedy line is ours, not Bambu's: their wiki says to update Studio or Handy, which does not apply to a print sent from Bambuddy. The developer-mode probe made it worse. It read anything that was not an explicit refusal as confirmation, and this firmware answers the probe with an empty result while refusing everything else — so an inference drawn from a non-answer became "developer_mode: pass" in the support bundle of a printer that had not accepted a command all day. The probe now has three outcomes: explicit success enables, explicit verify-failure disables, anything else stays unknown and the diagnostic reports skip. The HMS is authoritative over that inference in both directions. It forces developer_mode False when present, and clears back to unknown when the printer stops reporting it, so enabling Developer Mode and restarting the printer is picked up without restarting Bambuddy. Dispatch no longer treats a refusal as a wedge. The watchdog latches the HMS across both phases and fails the item on the first attempt naming the code and the fix, rather than spending three uploads and 270s a lap to arrive at a message about SD cards. The check runs after the active-state exit in both phases, so a lingering HMS can never abort a print that is visibly running. Also: the "wrong or mis-cased serial number" hint no longer fires in the moment after a reconnect. _report_messages_since_connect is reset by _on_connect, so a reconnect landing microseconds before the staleness check leaves it at 0 for reasons that have nothing to do with the serial — this reporter's healthy printer was told to go check its serial 1 ms after reconnecting. |
||
|
maziggy
|
ce807fb1cc |
fix(queue): upload to printers in parallel, cap wedge retries, make debug logs survive a farm
The reporter's 19-printer farm started prints "one by one", up to an hour apart. check_queue awaited each dispatch inline, and a dispatch includes the FTP upload, so every printer queued behind every other printer's transfer despite being an independent machine. His logs give the arithmetic: 40978500 bytes in 254.1s, 157 KB/s - a Bambu printer's SD write, not the network, is the bottleneck. Nineteen of those in series is ~80 minutes, and the next upload started 131 ms after the previous one finished. The delay is linear in fleet size, which is why it got worse the more printers he selected. Dispatch is now collected during the (still sequential) selection loop and run concurrently afterwards, capped by queue_max_concurrent_uploads - Settings -> Workflow -> Queue & Dispatch, default 4, 1 restores the old behaviour. Every gate is untouched; only the transfers overlap. The pass still awaits its uploads before returning: _start_print flips the row pending -> printing only after the upload, so an early return would let the next tick re-dispatch the same rows. FTP work moves to its own thread pool. It was on asyncio's default executor - min(32, cpu+4), six threads on a 2-core NAS, shared with everything else - which was survivable only while uploads were serial. Two problems the same bundle exposed: A printer that accepts project_file but never starts (#1678) was retried forever: 270s watchdog, revert to pending, re-upload the whole file, repeat. Hence his "printer who, since the morning, still not launch" - and on a farm each lap also eats an upload slot the other printers are waiting on. Attempts are now counted on the queue item; after three it fails with a message pointing at the printer instead of queueing a fourth re-upload. The debug bundle we asked him for held 4m49s of history. The push_status dumps fired on every frame rather than on change - several while their own comment claimed otherwise - which is 27,727 of the bundle's 29,830 lines and rolls 5 MB in under five minutes on 19 printers. They now log transitions only. The bundle also read just the live log while three rotated backups sat next to it, under a byte budget four times larger than the file it was reading. Migration verified on SQLite and Postgres: idempotent, backfills legacy NULLs (dispatch_attempts + 1 is NULL for a NULL row, which would silently disable the cap). Tests: 6 on concurrent dispatch (overlap, cap honoured, 1 == serial, default applies with no settings row, a failed printer does not cancel its siblings, no early return), 4 on the retry budget, 6 on the bundle's rotated-log span, 7 on the debug gating. Each verified to fail against the unfixed code - the first end-to-end log assertion I wrote passed without the fix and had to be tightened. |
||
|
maziggy
|
fdaff37975 |
fix(queue): two-phase dispatch watchdog so a printer that accepts project_file but never starts doesn't wedge the queue (#1678)
_watchdog_print_start now treats subtask_id-advance as Phase A "command landed", not as final success. Phase B (180s) keeps watching for the active-state transition; if it never arrives — printer accepted the file but stalled (cloud+LAN re-auth after a power cycle on old firmware was the reported trigger) — revert the queue item to 'pending' instead of leaving it stuck in 'printing' until container restart. Phase B skips force_reconnect because subtask_id-advance proves the project_file landed; a reconnect mid-parse would trigger 0500_4003 (#1150). Phase A's H2D 50s FINISH→PREPARE tolerance (#1078) is preserved by Phase B's 180s headroom. |
||
|
maziggy
|
5680f5d34b |
fix(scheduler): watchdogs no longer falsely treat FINISH->IDLE as "print landed" (#1370)
Both the queue-side _watchdog_print_start and the direct-dispatch _verify_print_response used `status.state != pre_state` to decide whether a project_file command had been accepted. When a printer was in FINISH at dispatch time (un-dismissed post-print prompt from a prior job), the firmware silently rejected the new command; if the user then dismissed the screen prompt, the printer moved FINISH -> IDLE and the watchdog returned early as "command landed" — leaving the queue row stuck at status='printing' indefinitely and the scheduler permanently marking the printer as busy. Narrow the "command landed" check in both verifiers to an allow-list of active-print states (PREPARE / SLICING / RUNNING / PAUSE). Inactive transitions (FINISH -> IDLE, etc.) no longer short-circuit the revert. The subtask_id-advance signal stays in place for H2D's slow FINISH -> PREPARE transition (#1078). Also wrap _watchdog_print_start's revert commit and printer_manager._persist_awaiting_plate_clear in run_with_retry so SQLite single-writer contention can't silently drop these writes. The revert path returns a tristate sentinel so the post-revert MQTT session-recovery logic only runs when we actually reverted (or the commit failed) — not when on_print_complete had already cleared the row, where a forced reconnect could break a healthy concurrent print. |
||
|
maziggy
|
69b6b5a334 |
fix(#1150): skip MQTT reconnect on watchdog timeout when project_file landed
Background: P1P firmware can take ~135 s after a project_file MQTT publish
to actually start parsing the uploaded .3mf — gcode_state stays IDLE and
subtask_id doesn't advance until parse completes. The dispatch watchdogs
treated the missed transition as a #887/#936 half-broken session and called
force_reconnect_stale_session, which interrupts the printer's in-progress
parse and triggers 0500_4003 ("can't parse print file") on the printer side.
Both #1150 (slow parse) and #887/#936 (zombie session) look identical from
state and subtask_id alone — both have stale state and stale subtask_id with
fresh telemetry. The distinguishing signal is the printer's gcode_file
field: it updates in push_status when the project_file command actually
lands on the printer, but stays unchanged when the publish was silently
swallowed.
Both watchdogs (_verify_print_response in background_dispatch and
_watchdog_print_start in print_scheduler) now capture pre_gcode_file from
printer_manager.get_status() before sending the publish, then on timeout
compare it against the last good status seen during the poll loop. If the
file changed, the command landed → log a #1150 warning, skip the forced
reconnect to avoid 0500_4003 mid-parse. If unchanged, fall through to the
original force_reconnect_stale_session call so the half-broken-session
recovery is preserved exactly.
Caveat documented in code: in a retry-same-file slow-parse scenario the
gcode_file looks identical pre/post-publish, so the watchdog falls through
to the reconnect path and the user still hits 0500_4003 on that retry.
Accepted to avoid breaking the half-broken-session recovery, which is the
more impactful regression of the two.
The new pre_gcode_file kwarg has a default of None on both watchdog
functions, so any caller that doesn't pass it keeps the original
reconnect-on-timeout behavior verbatim.
4 new unit tests cover both watchdogs: skip on gcode_file change (#1150
fix), reconnect when unchanged (#936 protection preserved), skip when
pre=None and current is non-None (printer just connected), reconnect when
pre_gcode_file arg is omitted (backward-compat). All 439 existing
dispatch / scheduler / mqtt tests pass unchanged.
|
||
|
maziggy
|
0918907dab |
fix(scheduler): watchdog falsely reverts slow H2D dispatches, causing reprints (#1078)
_watchdog_print_start reverted queue items to "pending" at 45 s if gcode_state hadn't changed, assuming the MQTT project_file was swallowed by a half-broken session (#887/#967). H2D Pro firmware (01.01.00.00) routinely keeps state=FINISH for 48-55 s after actually accepting the command before transitioning to PREPARE. The watchdog reverted items the printer had already started physically printing; the archive updated normally via _active_prints, but the queue item was now "pending" again, and the next scheduler tick after plate-clear re-dispatched the same item as if it had never run. With one item left in the queue that looked like a reprint of the just-finished job; with multiple items the symptom was masked by item N+1 getting dispatched during the race. Add a second "command landed" signal: subtask_id advancing past the pre-dispatch value. Bambuddy already mints a unique submission_id per project_file publish (#1042) and the printer echoes it back on the next push_status as soon as it starts processing the command - well before gcode_state transitions on slow-transition models. _start_print now captures pre_subtask_id alongside pre_state and passes both to the watchdog, which exits early on either a state change or a subtask_id advance. Raise default timeout 45 s → 90 s as belt-and-braces for printers that neither flip state nor echo subtask_id inside the polling window. Genuinely half-broken sessions (both signals unchanged across the full 90 s) still revert + force-reconnect exactly as before. Transient subtask_id=None during reconnect is not mis-detected as a change. pre_subtask_id=None falls back to state-only checking so the fix is safe for printers that haven't reported a subtask_id yet. New test_scheduler_watchdog.py pins the eight behaviours that matter: pickup via state change; pickup via subtask_id change with state still FINISH (the exact #1078 case); revert when neither signal changes; default timeout is 90 s; pre_subtask_id=None state-only fallback; current subtask_id=None not treated as change; printer disconnect mid-watchdog leaves DB untouched; item that already moved on is not clobbered. |