A concurrent trigger writer can adjust Event_Summaries between our
snapshot SELECTs and the per-monitor UPDATE; the UPDATE then overwrites
that adjustment with the older snapshot. Drift is bounded by the
zmstats/zmaudit interval and corrected on the next pass, because the
incremental triggers continue to maintain ES correctly between resyncs.
Locking ES before reading aggregates would invert the canonical lock
order and re-introduce the deadlock cycle the resync rewrite eliminated.
When zmDbDo is called inside a caller-managed transaction (AutoCommit off),
max_attempts is 1 and the loop falls through to Error on a 1213 deadlock —
which is misleading, because the caller (Event::delete, the zmstats
prune+resync TX) has its own outer retry loop that will roll back and
succeed. Downgrade to a Debug message in that path; Error is still emitted
for non-deadlock failures and for autocommit calls that exhaust their
retries.
Inline comments at every occurrence so future readers don't have to look up
the errno. No behavior change.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
selectcol_arrayref returns undef on a DB error. The previous code only
checked truthiness of the result before deciding to DELETE, so a transient
SELECT failure would silently skip the prune for that bucket and let the
transaction continue to commit an incomplete state. Capture \$dbh->err()
after the SELECT and bail out the same way the DELETE error path does.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Previously, if any of the five aggregate SELECTs (Events, Events_Hour/Day/
Week/Month) failed transiently, the per-monitor UPDATE phase still ran and
wrote `// 0` for every column from those failed groups, destroying valid
counters across all monitors.
Track per-aggregate success and build the UPDATE SET clause from only the
column groups whose SELECT succeeded. zmaudit re-runs on its normal
interval, so a missed group is corrected on the next pass instead of being
overwritten with zeros now.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The previous comment block claimed event_update_trigger fires BEFORE UPDATE
and that the lock order was buckets -> Event_Summaries -> Events. Neither
matches the code: triggers.sql defines event_update_trigger as AFTER UPDATE,
and InnoDB X-locks the matched Events row during WHERE evaluation before
either BEFORE or AFTER trigger bodies fire — so the canonical chain is
Events[Id] -> buckets[EventId] -> Event_Summaries[MonitorId]
which is what triggers.sql already documents. Comment-only change.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The previous resync code in zmstats and zmaudit used multi-table
UPDATEs against Event_Summaries that joined the bucket tables:
UPDATE Event_Summaries es
LEFT JOIN (SELECT ... FROM Events_Hour ...) h ON ...
LEFT JOIN (SELECT ... FROM Events_Day ...) d ON ...
... SET es.HourEvents = h.c, ...
zmaudit additionally used scalar correlated subqueries against Events
for the Total/Archived columns and against Events for Storage.DiskSpace.
MariaDB takes S-locks on the joined and sub-queried rows for the
duration of any multi-table UPDATE statement, regardless of isolation
level. event_update_trigger and event_delete_trigger hold X-locks on
those same bucket rows while they walk the trigger body, so the resync
deadlocks against active event lifecycle traffic. Captured a textbook
example in SHOW ENGINE INNODB STATUS:
TX(1) zma: HOLDS X Events_Hour[42229643]
WAITS X Event_Summaries[28]
TX(2) zmstats: HOLDS X Event_Summaries[2,4,5,...,28,...,73]
WAITS S Events_Hour[42229643]
Replace the JOIN/subquery pattern in both scripts with a snapshot phase
followed by per-monitor UPDATEs:
1. SELECT MonitorId, COUNT(*), SUM(DiskSpace) FROM each bucket
(and the equivalent Total/Archived aggregate from Events).
Plain SELECTs do consistent reads and take no row locks.
2. SELECT MonitorId FROM Event_Summaries to widen the universe so
monitors with empty buckets still get zeroed out.
3. For each monitor, UPDATE Event_Summaries SET ... WHERE MonitorId=?.
Each UPDATE only X-locks one ES row and reads no other table.
zmaudit's five separate UPDATEs collapse to one snapshot phase plus
one UPDATE per monitor. Storage DiskSpace gets the same treatment.
zmstats keeps the same outer transaction (BEGIN ... COMMIT, RC
isolation, retry on 1213) so the bucket DELETEs and the resync stay
atomic, but the resync no longer reads the bucket tables under lock.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Three issues in the existing Stats/Event_Data/Frames/Events delete
sequence:
- On any zmDbDo error inside the transaction the code called
dbh->commit() instead of dbh->rollback(). The server-side
transaction was already rolled back when InnoDB picked us as the
deadlock victim, so the commit() was effectively against a fresh
auto-started TX, but the bug pattern leaked through as confusing
state and prevented any retry.
- There was no retry. errno 1213 is expected under contention with
zmstats and zma touching the same Event_Summaries[MonitorId] row,
and the loser is supposed to re-run.
- At REPEATABLE READ, two concurrent filter workers deleting events
with adjacent EventIds take next-key/gap locks on each other's
rows in the bucket tables.
Rewrite the delete block as a retry loop: SET TRANSACTION ISOLATION
LEVEL READ COMMITTED, begin_work, run the four DELETEs, commit on
success. On any error rollback (was: commit). On errno 1213 retry up
to 5 times with backoff. Skip both the isolation switch and the
rollback-then-retry when the caller is managing their own transaction
(in_transaction); they would be the wrong scope to act in.
Falls through to the storage DiskSpace adjustment only on commit, so
a deadlocked delete leaves the event for the next filter pass instead
of orphaning the row with stale storage accounting.
Note: do NOT pre-lock Event_Summaries[MonitorId] FOR UPDATE here, even
though the trigger touches it last. Pre-locking puts ES before
buckets[Id] in the lock acquisition order, which inverts against zma's
event_update_trigger path (Events[A] -> buckets[A] -> ES[N]) and
re-introduces the cycle the rest of this work is removing.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Deadlock detection (errno 1213) is part of normal InnoDB operation
under contention; the engine rolls back the loser and expects the
caller to re-run the statement on a fresh transaction. Most callers
into zmDbDo go through autocommit, where there's no caller-managed
transaction state for a retry to disturb.
When AutoCommit is on, retry the statement up to 5 times with
exponential backoff (~100ms -> ~1.6s, jittered). When AutoCommit is
off, the caller owns the transaction and a unilateral retry would
silently succeed against a TX that no longer reflects the work the
caller staged before this statement; preserve the existing behavior
of logging and returning undef so the caller can rebuild the TX
itself.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
zmDbDo built log messages by s/\?/'%s'/g on the SQL and then passing
the result to sprintf with the bind values. Any literal % in the SQL
(LIKE '%foo%' patterns, or the disk-percent substitution used by
dynamic filters) was interpreted as a sprintf format spec, producing
garbage output or an uncaught sprintf error.
Replace the two-step approach with a single regex that substitutes
bind values directly, so literal % in the SQL is preserved verbatim.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Follow-up to 419846c87 (GHSA-g66m-77fq-79v9). The Device path check was
applied to all monitor Types in three places, but the Device column is
only passed to a shell for Type='Local'. Non-Local monitors (Ffmpeg,
Remote, Libvlc, cURL, VNC) may legitimately hold legacy values such as
an RTSP URL in that column and should not be rejected or warned about.
- scripts/ZoneMinder/lib/ZoneMinder/Monitor.pm: control() dropped the
spurious Warning for non-Local monitors that was flooding zmwatch
logs. The Error/early-return path is preserved for Local.
- web/includes/actions/monitor.php: save action only runs
validDevicePath() when Type=='Local'.
- web/api/app/Model/Monitor.php: replaced the unconditional regex rule
with a validDevicePath() method that checks Type before enforcing
the /dev/ pattern.
Also add client-side validation matching the server rule, so Local
monitors get immediate feedback instead of a round-trip error:
- web/skins/classic/views/monitor.php: HTML5 pattern attribute on the
Device input. Escaped for the v-flag regex engine used by pattern=.
- web/skins/classic/views/js/monitor.js.php: validateForm() now also
rejects Device values that don't match the /dev/ pattern.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Bug 376 (2006) closed all FDs starting from 0 when daemonizing. This
caused FD reuse problems: libx264 writing to a reused stderr FD led to
memory corruption (fixed in child spawn by 66f11435b, but not in the
parent's run()). It also meant children inherited closed FDs 0-2, so
any Perl die/warn output was silently lost, making daemon crashes
impossible to diagnose.
Redirect 0-2 to /dev/null (standard daemon practice) and close only
FDs 3+ for inherited sockets/DB connections. Children now inherit valid
FDs that won't crash or corrupt on write.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Warn() is a natural shorthand that's easy to reach for. Its absence
caused a silent crash in zmfilter when Warn() resolved to Perl's
built-in warn(), which wrote to a closed stderr under zmdc and killed
the process.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Each filter now has its own ExecuteInterval column, making the global
ZM_FILTER_EXECUTE_INTERVAL unused. The DB row will be cleaned up
automatically by zmupdate.pl on the next upgrade.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The sleep delay was overwritten by each filter in the loop, so only the
last filter's timing controlled the sleep. Now tracks the minimum delay
across all filters so no filter oversleeps its ExecuteInterval.
Also removes the unused ZM_FILTER_EXECUTE_INTERVAL reference since each
filter has its own ExecuteInterval, and fixes the overdue warning to
check the unclamped filter_delay and include the filter name.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
The SQL changes every cycle due to zmDiskPercent/zmDiskBlocks/zmSystemLoad
string substitution with live values, so prepare_cached never actually
returns a cached handle. Instead it leaks one cache entry per distinct
substituted value over the process lifetime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PostSQLConditions, HasDiskPercent, HasDiskBlocks, and HasSystemLoad were
set during Sql() term processing but never cleared between rebuilds.
Each Execute() cycle pushed duplicate ExistsInFileSystem terms onto the
PostSQLConditions array, causing it to grow unboundedly over the process
lifetime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Reolink cameras may return a 302 redirect from HTTP to HTTPS.
LWP::UserAgent does not follow redirects for POST requests, so the
login would fail. Detect the redirect, update protocol/host/port from
the Location header, disable SSL verification for self-signed camera
certs, and retry the POST to the HTTPS endpoint. Subsequent API calls
use the updated protocol.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Add a boolean web config option to control whether server statistics
(load, CPU, DB connections, storage, RAM) are rendered in the navbar.
When disabled, the HTML is not output at all, saving the polling overhead.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a boolean web config option to control visibility of the Back and
Refresh navigation buttons shown at the top of most views. Uses a body
class and CSS rule so no individual view files need changes.
Also remove the ZM_WEB_BUTTON_STYLE INSERT from the migration SQL since
zmupdate.pl handles Config table inserts from ConfigData.pm.in.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a web config option to control toolbar button display:
- icons+text (default): show both icon and label
- icons: show only the icon, hide text labels
- text: show only the label, hide icons on buttons that have labels
Body class (btn-icons-only / btn-text-only) is set in getBodyTopHTML() and
CSS rules in skin.css toggle visibility of .text spans and icon elements.
Add title tooltips to console.php buttons so they remain usable in icon-only
mode. Migration appended to zm_update-1.39.4.sql.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Change ZM_OPT_USE_REMEMBER_ME from a boolean to a tri-state string:
- None: checkbox hidden, sessions persist for ZM_COOKIE_LIFETIME (old disabled)
- Yes: checkbox shown and pre-checked by default
- No: checkbox shown and unchecked by default (old enabled behavior)
Update ConfigData.pm.in with new type definition, login.php to honor the
checked state, and session/action handlers to recognize the new values.
Migration in zm_update-1.39.4.sql maps old '1' to 'No' and '0' to 'None'.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a new config option (requires ZM_WEB_NAVBAR_TYPE=left) that controls
whether monitor filter settings are embedded in the left sidebar extruder
(default) or displayed inline at the top of the page. When set to inline,
the sidebar CSS no longer hides the filter elements, insertControlModuleMenu()
skips moving them, and the toggle icons are omitted. A `filter-inline` body
class drives the CSS scoping via :not(.filter-inline) selectors in sidebar.css.
Affects Console, Watch, Montage, and Montage Review views.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The Device field from the Monitors table was interpolated directly into
shell commands (qx(), backticks, exec()) without sanitization, allowing
authenticated users with monitor-edit permissions to execute arbitrary
commands as www-data via the Device Path field.
Defense in depth:
- Input validation: reject Device values not matching /^\/dev\/[\w\/.\-]+$/
at save time in both web UI and REST API
- Output sanitization: use escapeshellarg() in PHP and quote validated
values in Perl at every shell execution point
Affected locations:
- scripts/ZoneMinder/lib/ZoneMinder/Monitor.pm (control, zmcControl)
- scripts/zmpkg.pl.in (system startup)
- web/includes/Monitor.php (zmcControl)
- web/includes/functions.php (zmcStatus, zmcCheck, validDevicePath)
- web/includes/actions/monitor.php (save action)
- web/api/app/Model/Monitor.php (daemonControl, validation rules)
- web/api/app/Controller/MonitorsController.php (daemonStatus)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When ControlAddress does not contain authentication info, fall back
to Monitor->ONVIF_Username/ONVIF_Password, then Monitor->User/Pass.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add ZM_OPT_USE_REMEMBER_ME config option (auth section, requires
ZM_OPT_USE_AUTH) that controls whether a Remember Me checkbox appears
on the login form. When enabled and unchecked, the session cookie
lifetime is set to 0 so the browser discards it on close, logging the
user out. When checked, the session persists for ZM_COOKIE_LIFETIME.
When the option is disabled, behavior is unchanged.
- ConfigData.pm.in: new ZM_OPT_USE_REMEMBER_ME boolean option
- login.php: checkbox between password field and reCAPTCHA/submit
- session.php: use lifetime=0 when remember me is off
- actions/login.php: set/clear ZM_REMEMBER_ME cookie on login, also
update $_COOKIE so zm_session_start sees it in the same request
- auth.php: clear ZM_REMEMBER_ME cookie on logout
- en_gb.php: add RememberMe translation string
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When cameras use HTTPS with self-signed or invalid certificates, LWP's
default SSL verification causes connection failures. Add automatic
fallback to all three HTTP methods (get, put, post) in the base Control
class: if a request fails with an SSL/certificate error, disable
verification and retry. Once disabled, the relaxed setting persists for
the rest of the control session via the ssl_verify_disabled flag.
This follows the pattern already used in Dahua.pm and
TapoC520WS_ONVIF.pm but applies it universally so all control modules
benefit without needing individual SSL handling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Change default training data directory from @ZM_CACHEDIR@/training to
@ZM_CONTENTDIR@/training so it lives alongside events storage instead
of inside the web cache folder. Update PHP fallback to use
dirname(ZM_DIR_EVENTS) accordingly.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Fix path traversal via directory prefix matching: append
DIRECTORY_SEPARATOR to base path before strpos containment check
(2 locations in training.php)
- Fix temp file leak: rename tempnam() base file to .jpg instead of
creating a second file, leaving the original orphaned
- Remove raw_output from detect response (information disclosure)
- Fix double _pushUndo on Delete key (keydown handler + deleteAnnotation)
- Fix help text: "alongside events storage" → "inside the ZoneMinder
cache directory" to match actual ZM_DIR_CACHE default (2 locations)
- Add missing .frame-total span element in event.php
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add ZM_TRAINING_DETECT_SCRIPT config option for specifying the path
to an external detection script (e.g. zm_detect.py). When configured,
a Detect button appears in the annotation editor that:
- Runs the script with -f <image> -m <monitor_id>
- Parses the --SPLIT-- JSON output for labels, boxes, confidences
- Displays results as orange (pending) bounding boxes
- Users can accept (checkmark) or reject (x) each detection
- Only accepted annotations are saved to the training set
Also adds accept/reject UI in the sidebar object list, with orange
color for pending detections and normal colors for accepted ones.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add canvas-based bounding box annotation editor to the event view for
correcting object detection results and building YOLO training datasets.
- Two new config options: ZM_OPT_TRAINING (toggle) and ZM_TRAINING_DATA_DIR
- AJAX backend (training.php) with load/save/delete/status actions
- Canvas annotation editor (training.js) with draw/resize/relabel/undo
- Frame navigation (alarm/snapshot/objdetect/numbered frames)
- Roboflow-compatible YOLO output (images/all/, labels/all/, data.yaml)
- Training data statistics with per-class image counts and guidance
- Full i18n support via en_gb.php SLANG/OLANG entries
- Label validation, YOLO coordinate clamping, audit logging
- DB migration for existing installs (zm_update-1.39.2.sql)
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The error path was missing the /g flag on the regex replacing ? with %s,
so only the first placeholder was replaced. It also dumped bind values as
a space-separated string instead of using sprintf to substitute them into
the query. Now matches the existing debug path pattern.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
