Fixes
- RustDesk 1.5.0 clients stored every audit record three times. Client 1.5.0 retries an audit post unless the server answers with an empty body, and CortenDesk answered
{}. Connections, file transfers and alarms were each stored three times, and the extra connection rows showed as active sessions that never ended. The endpoints now answer as the client expects and ignore a repeated record. To remove copies already stored, runphp artisan cortendesk:dedupe-auditsto see what it finds, then add--force(in Docker:docker compose exec cortendesk php artisan ...) (#88). - File transfers by copy and paste showed an empty path. They now show the file names with a "Clipboard" badge, and search finds them (#88).
- Larger fleets: the container ran PHP with 5 workers and nginx with 1,024 connections, which took the console and client API down at around 700 devices. The defaults are now 24 workers and 4,096 connections, the container raises its open-file limit, and both are adjustable. See "Sizing for larger fleets" in the README (#93).
- PHP-FPM inside the container now only listens on localhost.
- Alarm type 10 from RustDesk 1.5.0 shows as "ID whitelist block".
New
- Front end on stock Bootstrap 5.3.8. The commercial theme is gone, replaced by our own layout CSS and JavaScript. Everything shipped is open source now (see
NOTICE), the image is about 35 MB smaller, and the console looks the same. - Notification retries. A notification that times out or gets a temporary error from Apprise (429, 424 or 5xx) is retried about 1 and 5 minutes later. Scheduled alerts wait up to 15 seconds (
CORTENDESK_NOTIFY_TIMEOUT). An offline alert is never re-sent once the device is back (#87). - Web client remembers your settings. Image quality, custom quality, FPS, codec, scaling, cursor options, clipboard and audio carry over to the next session, with "Reset settings to defaults" in the menu. Fullscreen moved next to the monitor selector (#92).
- New container settings:
CORTENDESK_HTTP_PORT(listen port),CORTENDESK_UPDATE_CHECK=false(no update check),CORTENDESK_LOG_LEVEL(one level for the console and hbbs/hbbr),CORTENDESK_FPM_MAX_CHILDREN,CORTENDESK_NGINX_WORKER_CONNECTIONS(#89, #93).
Notes
- One migration runs on start (retry columns for notification deliveries).
- If you put a reverse proxy in front for a large fleet, size its connection limits too.
docker pull marcpope/cortendesk:1.11.0
docker pull ghcr.io/marcpope/cortendesk:1.11.0