Skip to main content
Version: 2.0.16

2.0.17

Not yet released.

🐛 Bug Fixes​

  • PHP2Shell: 2.0.16 can leave PHP vulnerable (no version check; shinsenter 8.3/8.2 not rebuilt yet), and phpMyAdmin/Roundcube are not covered #1203: the update pulls every PHP version in use again, and also updates phpMyAdmin and Roundcube to images with the fixed PHP. opencli docker-update now reads the PHP version inside each pulled image and doesn't recreate containers on an image that still has a PHP older than its branch's security release (8.2.35, 8.3.36, 8.4.27, 8.5.12); it reports the run as failed instead, so run it again once the image is rebuilt. Docs
  • docker-update, autostart: give the shared image store its own runroot (shared-store pulls unmount root's containers) opencli#224: pulls into the shared image store used root's runroot, which unmounted the layers of every running root container (caddy, openpanel, mysql, the mail server...) so their next restart failed with executable file '/init' not found. opencli docker-update and opencli docker-autostart now pull into the shared store with its own runroot, and the update remounts the layers of affected containers without stopping them.
  • Roundcube kept its SQLite database (contacts, identities, settings of webmail users) inside the container, so recreating it for a new image lost them. The update moves it to /usr/local/mail/openmail/docker-data/roundcube/db and mounts it from there. The mail server is restarted once for this, mail is unavailable for about a minute.
  • Users' podman stopped responding for some time after opencli docker-update or opencli docker-autostart pulled into the shared image store, because the store stayed readable by root only until all pulls were done. Permissions are now fixed after every pull, and a user's podman socket that gave up meanwhile is started again.
  • podman_ensure_running, used to bring back a stopped caddy, FTP, DNS, mail or OpenPanel container, could remove other containers of the same stack (e.g. ClamAV and phpMyAdmin) and not start them again.
Was this helpful?