systemd (256~rc3-3) unstable; urgency=medium - /tmp/ is now by default a tmpfs, via the tmp.mount unit provided upstream. The old default setup can be retained simply by masking the unit locally with (do not do this if you are defining /tmp/ manually in /etc/fstab): systemctl mask tmp.mount or: touch /etc/systemd/system/tmp.mount It is recommended to check /tmp/ for any leftover files before rebooting after installing this update and manually cleaning up, as the directory will longer be cleaned up automatically on boot, as it is overmounted. It is always possible to remount the root filesystem in a local directory and remove leftovers manually after rebooting, but this will not be done automatically to avoid unintential removals. This situation can be easily detected by checking the journal after a reboot, as there will be a log message that says: tmp.mount: Directory /tmp to mount over is not empty, mounting anyway - /run/lock/ is no longer created with a patch before units start, but by a standard early-boot run-lock.mount unit that is ordered before local-fs.target. Any service needing to use /run/lock/ and running before sysinit.target (ie, they likely define DefaultDependencies=no) will need to be explicitly ordered with After=run-lock.mount. The two known cases where this happens in the archive have a bug+MR filed already. - On new installations, tmpfiles.d will now cleanup by default files that have not been changed or accessed on /tmp/ for 10 days, and /var/tmp/ for 30 days. The legacy behaviour can be configured with a local override if needed: echo 'D /tmp 1777' > /etc/tmpfiles.d/tmp.conf This override will be automatically provided for upgrades of existing systems from previous releases to Trixie. As a reminder, individual files and directories can be marked for exclusion from cleanups with the 'x' type configuration line as described in the tmpfiles.d manpage, for example: echo 'x /tmp/my-precious' > /etc/tmpfiles.d/precious.conf - coredumps are now disabled by default via configuration files rather than an out-of-tree patch (installing the optional systemd-coredump package will enable them as before). As always, overriding via local drop-ins is possible if desired. The configuration files that respectively affect the system systemd instance, the user systemd instances and PAM sessions are: /usr/lib/systemd/system.conf.d/10-coredump-debian.conf /usr/lib/systemd/user.conf.d/10-coredump-debian.conf /usr/lib/sysctl.d/10-coredump-debian.conf /etc/security/limits.d/10-coredump-debian.conf -- Luca Boccassi Tue, 28 May 2024 00:07:57 +0100 systemd (256~rc2-1) unstable; urgency=medium In the rare case a scheduled shutdown fails to be enqueued (most likely, D-Bus daemon/broker is not installed), the system will now immediately reboot, restoring the default behaviour intended upstream. -- Luca Boccassi Wed, 15 May 2024 00:40:56 +0100 systemd (253~rc2-1) experimental; urgency=medium TuxOnIce is no longer supported directly (via an out-of-tree patch) for hibernation. TuxOnIce users should instead use the environment variable $SYSTEMD_BYPASS_HIBERNATION_MEMORY_CHECK=1 (i.e.: set it on the kernel command line). The Debian-specific /etc/default/locale will now be made into a symlink to the distribution-agnostic /etc/locale.conf. The content format is compatible, so at upgrade time if /etc/default/locale exists it will be copied to /etc/locale.conf, if the latter does not exist. The Debian-specific /etc/timezone will no longer be supported by systemd-timedated, as it is no longer useful and its functionality has been subsumed into /etc/localtime, and tzdata will no longer create it. The journal 'audit' feature is no longer patched to be disabled by default, and instead we rely on the fact that the unit 'systemd-journald-audit.socket' is disabled by default. If you enabled this in the past by setting 'Audit=yes' in a configuration file, you will also need to enable the socket unit. -- Luca Boccassi Wed, 08 Feb 2023 20:36:58 +0000 systemd (251.3-2) unstable; urgency=medium systemd-resolved has been split into a separate package. This new systemd-resolved package will not be installed automatically on upgrades. If you are using systemd-resolved, please install this new package manually. -- Luca Boccassi Thu, 05 Aug 2022 20:26:12 +0100 systemd (251.2-3) unstable; urgency=medium systemd-boot has been split into a separate package. This new systemd-boot package will not be installed automatically on upgrades. If you are using systemd-boot, please install this new package manually. The default boot loader in Debian is grub2. If you have not set up systemd-boot manually, no action is required on your side. -- Michael Biebl Wed, 08 Jun 2022 21:49:47 +0200 systemd (251.1-1) unstable; urgency=medium systemd-journal-gatewayd and systemd-journal-remote are now built without the --trust option, in order to be able to switch away from gnutls to openssl. -- Luca Boccassi Thu, 26 May 2022 00:55:39 +0100 systemd (247.2-2) unstable; urgency=medium systemd now defaults to the "unified" cgroup hierarchy (i.e. cgroupv2). This change reflects the fact that cgroupsv2 support has matured substantially in both systemd and in the kernel. All major container tools nowadays should support cgroupv2. If you run into problems with cgroupv2, you can switch back to the previous, hybrid setup by adding "systemd.unified_cgroup_hierarchy=false" to the kernel command line. You can read more about the benefits of cgroupv2 at https://www.kernel.org/doc/html/latest/admin-guide/cgroup-v2.html -- Michael Biebl Mon, 21 Dec 2020 18:40:10 +0100 systemd (247.2-1) unstable; urgency=medium KERNEL API INCOMPATIBILITY: Linux 4.14 introduced two new uevents "bind" and "unbind" to the Linux device model. When this kernel change was made, systemd-udevd was only minimally updated to handle and propagate these new event types. The introduction of these new uevents (which are typically generated for USB devices and devices needing a firmware upload before being functional) resulted in a number of issues which we so far didn't address. We hoped the kernel maintainers would themselves address these issues in some form, but that did not happen. To handle them properly, many (if not most) udev rules files shipped in various packages need updating, and so do many programs that monitor or enumerate devices with libudev or sd-device, or otherwise process uevents. Please note that this incompatibility is not fault of systemd or udev, but caused by an incompatible kernel change that happened back in Linux 4.14, but is becoming more and more visible as the new uevents are generated by more kernel drivers. To learn more about the required udev rules changes please check the "CHANGES WITH 247" section of /usr/share/doc/systemd/NEWS.gz. -- Balint Reczey Fri, 11 Dec 2020 18:22:42 +0100 systemd (241-4) unstable; urgency=medium DRM render nodes (/dev/dri/renderD*) are now owned by group "render" (previously group "video"). Dynamic ACLs via the "uaccess" udev tag are still applied, so in the common case things should just continue to work. If you rely on static permissions to access those devices, you need to update group memberships accordingly to use group "render" now. -- Michael Biebl Fri, 17 May 2019 19:15:32 +0200 systemd (236-1) unstable; urgency=medium DynamicUser=yes has been enabled for systemd-journal-upload.service and systemd-journal-gatewayd.service. This means we no longer need to statically allocate a systemd-journal-upload and systemd-journal-gateway user and you can now safely remove those system users along with their associated groups. -- Michael Biebl Sun, 17 Dec 2017 21:17:32 +0100 systemd (231-1) unstable; urgency=low This version drops support for running /etc/rcS.d SysV init scripts. These are prone to cause dependency loops, and almost all Debian packages with rcS scripts now ship a native systemd service. If you have custom or third-party rcS scripts you need to convert them or change them to run in rc2.d/ - rc5.d/; see this page for details: . -- Martin Pitt Thu, 14 Jul 2016 12:54:34 +0200 systemd (224-2) unstable; urgency=medium This version splits out systemd-nspawn, systemd-machined, and machinectl into the new "systemd-container" package. That now also enables systemd-importd. -- Martin Pitt Sat, 22 Aug 2015 15:58:43 +0200 systemd (220-7) unstable; urgency=medium The mechanism for providing stable network interface names changed. Previously they were kept in /etc/udev/rules.d/70-persistent-net.rules which mapped device MAC addresses to the (arbitrary) name they got when they first appeared (i. e. mostly at the time of installation). As this had several problems and is not supported any more, this is deprecated in favor of the "net.ifnames" mechanism. With this most of your network interfaces will get location-based names. If you have ifupdown, firewall, or other configuration that relies on the old names, you need to update these by Debian 10/Ubuntu 18.04 LTS, and then remove /etc/udev/rules.d/70-persistent-net.rules. Please see /usr/share/doc/udev/README.Debian.gz for details about this. -- Martin Pitt Mon, 15 Jun 2015 15:30:29 +0200