​systemd 227 发布,Linux 的 init 程序

jopen 9年前

systemd是一个init程序。根据之前 Upstart的一些介绍大家应该简要了解了传统Sysvinit系统的不足之处。不过,这次的systemd的设计思想更为超前,也可以说是借鉴了OSX上launchd的思路。

systemd 227 发布,此版本更新内容如下:

        * systemd now depends on util-linux v2.27

        * libmount mandates /etc/mtab not to be regular file, and
          systemd now enforces this condition at early boot.

        * Support for the "pids" cgroup controller has been added.

        * Support for the "net_cls" cgroup controller has been added.

          This adds a new config directive called NetClass= to CGroup
          enabled units.

        * 'systemctl is-system-running' now returns 'offline' if the
          system is not booted with systemd. 

       Watchdog timeouts have been increased to 3 minutes for all
          in-tree service files.

       
'machine-id-commit' functionality has been merged into
          'machine-id-setup --commit'.

        The WorkingDirectory= directive in unit files may now be set
          to the special value '~'.

       
"machinectl shell" will now open the shell in the home
          directory of the selected user by default.

        The CrashChVT= configuration file setting is renamed to
          CrashChangeVT=, following our usual logic of not
          abbreviating unnecessarily.

       
The PrivateTmp=, PrivateDevices=, PrivateNetwork=,
          NoNewPrivileges=, TTYPath=, WorkingDirectory= and
          RootDirectory= properties can now be set for transient
          units.

        The systemd-analyze tool gained a new "set-log-target" verb
          to change the logging target the system manager logs to
          dynamically during runtime.

       
In nspawn /sys is now mounted as tmpfs, with only a selected
          set of subdirectories mounted in from the real sysfs.

        Support for USB FunctionFS activation has been added.

       
The "systemctl exit" command now optionally takes an
          additional parameter that sets the exit code to return from
          the systemd manager when exiting.

        sd-bus gained the new API calls sd_bus_path_encode_many()
          and sd_bus_path_decode_many() that allow easy encoding and
          decoding of multiple identifier strings inside a D-Bus
          object path.

       
systemd-cgtop gained support for a -M/--machine= switch to
          show the control groups within a certain container only.

        "systemctl kill" gained support for an optional --fail
          switch.

       
A new systemd.crash_reboot=1 kernel command line option has
          been added that triggers a reboot after crashing.

        The RuntimeDirectory= setting now understands unit
          specifiers like %i or %f.

       
A new (still internal) libary API sd-ipv4acd has been added,
          that implements address conflict detection for IPv4.

        File descriptors passed during socket activation may now be
          named. A new API sd_listen_fds_with_names() is added to
          access the names. 

       
systemd-networkd gained support for:

            - Setting the IPv6 Router Advertisment settings via
              IPv6AcceptRouterAdvertisements= in .network files.

            - Configuring the HelloTimeSec=, MaxAgeSec= and
              ForwardDelaySec= bridge parameters in .netdev files.

            - Configuring PreferredSource= for static routes in
              .network files.

        The "ask-password" framework used to query for LUKS harddisk
          passwords or SSL passwords during boot gained support for
          caching passwords in the kernel keyring, if it is
          available.

       
When downloading tar or raw images using "machinectl
          pull-tar" or "machinectl pull-raw", a matching ".nspawn"
          file is now also downloaded, if it is available and stored
          next to the image file.

        Units of type ".socket" gained a new boolean setting
          Writable= which is only useful in conjunction with
          ListenSpecial=. If true, enables opening the specified
          special file in O_RDWR mode rather than O_RDONLY mode.

       
systemd-rfkill has been reworked to become a singleton
          service that is activated through /dev/rfkill on each rfkill
          state change and saves the settings to disk.

        The journal daemon gained support for vacuuming old journal
          files controlled by the number of files that shall remain,
          in addition to the already existing control by size and by
          date.

       
udev will now create /dev/disk/by-path links for ATA devices
          on kernels where that is supported.

        * Galician, Serbian, Turkish and Korean translations were added.

更多详细改进内容请看发行说明

来自:http://www.oschina.net/news/66945/systemd-v227