Operating systemsm23 supports:
m23 supports Debian, Ubuntu, Kubuntu, Xubuntu, Fedora, Linux Mint, CentOS and openSUSE. The choice of the Linux distribution is a matter of taste and a question of the functional range. m23 gives the administrator the chance to choose the right distribution for each client individually which can be installed automatically. m23 allows to manages all clients via the same m23 administration interface, no matter which distribution is installed.
Free partitioning and formatting
In addition to the automatic partitioning and formatting, m23 supports individual harddrive setups. The resulting partitions can be used for the operating system, to store data and for swapping. The m23 interface includes all necessary functions for deleting, adding and formatting of partitions and for assigning them to mountpoints, installation or swap usage. In addition, it is possible to reuse the existing partition layout for the installation.
Support for software RAIDs
Partitions or entire harddrives can be combined into software RAIDs. m23 supports the RAID levels 0, 1, 4, 5, 6 and 10. These RAID levels have advantages and disadvantages in relation to speed increase and data reliability. RAIDs can be used like normal partitions to install operating systems to or to use them as swap or storage space. You can install the Linux kernel and its modules on a separate (non-RAID) partition to be able to install an operating system on all possible RAID levels.
User management with LDAP
User accounts can be managed with a central LDAP server. This increases the ease of use in environments with many clients and a lot of users. The OpenLDAP server is shipped with the m23 server and can be administered via the highly integrated phpLDAPadmin. m23 can use existing LDAP entries for new m23 clients or add extra entries on adding an m23 client.
NFS for storing home directoriesAn NFS server for storing the user's home directories is a good supplement to an existing LDAP server. Using an NFS server simplifies the making of backups dramatically in contrast to a lot of clients storing the home directories on local harddrives. As a benefit of this combination the user has the alternative to log in into any client and find her or his own desktop and files.
Thousands of software packages like office packages, graphic tools, server applications or games can be installed via the webinterface. Software installation or removal jobs are carried out at boot time or while the client is in normal use.
Graphical desktopsSupport for:
The m23 clients can be installed as workstation with the graphical user interfaces KDE, Gnome, XFce, LXDE and pure X11 or as a server without graphical subsystem. In most server setups, the server doesn't need a user interface because most of the server software runs in text mode. If you choose the Ubuntu distribution you can also install Unity, for Linux Mint Mate or Cinnamon and for Kubuntu "Kubuntu desktop" (KDE based).
Imaging: Installation using image filesAn image file of a partition or entire harddisk can be used to install other clients. These image files are taken from installed clients with all their software packages and settings. The creation of these image files is included in the m23 administration interface for your comfort. Image files can be compressed to save storage space and bandwidth on rollout. An image file can be used to install any number of clients. After the end of the transfer, the client will be adjusted like a client that was installed from software packages. This is done on Debian, Ubuntu, Kubuntu and Xubuntu clients. These "image clients" can be managed like "normal clients" with m23, this includes software installation and other administrative tasks.
Personal data and additional settings can be backupped and restored with the integrated software BackupPC. This is necessary for all personal data which is not handled by m23.
A client can be reinstalled with excactly the same software selection and configuration as it was installed with m23 before. This is done with the recovery feature directly from the m23 interface. The recovery includes the complete installation with partitioning and formatting. All settings done with m23 will be restored. Manual changes need a recovery from a backup medium.
The m23 rescue system is a minimal Linux environment that can be booted on the clients via the network. You can use it for on-location maintenance or via an SSH console. The administrator has full access to the client and can do all administrative work. Additional tools (e.g. the client log) support the administrator with his/her fault analysis.
The free virtualisation solution VirtualBox OSE is now (optional) part of m23 and can be installed directly on the m23 server as well as on the managed m23 clients. Required for virtualisation is a current Debian system (Debian Lenny or Squeeze). Virtual m23 clients which act on the network just like normal m23 clients are managed via the m23 server as usual. These virtual clients can be partitioned and formatted like a normal PC and of course the operating system and additional software can be installed as well.
VirtualBox OSE packages are available for 32 and 64 bit machines, that can be installed directly via the m23 web interface on the m23 clients. These m23 clients are virtualisation hosts for virtual m23 clients after the installation. Virtualisation may be useful to use the existing ressources of your computers better (e.g. office PCs that usually wait for user input most of the time or on a web server).
You can check the usage of harddisk and memory on the virtualisation host in the "VM creation dialog" before you create a new virtual m23 client. So you can take care not to assign too much harddisk space or memory to the VM. The newly created VM runs through the usual 3-step-setup after the creation is done.
Virtual m23 clients can run in graphical mode or without visible output. A modified VNC server was integrated for the graphical output to replace the VRDP feature of the non-free VirtualBox edition. It can be used to access the virtual clients from the boot on and allows to engage if problems occur. The VNC session contains the VirtualBox window with the running virtual m23 client and is based on the lightweight window manager flwm. An analog clock, a terminal and a button to shut down the session together with the VirtualBox machine are visible too.
The m23 interface and the entire m23 system can be improved with extra functions.
The MDK (m23 Development Kit) is a powerful environment to adjust m23 to your needs. E.g. you can create your own m23 server installation CD or build new network boot images with new modules.
Using the m23 extension halfSister, the ambitious administrator can easily make his or her favourite Linux distribution installable with m23, like it was already done with CentOS, openSUSE and Fedora.
How does it work?
m23 differentiates between servers and clients. An m23 server is used for software deployment and the management of the clients. Computers which are administered by the m23 server are the clients.
The client is booted over the network during the installation of the operating system. It is possible to start the client with a bootrom on its network card, with a boot disk or with a boot CD. The client's hardware is detected and set up. The gathered hardware and partition information is sent to the m23 server. Afterwards, this information is shown in the m23 administration interface. Now the administrator has to choose how to partition and format the client. Other settings include e.g. the distribution to be installed on the client.
A GPG key management was integrated into the m23 webinterface, too. After a data fault the restore script will first install an "empty" m23 server and use the backup to bring it into a previously saved state. A short guide showing the restoring steps can be printed out directly from the m23 web interface (and should be before an emergency occurs!).
A variable firewall makes sure that only safe values will be handed to the database. This should make SQL injection attacs impossible.
The IP management allows you to connect network settings to a MAC address, so that the necessary values are sent to the respective device via DHCP. Aside from this direct assignment, IP ranges can be defined, within which the IPs can be assigned dynamically (e.g. to notebooks). Specific settings (e.g. client name, MAC address) or specific IP ranges can now be blocked, so m23 cannot wrongly use these for m23 clients.
Copyright and usage notice for this textYou are allowed to use this text entirely, partly or in a modified form (e.g. for a press article or for product description in CD shops).
This text is licensed under the Creative Commons Attribution 2.0 License. And of course we would be very happy, if we got a copy of your article or a URL where to find it ;-)
Of course, there will again be a talk about m23 at this year's CeBIT. The slightly "misty" talk bearing the title "m23 Softwareverteilung - Auch fÃ¼r die Cloud (m23 software distribution - also for the cloud" will be given on Monday, 10 March 2014, at 3:15 p.m. at the Special Conference Open Source 2014 (Hall 6, Block H01).
A live stream of this (and all the other talks) will be available at the Linux-Magazin web site, and later, the videos will be made available in their video archive. Please don't let this discourage you from being present ;-)
In my talk, I will be reporting on the new m23 version which offers support for Apache CloudStack® serving as a platform for m23, using a customer's project as an example. I'm not going to talk you into buying a "cloud" product, but instead will explain the processes in the background on the source code level. If everything works as planned, I will even show you a live demo of the m23 interface on a little m23 Raspberry Pi server.
Aside from the m23 talk, there will of course also be many other talks ;-) Information about all talks can be obtained at "Special Conference Open Source 2014".
I'd be delighted to also meet you before or after my talk at CeBIT. Please notify me about your wish to meet (e.g. by using the 'message to developer' box at m23.sf.net), so I can coordinate the meet-ups ;-)
Have fun and maybe we'll meet at CeBIT!
The latest m23 release focuses on two main new features: For one, on support for Apache CloudStack® and on the other hand on the extended options to clone machines.
From this version on, m23 supports Apache CloudStack® as a platform for the m23 server and the m23 clients. This way, an m23 server, which can add further instances for the m23 clients, can now be installed on an Apache CloudStack® instance. The virtual clients can be created via the m23 interface and afterwards they can be installed and administered with m23, just like this was already possible with the integrated solution for VirtualBox. The Apache CloudStack® credentials can be entered directly into the m23 web interface. An extended version of the CloudStack® PHP Client is the basis of this new functionality.
As access to cloud systems typically™ is a remote access ;-), m23 now contains the option to install the X2Go server automatically on clients with Debian, Ubuntu and Linux Mint. Using an X2Go client for the local operating system, the virtual client can be accessed graphically. The desktop image of the m23 client is transferred to the X2Go client. Mouseclicks and keyboard strokes are transmitted in the other direction. In contrast to the X11 network protocol, X2Go offers an elaborate compression to be able to transmit the image information fast enough even using small bandwidth network connections.
A lot of time was invested into upgrading the cloning functionality. Now, you can also use Partclone instead of dd as file format for partition images. Contrary to dd, Partclone only saves the used sectors if the filesystem of the partition is supported. This way, the size of the image file is diminished, especially for big partitions and the process is far faster than saving every single sector. Similarly, the recovery of a client is sped up. We also added the automatical adaptation of the file system size to partitions bigger than the original one when using dd. Another development ensures that now also systems which have been integrated into m23 after their installation can be saved into image files.
Aside from these two big 'construction sites' there were, of course, many smaller ones. Many of our users will be delighted that the package search for Debian and its derivatives via the m23 interface has severely been sped up (feels like the factor of 100). During the update of the m23 server via the m23 interface, a new live status now shows the output of the update program. When installing a new m23 server, from this version on half of the free disk space is used as the maximum size of the Squid package cache. The page containing the server settings was completely reworked and extended by a help page. We also corrected client installation status bar, which now should really stop at 100% after the installation is completed ;-). The update interval was prolonged to five seconds (instead of every second as before). The creation of new local users was now separated into a single job for every distribution, which will soon allow to create additional users easier. In the background, the m23 API 2.0 was extended by everything necessary to add and boot a client.
Concerning bugfixes: As most corrections already found their way into the stable branch of m23 via updates, this time there is little to write about. New is the automatical deactivation of Plymouth, as this program led to crashes of the X-server at multiple times. In addition, the m23 administrator accounts can now be deleted, until only a single one remains (before, there wrongly had to remain two).
The new version is available as update via the m23 interface, via APT, as ISO to burn the m23 server installation CD as preinstalled virtual machine and as Raspberry Pi image file (the last three can be found in the download section).
We are always happy to welcome new team members! For example, you could translate m23 into another language, test the software, add new features, prepare new distributions for their installation with m23, check m23 for security and performance or contribute your own ideas to improve m23! Just contact us via the contact form!
The end of the year 2013 is there faster than we would have thought ;-) This is the reason why the version m23 rock 13.3 previously planned for the end of this year will be released early in 2014 - freshly renamed - as m23 rock 14.1.
However, you will not have to wait too long until the release: The functions of the coming version are ready. Now, we only need to do some final testing, update the documentation and the translations and build the different installation media (ISO, VirtualBox appliance, Raspberry Pi image, Debian packages).
See you soon and best wishes for 2013 from the m23 team!
Give m23 your reference!
Send message to developer
fix: group_actions.php: Corrected
reading of group name and group
action to re-enable renaming of
groups. (THX CLUBSKIES)
fix: m23Backup / createBackupList:
Added missing "then" and "fi". (THX
cha: createNewBazaarRepo: Now uses
a hand-crafted .bzrignore file.
cha: prepareOSDorCompression: Now
sure that m23debs.list is deleted.
patch 100740: Release of m23 rock
cha: FDISK_genPartedCommands: Now
fix: linuxrc: Now links
dhclient-script from /bin to
m23 was updated with the m23 update manager 6440 times. All m23 files were downloaded from Sourceforge 25936 times with a total size of 6614.78 GB.