This article is contributed by Kamila Součkova

Introduction

As the btrfs wiki says: “Btrfs is a new copy on write filesystem for Linux aimed at implementing advanced features while focusing on fault tolerance, repair and easy administration.” Although under heavy development, it has become stable enough for personal use, and there are plenty of reasons to try it. What distinguishes it from earlier filesystems is that it has been designed with scalability and robustness in mind: it can handle huge files (up to 16EiB — a lot!), it can pack lots of files and directories efficiently, has built-in error detection methods (checksums of data and metadata), support for transparent compression, integrated multiple devices support (RAID-0, RAID-1 and RAID-10 so far) and more — see here for a more complete list.

In this how-to I will focus on one particularly neat feature: snapshots. Btrfs allows you to make read-only or writable snapshots of the state of your filesystem without wasting space with redundant data. Together with YaST’s Snapper module, this makes tracking FS changes and undoing undesired modifications a breeze.

Some Background

So how does btrfs do all that snapshot awesomeness? The answer is, it uses subvolumes. Subvolumes are something like separate filesystems within one partition: you can have several filesystem roots in the same container. With copy-on-write (COW) this means that “almost-separate filesystems” can share duplicate data and therefore save space. (Note: Subvolumes do not have a fixed size, they grow and shrink with the data they contain.)

With snapshots, one creates a copy of the whole filesystem within a new subvolume on the partition. As data is actually duplicated only once it is changed, making snapshots is both time- and space-efficient.

Time to Test It

Preparations

You will need the btrfs tools (package btrfsprogs). The Snapper YaST module provides a nice interface for btrfs’s snapshot functionality. You may create a new btrfs partition just for experimenting, or convert an existing ext3/ext4 filesystem. As for conversion, this is done using btrfs-convert; the good thing is that it can preserve the original ext FS and one can easily go back to it (see the wiki page for more info). Alternatively, openSUSE offers btrfs as an option during installation since 11.3.

Listing Snapshots

Once you have a partition, you can look at its snapshots with
snapper list
or view plain btrfs submodules/snapshots with
btrfs subvolume list

Note that in the snapper listing snapshot #0 is always the current system. Also note that for btrfs snapshots and submodules are the same thing.

Snapper Configs

Snapper keeps configs for your partitions; you can view them with
snapper list-configs

Configs allow you to specify when to make snapshots, how many of each tipe should be kept etc. They are stored in /etc/snapper/configs; config templates are in /etc/snapper/config-templates.

In case your root filesystem is btrfs, Snapper will have created a config that makes a “pre” and “post” snapshot pair whenever you use YaST or zypper in addition to the “timeline” snapshots.

You can create your own configs (e.g. for /home or whatever) with snapper create-config:
snapper -c create-config [-t template]
You need a config whenever you want a path other than the default /. (Always specify it with -c.)

Making Snapshots

To make snapshots manually use either
snapper [-c config-name] create --description "something that tells you what this is"
(and see snapper –help for more options); or the vanilla

btrfs subvolume snapshot /[subvolume-name]

A snapshot made with the former command will show in snapper list; one made with the latter will appear as a directory named in. Side note: Snapper actually stores its snapshots in/.snapshots/.

Viewing Differences

To see the differences between snapshots you may either mount the snapshots somewhere (with snapper mount , or simply browse to the snapshot location) and see for yourself, or use the very convenient snapper diff:
snapper diff
will give you the list of files changed between the two given snapshots;
snapper diff
will show you the files’ diff.

Restoring Previous Snapshots

To restore a snapshot made with snapper use
snapper undochange []...
You can also specify ranges, as in snapper undochange 42..47.

To do the same with vanilla btrfs use
btrfs subvolume set-default
where can be found out with btrfs subvolume list and then unmount and remount.

Snapper GUI

Snapper also provides a YaST GUI for comparing and restoring snapshots or individual files.

Final Words
Btrfs does not have an error-correcting fsck yet, so just now it is not ready for systems that require high reliability. But that doesn’t mean you shouldn’t use it ­— it merely means that you should keep backups, which you should do anyway. So go ahead and give it a try! Its snapshots feature (together with all the other awesomeness) makes it a really worthy filesystem for both personal use and servers, and with Snapper managing snapshots is really convenient.
Note: Kernel 3.1
Fixes from kernel 3.2 have not been backported into 3.1 yet, so you may want to use the 3.2 kernel. See e.g. https://bugzilla.novell.com/show_bug.cgi?id=733843 in case you get crashes with 3.1.

References and Further Reading

BTRFS wiki

Linux.com Tuturial

openSUSE Snapper Portal

Booting your system and services

In openSUSE 12.1, systemd starts up and supervises openSUSE operating system and its services. This means a number of steps usually including the following:

• mounting your disks (checking for errors if needed),
• starting the firewall and other security tools,
• starting services like mail, databases or web servers,
• connecting the network and networked filesystems,
• starting the graphical system and the login manager which brings you to your desktop

When up and running, systemd keeps tabs on the services and starts new ones (like printing) on demand, when they are needed.

Ye olde days (SysV init)

systemd replaces the old SysV init system, which is basically a loose collection of shell scripts in /etc/init.d (sometimes /etc/rc.d in other distributions). Those scripts would be called depending on the “runlevel” of the system, based on symbolic links to the scripts themselves in the /etc/init.d/rcX.d directories. Each directory there represented a runlevel number, with their order and interdependencies managed through the naming of symbolic links to those scripts, using numbers from 00 to 99. Upon booting, the system would go to a designated runlevel, starting all the scripts associated to that runlevel. While a bit brittle, this system was reasonably clear and understandable, although the scripts were usually quite big and complicated.

New and improved: systemd

Instead of shell scripts, systemd uses “.service” files that define various features provided by systemd, a bit like the old SysV init scripts but with an easier and more declarative syntax than bash.

Dependencies and events

Instead of starting the scripts based on a predefined number of runlevels, systemd uses dependencies and events to decide when to start a service. For example, once a network device is initialized, systemd starts the network service. Once a connection is made, network based services like NFS (Networked File Storage) and Samba (file sharing) can go ahead and start up. If there is no network, systemd won’t bother starting them. And without a printer attached to the system, printer service CUPS isn’t needed either.

systemd uses sockets and dbus for starting services and in this way allows on-demand starting of daemon processes, watching the sockets for requests from applications. For example, only when you try to print will the print service be started.

Starting and stopping services

systemd can start as many services as needed at the same time, respecting the dependencies and needs of the system. This, combined with the fact that it does not need to start a shell interpreter for each individual script, results in faster boot times.

Services in systemd can be disabled, enabled, started, stopped, restarted and reloaded just as the scripts in SysV but unlike SysV, the applications controlled by the script are closely monitored and controlled. Where SysV would not notice services going down (due to crashes for example) systemd will attempt to re-start them. And when systemd stops a process, you can be certain it has been stopped (thanks to using the cgroup kernel feature), whereas in SysV, processes sometimes managed to wrestle themselves out of control of the scripts.

Standardization

systemd strives towards standardizing init scripts which used to be different for each distributions. This should reduce the maintenance burden for distributions as they will share more work but it of course requires them to adapt to a common standard first. This means using the standard “systemctl” for controlling services instead of the /etc/rc* files. Currently, openSUSE still supports the rc-style files (providing symbolic links in /etc/rc.d ) so you can still run services by hand. systemd is also backwards compatible with the old-style init scripts which means that “non” standard (not included in systemd) scripts will keep working. For more details and a list of limitations in that regard please refer to here.

Using systemd

Let’s go over some features of systemd which are relevant for ‘normal’ users as well as system administrators.

Managing services and (auto)mounting devices

Devices marked via udev rules are marked as units in systemd and their udev properties can be used as configuration sources to set dependencies for device units. systemd also maintains mount and automount points and (un)mounts devices where needed. For every automount point we have a matching mount point. To let systemd automount a device, add “comment=systemd.automount” to an fstab line option, and the mountpoint is automatically handled as a automount point. This can be used for networked filesystems nfs and cifs. /etc/fstab is used as an extra configuration for these mount points. systemd also supports target units which are logically grouped and in turn these groups can be controlled together.

Stability and emergency features

Systemd automatically computes dependencies between services when running and is able to break dependencies cycles that is when a loop between services happen, systemd will break it once detected, your system will continue to work without user interaction. It also supports creating a snapshot and later restoring the system state and can be used in cases like emergency shell and provide an easy way of suspending services. However this feature is not persistent, and it is only kept in memory and is flushed after a reboot.

New System tools

A system adminstrator will welcome the new tools like systemctl to administrate the system. For example to check the status of the mysql.service, run:
$ systemctl status mysql.service
mysql.service - LSB: Start the MySQL database server
Loaded: loaded (/etc/init.d/mysql)
Active: active (running) since Wed, 02 Nov 2011 10:53:34 +0100; 24h ago
Process: 3242 ExecStart=/etc/init.d/mysql start (code=exited, status=0/SUCCESS)
CGroup: name=systemd:/system/mysql.service
├ 3578 /bin/sh /usr/bin/mysqld_safe --mysqld=mysqld --user=mysql --pid-file=/var/run/mysql/mysqld.pid --socket=/var/ru...
└ 4191 /usr/sbin/mysqld --basedir=/usr --datadir=/var/lib/mysql --plugin-dir=/usr/lib64/mysql/plugin --user=mysql --lo...

What is really cool with systemctl is its tab completion. on first level for command like systemctl will show you wich action you are able to run. On second level it will show which service name you can use like systemctl start my will complete to mysql.service

Disable/Enable a service

 systemctl disable mysql.service
 systemctl enable mysql.service

Note that chkconfig is systemd aware and will show you when a service is systemd native or not.

Locations
systemd service files are mainly stored in /lib/systemd/system

Start stop your computer
 systemctl reboot
 systemctl halt
 systemctl poweroff
 systemctl emergency
 systemctl rescue

Get informed

 systemd-analyse --time: give your boot time of your system
 systemd-analyse --blame: find which service took too much time to boot
 systemd-analyse --plot > bootchart.svg : see your boot, in a graphical chart
 systemctl dot : Use a command line like systemctl dot --order | dot -Tsvg > systemd-order.svg to generate a graphical dependency tree

Debug
You can activate a debug mode on startup by adding on the boot line:

 systemd.log_target=kmsg systemd.log_level=debug

Moving backwards

While we do not recommend anyone using sys V init on 12.1, people who wish to choose it as the default can do so by installing sysvinit-init package (it will uninstall systemd-sysvinit package) or changing their grub options. Even if you don’t use systemd, don’t try to uninstall its package, it might break your system. You don’t need to remove systemd-sysvinit in case of emergency, if you let installed the sysvinit package you can boot with traditonnal init by adding on your boot line init=/sbin/sysvinit et voilà !

openSUSE 12.1 has been released a few weeks ago. A major new technology we introduce in this release is ownCloud, which we ship in a separate repository. ownCloud is a web application which lets you set up your own cloud – a place for you data where you can share it with others or use it over multiple devices. As YOU will own the data, it’s great from a privacy and security point of view.

However, setting up ownCloud, while not particularly complicated, is still vastly more difficult than navigating to a website which offers you convenient ways of giving them your personal data. If the convenience offered by companies like Dropbox, Canonical or Facebook is so much greater than what is offered by technologies which protect your freedom, you don’t really have a choice as common user.

openSUSE 12.1 offers a solution: mirall. While this tool has not yet solved all problems in the world, it makes deploying your ownCloud as easy as a few clicks and makes your files available for you off-line (a feature ownCloud itself lacks). Read on to learn what mirall has in store for openSUSE users!

Cloudy times

Today computer work is not limited to a single machine any more. People have the urgent need to share their data over various computers, ie. desktop computers, laptops and such.

For that the buzzword “cloud” is a promising solution. In the first step, the cloud offers not much more than storage that is accessible from everywhere, but nowadays the cloud also offers content and functionality. Lots of enterprises have jumped on the cloud and fly through the sky towards a billion dollar business… The idea is that users push their data into the cloud and feel the spirit of freedom from cellphone, laptop, computer, gps device, or any other device.

Privacy

Who owns the data that is pushed into the cloud? The user who pushed it of course.

But where is the data? Well, you can’t answer that question, because the only thing you know is an url where you pushed your data. Where on the planet the physical computer is the domain is pointing to, you have no idea. And even if you had its worthless because you still don’t have access to it. You don’t know where your data was mirrored to, you don’t know who has access to the machine, you can’t be really sure that your data is not investigated, used for whatever tests or analysis.

But wait, isn’t the data a very valuable thing for you? So you’d better control where its going, right? You do not want to push it to a server of a company from abroad where not even your countries laws work. And, even if you decide to delete your data, how can you be sure that the data is really going to be deleted? Maybe its just removed from the view you have on it… And who guarantees your data is not sold to the highest bidder?

You need a cloud that is controlled by you.

openSUSE and ownCloud

The freedom of software and data is very valuable to the openSUSE Project and we would like to help you escape the deceptive arms of those who offer you some convenience in exchange for control over your data. A first step was providing spideroak in openSUSE 11.4 which, unlike most competitors, encrypts your files and thus offers more protection for your privacy. But your data is still ‘somewhere else’ and we prefer to offer something you would really own.

ownCloud serving music

Fortunately there is a very appealing solution for that called ownCloud. This project, started two years ago by Frank Karlitschek as a part of the KDE family (but in no way limited to KDE users), offers a very up-to-date, slim but extendable and standard based cloud web app. Written in PHP and without difficult dependencies it installs on nearly every web space, even the cheapest one.

Just a week ago, Frank announced the start of ownCloud Inc. which “will help [the ownCloud team] to spread ownCloud and free cloud services in general – way more than we could have done without”. This will bring additional resources to ownCloud development, making it the most promising Free Software solution for protecting your Freedom in the cloud!

Vision of Integration between openSUSE and ownCloud

We want to integrate the ownCloud as good as we can into the openSUSE desktop to get the most out of it for the user.

First, it should be very easy to install an ownCloud. Webapps can be ugly to install for not so experienced users, but should you bother? The idea is to have an installation on either your local machine, a machine in your local network behind a firewall or on your domain just a few clicks away. If you want to remove it again, the de-installation should again be dead easy. This includes of course the wiping of the data which was on the cloud.

ownCloud and GNOME Calendar

Second, you might want to move your cloud. Lets say you worked with a local installation and like your ownCloud. Now you want to push it to the internet. That also should be just a few clicks, including all the data you already maintain in your ownCloud.

Furthermore, ownCloud will integrate into desktop applications where useful, for example as storage for contacts and calendar data, simply as file share via webDAV or as music storage for Amarok, as central bookmark server or notes platform.

All data you work with in the desktop apps should also be accessible via the ownCloud web interface in case you travel.

Mirall

As a starting point of integration openSUSE has a tool called mirall. It was started by Duncan Mac Vicar and extended by Klaas Freitag in a SUSE hackweeks to implement the described vision. The first and foremost target with mirall is to provide a very easy to use tool for un-experienced users which hides the complex client/server issues from its users completely.

Mirall starts in the system tray on your desktop and if you click on it, it checks if your system is already connected to an ownCloud. If not, it offers you to either connect to an existing ownCloud by providing a URL and credentials or if you want to install a new ownCloud.

For a new installation mirall gives you the choice between a local installation on your machine or an installation on a ftp accessible web space which you can rent from every other ISP for very little money. mirall will download and install the latest ownCloud on that webspace and automatically create a configuration for you.

With a connected ownCloud to your desktop, it is now possible to configure other express upload directories through mirall to mirror your data to the ownCloud. Moreover, mirall allows to fetch the data from the ownCloud to the directory on your local machine. For example, if you have been at work and changed some data on your cloud, you can fetch the directory on the machine at home to get the last data from work to home. This is a manual action due to some limitations in ownCloud, see the next chapter for more information about this.

These functions are working with mirall today. More functionality will be implemented following the described vision and based on user feedback. Of course, patches are welcome! You can contact Klaas directly or talk to the ownCloud team.

Limitations of ownCloud

ownCloud currently offers webDAV as file interface solution. As webDAV is a well known and commonly implemented standard it works on a large variety of devices and operating systems. Users will be able to access their files from pretty much anywhere, including most mobile phones, Windows, Mac OS, Linux, BSD and more. However, webDAV does depend on a network connection and leaves no trace of the files when working off line.

mirall offers a partial solution to this, offering the files for off-line use and providing basic syncing. But this syncing does NOT take place on the file level! When the online ownCloud offers newer files than the local one, mirall will notify the user but only offer the choice of replacing the local files with the ones from ownCloud or the other way around. It does not know which files have changed. This is due to the limitations of the webDAV protocol. We do expect ownCloud to work around this and provide more advanced syncing possibilities in the future but it is a complicated problem and there is no solution yet. If you happen to be interested in syncing algorithms and want to help us find a creative way around this limitation, get in contact!

How mirall works around this limitation

Mirall creates so called Express Upload Directories for its user.

What are these doing? They connect a local directory with a directory on the ownCloud. Whatever changes you do in the local directory, either connected to a network or not, is going to be mirrored to the ownCloud right away or as soon as you go online again. No matter if change, add or delete files or directories, the changes go into the cloud. The local directory is the master.

What happens if changes are made to one of your Epress Upload Directories on the ownCloud? Well, the local directory is the master. That means, the changes to files on the cloud are going to be overwritten with the next syncing mirall performs.

If new files are added on the ownCloud, mirall disables the Express Upload Folder for you in order to not delete files on the cloud, because, remember, local is the master, and if remote are more files than local, following the logic, they need to be deleted. However, mirall tries to avoid that and disables the folder and you have the change to fetch from the cloud, that is, get the contents of the remote folder to your local system.

This means you have to be a bit careful with using mirall on several computers at once and it is currently not suitable for working collaboratively. Changes could be overwritten or lost if there are changes in both the cloud and your local system!

Taking this into consideration, mirall already solves a lot of the “where are my files” problems people have when they move from university→home→boyfriend and back.

Getting ownCloud and mirall

After all this reading you might want to know how to get mirall and start using it. Well, this is as easy as the proverbial pie. Installing mirall only requires you to click the one-click-install on the right and (after agreeing with the installation by giving your administrator password) starting it from the menu…

For users of openSUSE 11.4 and older, you can find more (and up-to-date) instructions on getting ownCloud and mirall installed on the openSUSE wiki.

Getting your ownCloud

Once you have mirall up and running, click on the tray icon. If you have no ownCloud configured yet, a dialog wizard will show up and offer you two options:

1. connect my ownCloud
2. create a new ownCloud

If you pick the first option, connect my ownCloud you have to enter the valid ownCloud url into the text field and hit Next. The next dialog asks you for username and password to connect to the ownCloud. After hitting on Next again, a summary page displays the result of the operation.

ownCloud Installation

If you choose to create a new ownCloud on the first dialog page, the next wizard page asks you if you want to create it on the local computer or on an internet domain you control. In both cases Mirall downloads latest stable sources of ownCloud from the ownCloud project page, extracts it on the local machine and copies it to the target system.

If you go for the installation on the local machine, Mirall will ask you for the root password to be able to copy files into the webservers application directory.

If you choose to install on your ftp accessible site, enter a domain name into the text field. On the next wizard page, a FTP-Url and credentials for FTP need to be provided.

Mirall uses FTP to access the webspace and creates a directory called ownCloud on the ftp site and copies the files there.

On the next wizard page, you need to provide credentials for the admin user of your new ownCloud instance. During installation the user is created automatically and after the installation finished you can use the credentials to log in.

mirall now downloads the latest stable release from the official ownCloud page and installs it on the target machine. It creates an auto config setup for ownCloud so that you do not have to do the initial setup step on first login, as you already provided the data through mirall. Your ownCloud is configured completely with that, and the ownCloud gets connected to your machine as described above.

After a successful installation, Mirall tries to create an express upload folder in your home directory named Mirall. It is automatically mirrored to the cloud. This step is skipped for security reasons if the folder already exists in your home dir.

Express Upload Directories

As described before, Mirall can create so called Express Upload Directories for you.

They basically connect a local directory with a directory on the ownCloud. Whatever changes you do in the local directory, either connected to a network or not, is going to be mirrored to the ownCloud right away or as soon as you go online again. No matter if change, add or delete files or directories, the changes go into the cloud.

The local directory is the master. Always keep that in mind!

What happens if changes are made to one of your Express Upload Directories on the ownCloud? Well, the local directory is the master. That means, the changes to files on the cloud are going to be overwritten with the next syncing Mirall performs.

If new files are added on the ownCloud, Mirall disables the Express Upload Folder for you in order to not delete files on the cloud, because, remember, local is the master, and if remote are more files than local, following the logic, they need to be deleted. However, Mirall tries to avoid that and disables the folder and you have the chance to fetch from the cloud directory, ie. get the contents of the remote dir to local.
Warning This contains a risk: If you use the same ownCloud with more than one user or from more than one device at the same time, make sure to not lose data.

Check the Status of your ownCloud

Mirall opens a status dialog if you click on the tray icon and if there are express upload directories configured. The dialog gives information if the directories are in good status, ie. synced correctly, or they are disabled.

If they are disabled, the button fetch lets you pull the data from the ownCloud to your local directory. With that, local files can be deleted!

More information and the future

The latest information about mirall usage and features can be found on the openSUSE wiki and especially the Support Database article for ownCloud and mirall.

The plan is that mirall will continue to be developed and of course become available for more linux distributions. We hope to offer deeper integration in the desktops in openSUSE, expand on the functionality and take advantage of the other features of ownCloud like photo, music, agenda and system configuration synchronization.

Many thanks to Klaas Freitag for the extensive write-up this article is based upon (and of course the work on mirall)

Here is the e-mail from Greg K.H. announcing the removal of systemd from Tumbleweed:

Due to a number of inter dependencies on packages that are not ready for
Tumbleweed, and other interactions with the system that are causing
problems for some users, I’m going to remove systemd from Tumbleweed
today to allow the developers to spend more time on getting it stable
for Factory and 12.1 instead of having to chase down problems that are
specific to Tumbleweed only.

So if you have installed systemd in Tumbleweed, I suggest you now remove
it with a simple:
zypper rm systemd

thanks,

greg k-h

The ‘cloud’ has been a buzzword for quite a while. While some are still rather cynical towards the concept, products like mobile phones with Android have shown the value of putting your data in that huge, amorphous network of servers somewhere. Apple recently introduced their new cloud service and Microsoft has their cloud too. So with the other major players talking cloudy, what does Linux have?

Variety

Let’s define Cloud technology as ‘related to putting data online & sharing among devices’ which is a reasonable definition for our purposes. There is a huge number of technologies connecting openSUSE users to online services. However there is a distinction to be made between commercial or proprietary operating systems and ours. We don’t create a vendor lock-in scenario because we focus on tools that freely connect you to your choice of publicly available services. This is a key distinction because we’re not owning or controlling the cloud that you place your data in. You, the user, get to decide the place where it best fits your needs and comfort level. Yesterday we highlighted integration in our every day applications. Today we focus on file syncing services and especially the cool Free Software project ownCloud!

File sharing and syncing services

A file sharing and syncing service helps you keep the files on several computers in sync and allows you to share such files with others on the web. This is indeed what Apple recently introduced, but others have gone before.

SpiderOak offers such an online space for synchronizing files. Just drop the files in a specific folder on your drive and you can share them between your computers, mobile devices and even with other people. SpiderOak is included in the openSUSE repositories since openSUSE 11.3.

The popular Dropbox service offers a similar online space, and even comes with integration in the Nautilus file manager!

But both services depend on proprietary server components and need a proprietary daemon to run on your system. Dropbox also doesn’t encrypt the data on the server, something Spideroak at least does. SpiderOak also promised to at some point open source their client software. Still, the server side and web interface will remain closed. Luckily there are Free Software solutions out there, two of them I’ll talk about here.

Ye Olde iFolder

Introducing iFolder

iFolder is a Free Software file storage, backup and sharing solution written in C# and Mono. iFolder comes with a desktop client which allows you to manage your iFolders. iFolders can be located on any number of servers, each having a number of folders which you can manage independently. Each folder can be your own or shared with others and files dropped or modified in them will be synced right away. iFolder, like Dropbox and SpiderOak, has clients for Windows, Mac and Linux and new is a web interface so you can access your files anywhere and share them with others on-line.

iFolder does have some powerful capabilities, especially for corporate environments. For example, unlike most competing solutions it only transfers modified portions of files instead of the whole file, saving you some bandwidth. iFolder also offers integration with LDAP, disk quotas, and limits to the number of iFolders each user has. It integrates in the Desktop with a Nautilus plugin which allows you to turn any folder into a shared iFolder and iFolder has fully encypted file transfer and storage. Finally, a big plus of iFolder is the excellent documentation which you can find here.

So unlike Dropbox and SpiderOak, iFolder is fully Free Software and the source is hosted on sourceforge and can be found here. With 9 commits in the last 2 months it is not a hugely active project and the last release is from November 2009. Still bug fixes are being added and lots of improvements were landed in 2010. You can find more up-to-date versions by searching on OBS. And it should be noted that this ‘old’ version of iFolder has been offering all the features of Dropbox and SpiderOak, fully open and free, for years!

introducing ownCloud

ownCloud is an exciting and much more active project, developing file synchronization technology on a Free Software base. However, ownCloud is far more ambitious as you can read in this recent blogpost by lead developer Frank Karlitschek. Frank welcomed Apple to the Cloud, giving a feature comparision between ownCloud and their cloud offering. He mentioned that:

Apple finally [has] features which are essential for people who live in a connected world and have more than one device

Furthermore, after having a look at what Apple came up with, he concludes:

It seems that we support everything iCloud is doing and a bit more if you look into the feature set. And we have a fundamentally more advanced vision and architecture.

ownCloud offers all the basics of accessing files via a variety of means including a web interface and WebDAV. WebDAV support means integration in both Nautilus and Dolphin, as well as all KDE and most GNOME applications. ownCloud can synchronize application data like Bookmarks and settings and allows you to listen to your online music collection in a few ways including from Ampache and via the web interface. The same goes for movies and photo’s while PIM syncing is in the works.

Above and beyond iCloud, work is going on in ownCloud to offer versioning of all files through it’s git backend and the ability to share your files with whomever you want. More importantly, ownCloud is extensible. It is possible to write addons to ownCloud like RSS readers, todo lists and note syncing or even collaborative gaming. And ownCloud doesn’t limit you with DRM or digital fingerprints, allowing you to put your music and movies online without restrictions. Finally, where Apple’s cloud just works with Apple devices (as we’ve come to expect from them), ownCloud is not so picky, working with Windows, Mac and Linux as well as Android, MeeGo and of course the iPhone. This is possible because ownCloud uses open standards and works with Free Software where possible, including the mentioned webDav, HTML5, Open Collaboration Services and Ampache. An access library for applications, to offer better integration with ownCloud, is already available and part of the latest openSUSE.

Upcoming

And more is coming in ownCloud 2.0, scheduled for later this year. The UI has been completely redesigned and there is now media support to play audio and video via the web interface. The mentioned Ampache support, developed by Robin Appelman, will have a nice effect: the awesome Tomahawk music player we mentioned yesterday has Ampache support and thus will be able to play music directly from your ownCloud 2.0 server!

Moreover, user management has been improved as well, and you will be able to extend ownCloud with new functionality via apps and plugins! Finally, ownCloud aims to support not only Chromium and Firefox but also lesser used browsers like Konqueror. A really exciting feature is being worked on by a GSOC student: git versioning of files! There is some UI work to be done, which will then allow you to go back in time and see earlier versions of your files.

Aaron, one of the ownCloud developers, gives his thoughts on where ownCLoud should go:

With this, we can not only compete with Apple & Google, we can leave them behind or integrate with them from our own camp. There’s no reason you shouldn’t be able to play your music from the cloud in Amarok or Tomahawk, view your calendar events on your iPhone and in Kontact and access all of your files from Android or webOS or Gnome or Windows or OS X. You should also be able to back up your home directory and keep it in sync between multiple computers

Control

The most important feature of ownCloud, however, is control. In Frank’s words:

You don´t have to put your private data into the hosting center of a big company

You can install ownCloud on your own server, your desktop at home, a company server but also at a hosting provider supporting PHP, or run it as appliance in some cloud service under your direct control. Because it is Free Software, you fully own your data and only you can determine what happens with it.

All this doesn’t have to mean ownCloud is hard to set up. The upcoming 2.0 version will make it drastically easier to deploy but the current 1.2 release can also be had in a number of very easy ways.

Get ownCloud

install ownCloud

openSUSE 11.4

openSUSE 11.3

You can install owncloud on your own server or even desktop using the openSUSE repositories on OBS. Be sure to read the documentation on starting and managing your ownCloud. You can also use the how-to on Aaron’s blog to get the latest and freshest version installed from source! After following the basic installation steps, you just have to go to 127.0.0.1 in your browser to complete the setup from the web interface.

However, your ownCloud can be had easier. You can use the excellent ownCloud-in-a-box SUSE Studio image to download a virtual image of ownCloud and start running it right away. It gets even more interesting with the Amazon Elastic Cloud Computing support of SUSE Studio. Deploying your image on the Amazon cloud is easy as pie. You could use microinstance for which even a free 1-year subscription is available. Yes, that’s right, you can have your openSUSE ownCloud under your full control on Amazon EC2 for FREE for one year!

If running your own Appliance on EC2, another provider or your own hardware is still a tad complicated, there is an even easier option: packagecloud.com. They offer ownCloud hosting (including a free option, yup) with their convenient web interface. Yes, again, full control over your own data and easy to use.

Helping out

ownCloud is a Free Software project, which means you can get involved. The team is looking for anyone who has some skills in (or wants to learn) PHP, C++, HTML, CSS & Javascript. Drop by #owncloud on freenode or send an email to owncloud on kde.org and get involved today!

Conclusion

Frank closes his blog with:

I think ownCloud has the potential to become the cloud storage solution for people who don´t like vendor lock-ins and care about freedom and privacy

We would like to go a step further and claim that the wide variety of hosting options for ownCloud offers not only more security and privacy but also far more flexibility and, in time, lower prices than what Google, Apple, Microsoft and the other ‘big boys’ bring to the table. Combined with the excellent integration openSUSE desktops already offer with online services, openSUSE has a sunny future in the cloud.

The ‘cloud’ has been a buzzword for quite a while. While some are still rather cynical towards the concept, products like mobile phones with Android have shown the value of putting your data in that huge, amorphous network of servers somewhere. Apple recently introduced their new cloud service and Microsoft has their cloud too. So with the other major players talking cloudy, what does Linux have?

Variety

Let’s define Cloud technology as ‘related to putting data online & sharing among devices’ which is a reasonable definition for our purposes. There is a huge number of technologies connecting openSUSE users to online services. However there is a distinction to be made between commercial or proprietary operating systems and ours. We don’t create a vendor lock-in scenario because we focus on tools that freely connect you to your choice of publicly available services. This is a key distinction because we’re not owning or controlling the cloud that you place your data in. You, the user, get to decide the place where it best fits your needs and comfort level. Today and tomorrow we will highlight some of them here, starting with integration in our every day applications.

Integration with online services

install Tomahawk

openSUSE 11.4

openSUSE 11.3

openSUSE brings online services like Facebook, Flickr and Youtube to your desktop in a variety of applications. New applications like UMPlayer and Tomahawk go out of their way to deeply integrate twitter or last.fm, with Tomahawk even able to play music over a jabber (google-talk) chat connection. Of course, old-timers like Banshee and Amarok also feature music from music stores, pod casts and other on-line services.

Image viewers have engaged online services for a while. F-Spot can export to Flickr, 23, Picasa Web or SmugMug. Shotwell can even export to Youtube. KDE applications Digikam and it’s companion Showfoto add Facebook, piwigo, shwup, zooomr and some other sites to the list. They can also export to a HTML gallery and put the images with the locations they were taken at on Google Maps. This is shared KDE infrastructure so you can expect image viewer Gwenview and even screenshot application ksnapshot to be able to directly share over these services.

install UMPlayer

openSUSE 11.4

openSUSE 11.3

It’s not just media tools which integrate with online services. LibreOffice can export and import documents from online office suites like Google Docs and Zoho. And of course, mail/agenda/news clients like Evolution and KDE PIM have offered integration with online services like mail or calendaring forever. Feed readers like Liferea offer syncing your news feed with Google Reader and TinyTinyRSS.

Besides integrating with existing services, Free Software projects are starting their own. Prominent examples are GNOME’s Tomboy which synchronizes your notes between instances using an online service.

More

But there is more available for openSUSE. There are several file sharing and syncing services out there. Read about that in part II.

• How to install GNOME 3 easily using 1-click install from your favorite Web browser,
• How to switch from my previous repository (home:fcrozat:gnome3, which will be soon phased out) to the new repository. This is important to continue receiving updates for those who installed the live image on their system.

If you are interested in trying out GNOME 3 but don’t feel like upgrading or installing it right away, you can play with the live cd at http://gnome3.org/tryit.html .

Features in this release

To find an overview of what’s new in GNOME 3, read the official GNOME 3 website. We highlight a few features here, some of which are particularly relevant for openSUSE users.

Activity Bar

One of the most exciting features of GNOME 3.0 is the new activity bar. With the activity bar, you can track your todos and day to day activity along with cool features like accessibility, instant messaging, nifty application access and power options. Activity bar replaces your desktop and brings to you a new desktop experience.

NetworkManager integration

openSUSE ships with the latest 0.9 version of NetworkManager. This release brings simplified configuration, support for fast user switching and shared computers as well as integration between “System Connections” and “User Connections”. The new NetworkManager is well integrated in GNOME Shell.

Accessibility

The GNOME community has always put much effort in enabling the use of GNOME software for disabled users. GNOME 3 makes accessibility even more prominent by offering access to a wide range of basic settings directly on the ‘Activities bar’.

Messaging

A very nice feature of GNOME 3 is the pervasive integration of messaging with gnome-shell. You no longer have to run separate applications to be connected to the world. You can communicate right from within your desktop!

Tweakable

GNOME 3 on openSUSE comes with the GNOME 3 tweak tool allowing you to access some hidden features of GNOME Shell. However, due to the design of GNOME Shell, every aspect of it can be modified relatively easy. GNOME Shell is written in Javascript and building on other modern technologies like CSS. While this might not be as easy accessible as a dialog with buttons it is far more flexible for power users. And we expect that the average user will be more than satisfied with the default configuration options. Of course, future releases of GNOME Shell will probably re-introduce some of the functionality which used to be present in the GNOME 2 desktop. The developers simply have not had time to bring back all the most needed features and based on user feedback, priorities have to be set.

Enjoy GNOME 3 on 11.4, as so many others already have!

Note : Please read known issues before installing GNOME3 on your system.

Article Contributed by Nelson Marques

This meeting is a direct follow up of the Cross-distribution meeting on application installer which took place in the 3 days leading up to the Bretzn sprint. During this meeting developers from Debian, Fedora, Mageia, openSUSE, and Ubuntu Linux distributions decided to work together on common APIs and code for application stores on Linux.

The Bretzn Sprint was dedicated to the development of a proof of concept of this idea, using existing components like the MeeGo Garage client and libattica as starting points.

Attendees of this meeting where Will Stephenson, Dan Leinir Turthra Jensen, Frederik Gladhorn, Mateu Batle Sastre, Eugene Trounev, Vincent Untz, Pavol Rusnak and Frank Karlitschek. This team brought together a wide range of skills, as can be seen in the results of this highly productive sprint.

In just two and a half days, they managed to create a working application store client for openSUSE and KDE. The server part was based on OCS (Open Collaboration Services), which already did everything needed for the server side and just needed some data inserted to be shown in the client. The client in turn was created by using the MeeGo Garage client, and adapting it to be a more KDE-like application, by using the various KDE widgets and other tools, removing the duplication of functionality which was covered by the KDE libraries.

Finally, a new application view was created, to replace the existing which was never as pleasing as the original authors wanted it to be. This was done by using Qt Quick, which has allowed for the rapid construction of a much more pleasant look. The new interface is the result of a brainstorming session involving all members of the sprint, based on the results from thecross-distribution meeting. While this work is still in progress, it already represents an improvement over the old interface.

Article contributed by Frank Karlitschek

The issue

As a developer you want to create cool Free Software and you want to get the result out to millions of users. Unfortunately, to get the code to the user you need to do more than just the fun stuff.

After you have written your application you have to compile and package it for all operating systems you want to support. There is a large variety of Linux distributions out there and of course there is Windows and Mac too. Once you’ve build it and package it, you have to create a web page with information about the application and some features like commenting and a bug report tool. Next step is to promote your product on twitter, Facebook and at conferences; and try and get it into distributions. Once the distributions release some users hopefully turn up and start giving feedback; resulting in changes to the application. A new version is released and you have to start all over again with the boring stuff.

The ambition: introducing Bretzn

Frank wants to reduce this whole cycle to 10 minutes of work (well, except for the fun part: the coding). Meet project Bretzn. The goals of this project are to:

1. Make it easy for developers to release an application
2. Make it easy for developers to market and for users to learn about that application
3. Make it easy for users to install the application

Project Bretzn consists of three parts: the building of the package, the publishing of the application, the marketing to the users and the feedback cycle back to the developer. The building part consists of plugins for IDE’s and integration with Build Services and using Lubos Lunak’s OBS Generator (yes, still looking for a new name) to automatically generate cross-platform build files. Publishing sends the app to a variety of App Stores and then markets it through announcing it on Facebook, Twitter, via a RSS feed and the Social Desktop. Finally, users can provide feedback by rating the application and commenting on it, by writing bug reports and send messages to mailing lists or forums. And all of that with one click (well, maybe a few more).

Status

Currently, Sebastian Kugler, Dan Leinir Turthra Jensen and Frank himself have been working on this for about three months. The plans are older, however, as are some parts like the OCS based App Store.

The API is all open, part of the Open Collaboration Services which is on freedesktop.org. There are free server and client implementations supporting uploading and downloading of applications, abstracting Build Services and including all the social features. The API is supported and used by Intel, Nokia, OpenOffice.org, KDE and GNOME.

The new Open Collaboration Services AppStore currently has four different clients:

• the openDesktop series of websites (kde-apps.org, gtk-apps.org, qt-apps.org, gnome-look.org etc)
• a Qt client (supporting Windows, Mac and Linux)
• the MeeGo Garage
• the KDE GHNS app
• a GNOME library (no gui yet)
• Ovi, AppUp and more in development

Current features of the API include categories, screen shots, change logs, commenting, rating, search and update notification. Furthermore, applications can be either free or paid and the payment goes directly to the developer. Not all AppStore clients include all features right now, the KDE GHNS client is probably the most complete as it has been around for many years.

Social features include notifications directly to the desktop using the Social Desktop API, ‘what my friends like” and “what my friends develop” and Knowledge Base integration.

Over 70% of the work is done already and the team is aiming to ship in December. They are working with other openSUSE developers to make a proof-of-concept App Store for openSUSE using OBS and the other integration which will hopefully ship in the upcoming openSUSE 11.4 release.

Plans for the future

Future plans are to show commits of friends, link to bug trackers so users can file bugs, allow users to easily test newer versions so developers can gather feedback and the ability to give donations to developers.

The code will soon be available on Gitorious for all to see and of course anyone interested in helping out is more than welcome. The team is looking for integration in other DE’s like Eclipse, NetBeans, Visual Studio, MonoDevelop and KDevelop. But also work on more App Stores, further social media integration and code cleanup is needed! And especially work on a GNOME frontend, in the spirit of Collaboration Across Borders, would make the team happy.

The Meaning Of Bretzn

This makes the Build Service KDE:43 repository redundant, so it is being removed. KDE:43 users should remove this repository from their installation sources.

For more details and information on how to use other KDE versions on all supported openSUSE editions, see KDE Repositories.

Edit: KDE 4.3 users on older versions of openSUSE are not being left out in the rain.  KDE:KDE4:STABLE:Desktop contains the same KDE SC 4.3.5 that was added to openSUSE 11.2.