 It's OK to start. Welcome, everybody. I'm very impressed how full the room is. So there seems to be a big interest in disaster recovery. And as the first presentation, I will show you very basic things about the kind of disaster recovery which will further on presented today here. First of all, why does it matter? In particular, disaster recovery with relax and recover is currently used more in enterprise environments on thousands and thousands and thousands of server systems. So we are not talking about something which we plan for the future. We are talking about something which is in production use. The topics, what does disaster recovery mean, what do we mean here, this disaster recovery? How does it work? What is not meant with disaster recovery, as we understand it here? And now, the particular implementation of a disaster recovery tool, relax and recover, what is it, how does it work, how to set it up, its advantages and also its disadvantages. It's important to avoid false expectations, to know where its limits are, and then finally, a bottom line. What do we mean with disaster recovery? A disaster means the system got destroyed. It doesn't matter how it got destroyed. Messed up essential files, broken disks, a complete destruction of the system. For example, with ButterFs, you can do snapshotting and get it back. But this does not help if your ButterFs file system gets corrupted or if your disk breaks down. So you need something which works on bare metal hardware, on an empty system. There's no requirements, which kind of disaster can happen, everything. From this, it follows that to recover, it means you must be able to recover on an empty system on a bare metal machine. This means disaster recovery is reinstalling from scratch. There's nothing on the system. So it is essentially an installation. But there's a restriction. As it is now, let's say it works best on same or on full compatible hardware. It can be done on a little bit different hardware. Even more interest than expected. No, it's fine. For me, it's fine. Yes, but emergency exit is right there. And it looks like a window, but they prepared for something so you do not jump out and hope for the best. There are stairs outside. So either come in or please stay out. OK, now it's sufficiently silent again. So I said disaster recovery means to reinstall the system from scratch on new compatible hardware. And it means, because it's an installation more than restoring files, it's more than a backup. You must prepare the hard disk. Then you restore the files. And finally, a bootloader must be installed, basically an installation. How does it work? While your system is up and running, you must save the payload of the system. I call it the payload. It's the files. You create a backup. And we have talks here about backup solutions. This is one part of the whole disaster recovery. And it's, of course, the most important part. Because the files, your own files are your value. Anything else can be done. It needs more time manually. But if you lose your files, you are completely lost. Then while your system is up and running, you prepare for reinstalling from scratch. This means you need to create an installation system that can recreate your system. I call this installation system a recovery system. If you install your, or when you install your Linux, your Windows, your whatever, first of all, there runs an installer system. It runs in the main memory of your computer. And this, from there in, it prepares the hard disk, puts the files on it, installs a bootloader, and then you boot this system from the hard disk. The same kind happens here, too. You have an installation system which runs in the main memory of your computer. And then it does all the same things. This means a recovery system is a recovery installation system. I add the word recovery only to make a distinction between a new, real installer and this one. And there is an installer, a system installer, not an installer of an application software, a system installer. Now, when your system got destroyed, your boot, you're already prepared recovery installation system. And therein, you run the recovery installer. And the installer does what an installer does. It prepares the storage, stores the system payload. In this case, it means it restores the files from the backup. It installs a bootloader. Then you can reboot your new install system. This is all what it does. What disaster recovery is not meant to do, at least not how we discuss it here, is system configuration. The system configuration is the same as it was before. You get the same disk partitioning. You get the same mount point, the same file systems. And then the rest of the configuration, which is in files, is restored from the backup. So you have the same configuration files. And the bootloader is also reinstalled as it was before. There are very limited adaptations, only as far as needed, for example, unique IDs. Or if there is a well-known place where a MAC address is, then this is adapted, but nothing more. The next point is no system migration. You can use it to migrate a system on a little bit different hardware, but it's not the intended usage. And depending on how different the new hardware is, it may be much easier to start from scratch on the new hardware. It depends. If you have thousands of servers and you move them to new hardware, you can invest a few days to get it automated. So it depends. But on your laptop, I think it's a little bit overkill if you buy the next newest best from a different company. I think it's not meant for this use case. And therefore, it's no competition, this various kind of configuration or migration tools. We do something different. We do the bare metal initial things. And we try to do this thing good. Now, what is relax and recover? It is a disaster recovery framework. It is not meant as just install, just run, and just be happy solution. In some well-known generic cases, it works. It even just works. For example, now a little bit at what is meant on some SUSE systems, it just works. Because I run it on SUSE systems, I test it on SUSE systems, in particular, on SUSE Linux enterprise systems. When you install it as the defaults are, then it just works. When you do your special settings, then it may not work. Relax and recover complements the backup. It is not a backup tool, but it calls a backup tool. By default, it's plain tar. But it also, and this is very important for some more professional usage, it has built-in support for many third-party professional backup tools. This built-in support means it knows how to call the backup tool. And it knows how to include the important parts of the third-party backup tool into its recovery system. So you can restore your backup on the new hardware. It is meant for experienced users. It has a command line tool, but there is no graphical user interface. And now one of the most important things, it's written in pure bash scripts. Oh, that's not a programming language. That's ugly. That's terrible. And whatever it is. But this does not matter. When you are on the new bare hardware and you have nothing. You have always a bash. And it's the native language of the system administrator. In the end, you run bash commands. So it has no special requirements. It is meant to be adapted and enhanced as needed for your particular use case when you have a little bit of special use case, you may have to adapt it. How does it work? It's the same as before. There are two commands. The one command is called make backup, mcar backup. And it does two things. It makes an installation system, the rear installation system, the rear recovery installation system. And it calls a tool by default tar to make a backup. After the system was destroyed, I have to say something to the step before. The recovery system is made as a bootable EZO image. After the system was destroyed, you boot on the new hardware the bootable EZO image, the rear recovery system. And therein, you run the rear recovery installer. And it's called rear recover. What does it do? The same as described before. It prepares storage. It stores the payload. It restores the backup. It installs a bootloader. And then you are done. How to set up relax and recover. There's a configuration file. And it contains ETC rear local.conf. And it contains bash variables and bash arrays. And the values of this bash variables and bash arrays determine how it works. If you need to recreate special things that's not already there in rear, you may have to adapt the scripts, the bash scripts. I think it's now the advantages of relax and recover. It's generic. I already told this. It's all pure bash scripting. So if you're an experienced user, you can easily adapt and enhance it. But often, if you have special things, you must adapt and enhance it. It's small. There's no GUI. So you can run it on a system with not too much memory. The recovery system that was created on the already running system is specific for this one kind of system. So if you run it on two different hardware, a little bit different should work. But on two different hardware, it may just fail. And then you're a little bit of lost because it's generic and small. There is not all and everything included to run on all and every hardware. You can do this. There are options to include a lot of things if you know what the target hardware is. So if you have to migrate 1,000 servers to a new hardware, you can use rear to do this. But then you have to know what to do. It does not magically work there. And the last advantage, it's, from my point of view, the biggest advantage. It is fast. It is not exactly fast to configure it. You need to do something. But it is fast when you need your system back. So reinstaller runs directly shell commands. There's nothing in between. And my own experience on SUSE systems, rear is long, long done with recreating the whole system, a system not with a terabyte of data on it. So basic operating system, two or three gigabytes of basic stuff. It's long, long done. While just or out just is still in its startup phase and trying to get things set up and done. So you are done with rear. This is not something against just and out just. This is something against big and fat installers. There's a lot of graphical things that work on any system. They are, of course, they must be a lot slower. But rear is incredible fast. The limitations are, I already told it, what the recovery system can do. It's small. So if something is missing and you run it on the new hardware, let's say a driver for the network card is missing, then you are lost. It's not there. You cannot make your network card. You cannot use your network card on the new system. So again, you must prepare for it. As long as your system runs and you must have your new hardware already there, you prepare for it and then it runs fast as lightning. To deal with the limitations is you must verify it. And I would even recommend to do your actual deployment on the new hardware as a rear recovery installation. You test it, you try it, and then the final step is to get the system on the new hardware. It's a deployment with relax and recover. Then you know this step works. If now this hardware breaks down, you take the next one and you know this step works. And be prepared for manual intervention. It means know your system. If you are clueless how your partitioning is, how you're networking is, then it gets hard, this plane working on the command line. But now, for me, the most important part is the bottom line. And it's not about relax and recover. It's a general statement. Regardless how a system was installed and regardless what you use for disaster recovery, in the end, a disaster recovery installation will be your final system installation. All this graphical nice thing is before. Call it just whatever they're called. After a disaster, this is a final installation. You may decide to do your disaster recovery by running the installer of the Linux distribution. Then you basically do not have a disaster recovery procedure. You just start from scratch. Laptop falls down the stairs, crashed apart. You buy a new one, plug in the DVD from the Linux distributor, and start from scratch again. So this is not meant with disaster recovery procedure here. But if you have any kind of disaster recovery procedure at the end, this one will be the procedure which does the final system installation for you. If you have 1,000 servers, 10 of them get installed via the disaster recovery procedure. That's, for me, all I had to tell you as initial information. So thank you very much.