 Check. Good morning everybody. After the first talk, this is the second talk about to relax and recover. And Johannes, my colleague of the project, will give a live demo. We will switch perhaps first, because he needs some time to start the demo. Yeah, you can switch. I will let Johannes explain what you will do, not too close to your mouth, otherwise you will not like this. Okay, on this laptop, I have installed... We don't see anything yet. We don't see anything. What's wrong? It has worked. Maybe you can explain. Oh, yes. Okay. Now you see what's happening on my laptop. I have the relax and recover software installed. And now I let it do a backup of the data, plus some more things. I create an image, a bootable image, that I can later use to reinstall the laptop. And this takes some time. This is why we started right now. It's, you see, you don't see much. It's just doing its job. So, one minute, start it. And I start here. It's a clock. And now I hand over to Grazien, who will explain what Ria is, how it works while this just runs. But it's a live demo. This backup that is right now made will be later used to recreate the laptop. And I will destroy it. I will not physically destroy it because I only have one. But it makes the system completely... Where is the destination of your backup? Yes, the destination of the backup is something that you can use. It's a USB drive. The faster the drive is, the faster is your backup. So I used a spinning disk. It works better than a USB stick, a cheap one. If you have a high-end USB stick, it's also okay, but I prefer this spinning disk. It's reliable. Okay. We will switch now the display. Okay. All right. Johan has already explained that he... Okay. He already explained that he started his rear backup on the USB device. This is a more unthinking slide for why is the need of relaxer recover and to have a disaster recovery image of your Linux system. Whatever. You all here, you'll probably understand why you need it. Let's skip that. Let's go directly to relaxer recover. It is a disaster recovery solution. It works online. That is very important. So you don't have to take an offline backup. Systems can stay offline, online, I mean. It is completely written in bash. We have chosen, you have to... Rear was designed in 2006. Quite some time ago already. And before Rear there was a make CDROM recovery project also written in bash. So that's just extended it. The nice thing about Rear is it is completely modular written. So it means each basket can be very, very tiny. Can be bigger. But you can just make everything very modular and you can test it very easily like that. It is working on a lot of architectures now even on Zlinux on the IBM. So IBM is a big sponsor now for making Rear working fine on their mainframes like systems, Lbar systems and so on and so on. So everything with x86 is supported of course. PPCs are supported and some other architectures. Okay, I already said it's online. And it integrates of course with backup solutions, external backup solutions and internal back solutions. I come back to that. All right, Rear is... Well, it starts in 2006. I think in 2009 it gets a broader knowledge about the... In the communities, big customers that start using it. It is open source so they don't have to pay for it of course, but they use it. I think Suzy was one of the first that really shits relax and recover with their high availability solution. And after that Fedora and Rail came. I think in most operate systems it is now part of the mainstream executables or packages I have to say. The nice thing about Rear is that the integration is very well working with external backup solutions so that Rear is very concentrating itself on the disaster recovery only. And the backup is effect something else. So the purposes of Rear is to recover as fast as possible your basic operator system. And the rest can be recovered through external backup if needed. Or you can choose to do it directly with external backup and then you only have a very small part of a multiple image that only recovers your basic drives and file systems mounted that then the recovery starts through the external backup interfaces. And it skills very well. In the next talk I will talk about scaling. So that's also interesting. Already explained use Rear or use the best tool for your job effect. You can do Rear recovery and you can recover everything from A to Z in fact for your system and you can also do only the basic US recovery and then use something else to recover your data to recover your configuration by means of configuration management software management, et cetera, et cetera. So Rear is effect the best tool for a job and it is really meant to be excellent on the system devices operator system level. All the rest can be done or can be added but it's not necessarily needed. If you're fine with for example Pippet or Seth to do your configuration management or restore everything afterwards on system level you can do it, you can just integrate it. Okay, let's explain the backup and the output flow because that's important to understand. There are two flows, it speaks for itself. Backup is meant to send your backup of your operator system to a means. You have an internal one. The internal one that we define as using tar or r-sync. Everything which is external can be another program to do backup. For example, barials. We have done several barials examples in other presentations as an external backup solution. It's worked perfectly fine but there are so many. I think almost all backup solutions are supported. I can only think about one or two that's not yet inside the rear but for all the rest of backup solutions are already integrated. That's the backup. The output is effect your disaster image. The image that you create to boot from afterwards and the image can be an ISO image, can be on a USB stick, can be network booting. These are the possibilities. So you have to decide if you want to make a disaster recovery schema, you have to decide where do I want to put my backup? Do I want to boot from ISO? Do I want to boot from USB or network? We'll recite the backups. Is it on the NAS environment? Is it on the cloud? Because it's also integration in the cloud as possible or it's already in rear itself. So there are a lot of possibilities. So you have to decide and once you've made the decision, you can start configuring rear in the configuration files. So the basics of rear is of course there's a main script. It's a batch script. Everything in the rear is batch. So also the configuration files. So it's also batch script. Well, it's a variable there but if something missing, a code is missing, you will get an error in the rear. So remember that. So the main commands are make rescue or make backup. Well, it's a difference. Make rescue will only make the disaster recovery image and no backup will be made. The make backup includes the rescue image and the backup. If you're using an external backup for example, then normally make backup is the same as make rescue because you rely on the external backup interfaces to make your full backup of your system. So you don't need rear to take an extra backup. Although it is a possibility to force it. For example, barriers that made a very tiny script that forces a backup. That's a possibility. Everything is possible in the rear. Another important part is the recover part. Of course, recovery you only do when you boot off your rescue image and then you can do a recover. But here is clever enough that you can do a recover on your live OS, right? Documentation is available and there's a very good main page and the websites is also there to explain the things we saw, not clear. And there are also user mailing lists and stuff like that. And there's also a hit hub with issues, lots of them. And you can find a lot of information on the internet. So the methods explain the backup already which are some possibilities. Duplicity is a cloud backup solution for example. And the backup URL is in fact defining the location of your backup. So you say backup for example equals netFS. That means that you will use the internal network, more or less. What we call netFS is more NFS slide, okay? But a location has to be added and that you do in backup URL. The same is true for output. Output is then ISO, PXE, OBD, AIR. Is somebody still using bootable tape drives? I don't think so. And USB. And same output URL. If you don't define an output URL, it will be the same as the backup URL. That's important to know, but all right. This is an overview of the possibilities. If you say output is PXE and backup is netFS, then that means that you will be booting from a network somewhere. That you need a PXE environment to boot from. Rear can set it up for you. That is not a possibility that is there. USB makes a bootable USB drive. ISO image is booting from an image or a physical ISO or DVD is also possibility. You can make unbootable image including the backup and you can put it on an unbootable DVD. There is a few big scanner companies in the world. There are only two of them. And they're all using Linux for their scanning software and they all have a DVD delivered and that is made with Rear to do the recovery of their scanning machine. They don't know it, but I know what they're doing. Okay, some examples of the locations of the URLs. So it's quite easy, I would say. File, NFS, SIFS, Samba, USB, ISO and even tape. NetFS, this is an example that you can add in the local configuration file. Backup program is the default internal backup solution but you can replace it with R-Sync. If R-Sync is used internally because you also have R-Sync that you can point to somewhere, even to the cloud where we are speaking but then it's not an internal backup anymore. And there are some options. You can find all the options that are described here in the default configuration file. If you just find it default.conf and you find everything in there, everything which is possible within Rear is described in there. What is coming now is just example configuration files. The presentation is already online on the Fosden website. I will also publish it on our website but okay, there are possibilities. There's so many things that are already shipped with Rear. You can find them in the default configuration. Subdirectories in the examples. I think there are a lot of examples in there which are useful as a start. If you are not sure what to do, you can just have a look at it and then you have a clue effect. Another thing that within Rear, you can also type RearDump and then it will show you which configuration files will be read. And it is in this order. That means the last look, local.conf is the last configuration file that will be read. So everything which is added there will overrule everything else which is over there. That's important to know. Local.conf is the last one. Here are our examples. This is a very simple example. Copy as is, you can just add your own executables to the disaster recovery image. Result files, you can add anything to there which is also left there. Iso is the standard or the default, I would say. Not the standard, it's a default. But most of the time, so it's the most used. An interesting thing is here, the only include volume group. If you have a system disk, we will switch them. We will switch because the rest is only configuration examples. You can exclude stuff from your backup that you don't want. And you can add even users because that's what a customer can do that if they want to recover Oracle immediately and often disaster recovery, then it's there. And we do it automatically. SAP, we do exercises with SAP also. So you want to exclude the data files because you recover from external backup, et cetera, et cetera. Okay, I will switch back to Johannes. So now I made the backup. It took about 12 minutes. So, and the backup is now on the USB drive and there's also the USB drive is prepared to boot the recovery system. Both things are now on the USB drive. Now I destroy this system as good as possible with software, only one example because it's fast, I remove all libraries. Just, so now nothing works. Simple commands like list something doesn't work. Okay, let's do a reboot because perhaps it helps, also not possible. So just, yes, messed up. So I need to reboot the hardware, it's a very hard way. So now it's off. Now I boot the system that is on the USB drive. Escape, F9, what's, this is HDMI cable. I don't know what is going on. It has worked. I try the Windows Bay redo things. I don't know. So it will not be on the video, but for you, you can see, I boot from the USB drive, USB hard drive. Enter. Now something, I don't understand. What's wrong? And now I have a boot menu from the recovery system on the USB drive and I choose the topmost image. This one, and again, let's hope. I think when the kernel starts, then things get better. Now it loads the kernel from the unit, from the USB drive and the initial RAM disk. I would show. Let's wait. Now the kernel boots and nothing else. One cannot see it. Oh, okay. So for those on the video, I booted from the USB drive. Nothing was seen, there was no output on the HDMI cable, so there's nothing on the video. But what is now booted is relax and recover rescue system or recovery system. I log in there as root and type the recovery command. There's one question. If the disk layout is the same, it's to avoid that you get interesting effects if you install the new system on the USB drive. So there are now two disks, the USB disk and the system disk. So in certain cases, it can happen that the installation happens on the wrong disk. So there is a check that this does not so often happen. It's clear we have two drives. I have 10 minutes left. That's good because the backup took about 15 minutes, 12 minutes, but this is usually much faster. And now we have to wait to restore the backup. What it does, you see, it does a complete system setup. It does partitioning, creates partitions and all these things. And at the end, the backup is restored. And then I can reboot the system. There are other possibilities. In this case, I would not need to do the partitioning again or to reinstall the boot loader. I only messed up some files. And I have it prepared on the system, but I cannot show it because I have no time here. You can boot the recovery system as another entry of the regular grub that is on the hard disk. So you do not need to boot from the USB drive. You boot the recovery system from your normal drive. And then you can restore the backup also from your normal drive if you have it on a separate partition, on your normal drive. So this could help for software errors. And I show here the complete recreation of the system which you need if your hard drive gets broken, if it's physically broken. And what also should work is that you connect this USB drive to another laptop that is sufficiently similar and install it from there, from the USB drive. So at the end, it's also not needed. In this case, re-migrate some configuration files. For example, if you have different network, you have to do some little adjustments, but it's not a complete reconfiguration of the system. It's only the very minimum that is needed to boot the, to get the re-created system up and running. So if you have a, if you do a complicated migration, you have to do something more. You cannot just let re-do everything. But you can, of course, it's batch scripting. You can extend and enhance it as you need it. And I think this is the main reason why it's so much used on server systems and so. I think admins like it that if something needs to be done, they just hack it in. They need to, they need to do a complicated, to learn a complicated system. Just look what's a batch script. So, oh, this looks like the right place, hack it in. Perhaps it's not exactly the right place, but you'll get your thing done in a relatively short time. Now it installs the bootloader, which is also not needed in this case, but in general it's, it's done. And I could speed things up. I let re-automatically determine where to install the bootloader. This takes a little bit of time. And if I specify install the bootloader on dev SDAs, then things get a little bit faster. You can set up so many things. So many possibilities to find tuning as you need. But the more you, you configure the less it works in the general case, if you have to do it on a different laptop. But it's also not a big problem, because in the recovery system, you can adjust what you like. You have a complete Linux system. Only this command line, no, graphical interface. But theoretically, you can rear. Now I reboot, just a moment, reboot. So I unplug the USB drive when the system is off to show you that there's only the hard drive. There's again nothing on the video, but now it boots again. It's my standard system. It's the one that I, that I destroyed. I swear it's the one that, it's no, no trick. It took six minutes, something like this, 10 minutes. And I have here not a test system, a small one. It's a regular normal system with GNOME desktop and LibreOffice and all sorts of things. But of course, if I had two terabytes of videos, I would not back up them with rear. Because if you need to go, to be back soon, and the restore of the backup takes hours, that doesn't help. Usually you don't need your videos. So this is also something that Grazian explained. You can do the backups in parts. Your basic system, your basic system and your essential files that you really need and your optional files. So it's flexible. Bottom line, rear is flexible. Yes, I'm back again. Perhaps I should even start something. So a browser, it doesn't matter, it works. Stop, yeah. Only to show, it's the same system that was completely messed up. Now it's working, seven minutes later. Perhaps it helps you if you, that's not this time, next for them. You have rear with USB device. So if you mess up something, you just show something to someone and oh shit, everything messed up. No problem, your talk one hour later is safe. You reinstall your system and nobody noticed that you did something bad. Okay, that's all from me. I think the next thing is question and answers and I would give back the microphone to Grazian. If he wants, I should answer. So we do it both. Questions? Yes, we have support for what it is called. Looks, yes, of course. Yes, and if you have mega looks, something else, then it's shell scripting and it's not so complicated to add it. Oh, I forgot, I must repeat the question. The question was if disc, if encrypted discs are supported, yes, they are supported when they use looks as encryption mechanism. At the end, or there was some line, please type our MRF10 real something. What's stored in the template now? Why do I need to read it? The question was at the end, because some line, one should remove TMP rear something. I have a special setting that's not the default. In the temp folder, rear builds its recovery system and does its temporary stuff. And usually it gets deleted by default, but because I mainly develop rear, I usually keep it because then I can look what they're in. So that's not the default and of course as everything you can, there's a configuration variable and you set it and then things happen different. Migration from physical to virtual is possible, but as always when you migrate and things change too much then you have to do something. For example, you have a physical machine with two hard drives and a virtual machine with one. Then the hard drives configuration cannot be applied directly on the virtual machine. Maybe one important thing to mention. No, no, you can keep it for a moment. If you want to do physical to virtual, there is something as a disk layout. So the rear also creates a disk layout configuration. You can edit that file and map it to the destination effect if you have one disk less or something like that. I don't care if it's a product or if I need to add it to the rear. That's it, the time is up. Just one thing from the virtual machine. Yeah, that's us, we do that. For example, I think it was not Amazon, but Zalando did it. Yeah, one of the project leaders from in the beginning, he worked a long time for Zalando. He used a lot of times rear to do physical to virtual. You're welcome.