 So, we start this free, small session about boots with Harold Boyer from U.S. to work on still all the characteristics of how to improve the leg system with some of the patches that you are going to be using to get the other U.S. system, the heel education to get the patches, and finally, back at which we are going to be talking about today's talk. So, please welcome Harold. Thank you very much. I'll tell you about WREKIT, which is a generic cross-distribution in-and-round test generator. And the main benefits for them are to use WREKIT SR in-and-round test generator. WREKIT is the unification for to use the same tools on really all the distributions. And we can have content support to improve the software of course. WREKIT is able to handle very difficult storage combinations because it's event-based with U.S. and in the last year, it also system-dead capabilities and integrates nicely into the system-dead infrastructure. We also gain the benefits of having system-dead in-and-round tests. WREKIT is able to boot from LVM on anti-rate, on DM-rate. You can stack everything like you want. As long as you specify every parameters on the kernel for one line option, assemble those devices. Like specifying UUITs for the rate or specifying volume names for LVM. There's another mode, which is called host-only. They can celebrate it. Generally, an integrator of S, which is only for this machine with this layout. WREKIT wrote everything about it on the disk. And you don't have to specify anything on the kernel command, but except the root people, the root processor. Internally, WREKIT has a very extensive test suite. So it starts virtual machines, which boot even starts in a S servers or ice-passing servers. And they talk to each other, the virtual machines, so we can make sure that a release is stable and the root versions are put in the last code. WREKIT is very fast. If you boot it with Sys&D, because then there's no shell script involved in the single cases. And it's even faster than booting the kernel of the complicated kernel test. For the simple case, like a better test volume of X4, the kernel is slower than the combination of the kernel and the inner kernel test. Maybe they fixed the kernel to get the same speed, but currently at this point of time it's faster. WREKIT is very well documented for the user, but not so for the developer, because the shell code is very easy to read. So I can give you a peek of the documentation which is shipped and the package is very long. I think you will have an overview of how the boot is going in Sys&D setups. Yeah, in Sys&D we have some type of integration, because we work very closely together. If you have Sys&D in Vivian, you strike it to get all the benefits. Like early logging with the JOL, we have some fake introspection afterwards. If you are at the real boot, you can see which great service units have been started and what the output was. And what has never been done before, I think, on Fedora, if you put a Fedora, it disassembles all the root device components on shutdown. So you can tell the X1 server, OK, I'm done with that volume. Or the NFS server, this is finished now, and you get a clean state for the next boot. Also for MD rates, it's very crucial to wait until everything has synced before you disassemble it or reboot, so that you don't have to switch to the next reboot. Features of breakage are that they can build, like I said before, host only images, which are very small, very specific, and only include the panel modules that you need to do that, but only that specific machine, and with that specific layout or assembly layout. You can also build a generic one, which has all the modules you normally need, like LVM, anti-rave, whatever. And nothing is pre-configured in the Enable OS, so every configuration, how those, this RAID or this LVM is assembled, it has to be specified on the count of LVM, so you say, I want all RAID members with new RAID, new ID, blah, blah, blah, assemble. I want to activate these volume groups, or I want to activate these logical volumes. And that is the case of a generic one, which you can ship and just put it on most of the computers, just put anything which you have on this, as long as you know the U11s or the names. Then you can also build rescue systems with it, very easily. You can include boot charts, helpers with system D, there are system D boot charts, and for another system D, systems that was the old boot chart, people who maybe, which you can include in the Enable OS, so you get an idea of how the boot looks like even in your Enable OS. RAID supports FIPS, which is certification for government machines. You can drop the capabilities on one system D set-ups for the real boot, so maybe you want to load all the current modules in the Enable OS, which you want to use, and drop the capability of the whole system to load any module ever on the real boot. So you can have a fixed set of modules, which cannot be altered. It supports internationalization, not so much now on Debian, because some of the parts of the keyboard layouts and fonts are not that common between distributions. That's maybe one reason to have one layout for all distributions to have such tools work. It has a network stack in it, because it has to be cross distributions, we have to have some minimal network stack, not probably from Debian or whatever. We just made another one. We hope that maybe we get the system needs on the network capabilities at some point of time, so there's been specifically a wide range of common network set up with the real system. RAID supports boot splash programs like Windows, or generous splash, which is from Dental, which is playing nice splash screen, and even ask for the password. We can have a quick look at this. Break into ports as the 19 sets you use, with various IBM tests and networks. Supports mainly de-cache already, but it has, of course, multiple devices, which is not so common. It's about, of course, quick relax, de-embrace and de-rate. In squash live, it's something how we build our live CDs on Fedora, where you have a squash vest image, which you overlaid with some GN overlaid, with a persistent storage. When you have your live USB stick, you can have a persistent storage. Then, of course, it supports LVM, Multi-Path, ZFS, by the channel called iScalzy, MVE, MMS, and resuming from disk. Resuming from disk is not as easy as it looks like, because you have to wait until all the spot can show up so that you can scan if there is no signature from our evaluation image. It supports various security systems, by the state of the name, if you wanted that, but I wouldn't recommend that. The e-click system, as it is, of course, you can even have ZysLock in there for the non-SYSTEMD systems. And, of course, SYSTEMD is supported. And with our user match, we had to mount a slash user from within the environment. The same has to be done, maybe, on the internet, with the user, because it has to know how to insist on the environment, how to activate this slash user at the end, what to convolve you, because it seems like you can do that before it tries to mount every S-type entry. Great, it has been in Fedora since version 12, which is a long time model. And Retta had the question since version 6, which has been since 2010. And then Avian, since Screech, I think, and we are to talk about it. They are often users who want to switch to great, but they are lazy. They just want to keep what they have, but they always, every, every time they say, yes, yes, we want to switch, but nobody's doing it. And what they want to switch is any other one. There are over 100 individuals who should be users to great, and, for example, the last 12 months there, I think, 37. And now I'm going to show a small demo, which is always two people. So, I will put another one here. The process will be now. We'll come back to it with this, and with LVM, they all can come up with it. So, you see there's a switch to the user, and it's an LVM with multiple logical volumes. And just check out Screech it. Great possibility. And I'm going to run away from it and put it in the kit repository before installing it on the system. So, it doesn't template your installation. Wait, wait, wait, wait. No, you cannot read the Screech. Screech, are you going to read it all? Yes. That's true. Maybe I can change the console. Okay. So, that demo is, well, of some useful for you in the back row, so maybe this one. But just to show that it works, select the graph entry with my credit, and then our test. And you also have to put in the password, you put the disk, and then it works. That's what I said, yeah. Of course, I couldn't lie. Second thing, with the LVM installation, the system need installed, just put in the password, and we're done. And it even was faster than before. Maybe how to get into the environment as without, you can set break points in the environment as rd.break, you just put on the card and you have a shell in the environment as before, which is real good. You can introspect the journal, change the layout, which is very nice from, within the environment as you get all the nice system features of the journal, and of course, the system control also works. Well, basically, that's what I have. I will have to show you more if you can read it. You're welcome to contribute to the data project. It's on www.arc, it has a great page. They can find all the links. The download directly is also on www.arc and the page repository is also on www.pitchhub and it's also in front of the safe. So we have any questions about it? Now that I'm here, it's the right time to post any questions. So, I don't have any questions to you, but if somebody likes to use break hearts with their data, or wants to help, I'm the maintainer of the break heart package and the current status is that 031 is upstream version and this is also the version that we have in un-stable and the only thing I would say that's currently not working in the end is the I-18N internationalization support but I'm working on this in two weeks in this state. So, just grab the package and try it. I use break hearts for nearly all of my machines and no postings over one year. We're going to have no problems. What's also very nice is the debugging support without the break. Yeah, but just try it and if you have any questions which are definitely related to the package, you can just contact me. That's really, for me at the package maintainer it was really nice to see that the cooperation was very easy. So, between distributions, Fedora, David, that's really nice work. Thank you very much. You're welcome. Only because of Vivian which has another name for your rule or DM and for LVM that's probably currently with the current version that it doesn't include the unit rules. So, you might want to try the Hitchhack or the next version. It's 31. So, you want to check it out? Yeah, yeah. So, also my one is... We're about to make another one. So, you might want to stop re-name to make that another game. I'm sorry. I didn't make the other one. For some reason, besides to re-name upstream unit rules to something they think it's elegant we're working on it. And this is really it. It's not pointed. It makes something difficult. We've just seen no point. I think I can answer that. We're working on it, but it's a hard problem because, you know, you need to make sure that nothing breaks in process but we're confident that it will arrive at the actual rules of action. It's not only the unit package, it's also like the other game and the device that they're working on. People think they have nicer names but it's not the case in my opinion. So, I actually have a question maybe for you as the maintainer. As far as I remember, there are some packages that ship hooks for the internet. For example, there's the dock gear in the last interview on boot, which I actually use. It was pretty nice. We started working with Frick, currently. I did not use this so we have to try it. What's really nice working is the broken scripts for the regeneration of the datagram. It has a new kernel package and so on. Let's try this. And I can really say most things were very easy to adjust to be a process of generation in front of the generator. We put all the scripts on the packages to be invaded itself first and then we try to move that to the other packages. But I don't think you're flagging the word anymore. So we just make that clear. There's no hook. We can do that. There are hooks. Yes, but they are not the same like in your universe as the infrastructure we should currently have. There are hooks, of course. But it's a different system. So the modules themselves are kind of hooked. The modules also are hooked. The entirety of the track is only hooked. It's just different from what they are. Any more questions? Yes. Is RD.flake a good feature? Or is it a dynamic kernel feature? It's a right feature because the kernel has booted the environment. So it can't be a kernel feature. It's a black feature. Okay, thank you. I prefix all the command line options in RD. So we can easily tell them apart from the rest. I just have to be maintaining natural voting rights and I have pretty some support for the NPA in the universe. I know there is some support currently in Berkeley as well. Yes. Is there any documentation that I can make on that? So I can make at least the current command line options are well documented. Okay. Where is this allocation? You can show me how some test suites are running on working on a production success from the test suites. My test suite I run on Tobogno. Well, it should be easy to be if any of you run on Debian 2 because it uses just some of the tools. There's a weekly for the test suite which command lines you have to install like the servers, the Iceman server or the entity server or what else. Now how do you call them if you have the configuration parts that can happen in different places? That's always the problem. Changing names from the upstream so maybe there are diodes from the upstream maybe. So that's the only problem you will have for the test suite. Of course, there's one test with rpm which can skip, of course. Yeah. More questions? If not, thank you. And I hope you will adapt to it. I have one thing to say. I would like to break out the box. I think at 3.30pm today we are I like to discuss what should be done or what has to be done to make break out the default inner-gram address or do we want this? So if you would like to discuss this question please join the box and we can talk about it.