 Hola a tots! Gràcies per venir a l'estolc sobre desasterrecoverimentatge amb relaxant i recuperant desasterrecoverimentatges. El meu nom és Diago Liveira, soc un dels fundadors del projecte d'Irland projecte amb el meu amic, Pao Roura i Roen Carbonell. Aquest estolc consistirà en dues parts. El primer és l'introducció de l'estolc d'Irland projecte i l'estolc d'Irland itself. I després és un workshop. Vull mostrar a tu com m'enganxar les desasterrecoverimentatges per Linux amb relaxant i recuperant l'estolc d'Irland. Hi ha instruccions a GitHub. Si vols seguir-les, no tenim prou temps per preparar-ho ara, però potser, si he vist la presentació a l'estolc d'Irland, hi ha instruccions. Si l'estolc d'Irland és almost prou, vaig durar 5 o 10 minuts per acabar-ho. Si vols seguir-les amb els teus laptops, jo m'ho demostraré. I aquí, després. First of all, un moment, sobre l'introducció de l'Irland. Què són els features de l'estolc d'Irland projecte, com a contribució a l'estolc d'Irland i l'estolc d'Irland. First of all, què és l'estolc d'Irland? És una resolució de l'estolc d'Irland. T´ merchandising with relaxation recover is a central management tool for relaxation recover, and it focuses on diasterecovery of the network. Without using Isoimage or USB drives, just booting over the network and recovering the system. It helps the seasoned ninja de mantenir tots els serveis d'arribar i recuperar el sistema de Linux per a tu. I és el principal concern de l'AMM, és el que és més fàcil de utilitzar, i en el cas que les sèrmines poden menjar, en una manera més fàcil, les sèrmines de reconeixement per a Linux. És basada en la reconeixement i recuperar el framework. És també llit en baix. I, bé, volem veure com funciona ara. Well, l'AMM ha utilitzat serveis que són necessaris per reconeixement i recuperar els sistemes sobre la reconeixement. TFTP i DHCP són per la reconeixement de la reconeixement. I l'AMM agraït de la configuració i de la reconeixement per a tu, per mantenir aquestes configuracions, cada vegada que agraït l'AMM client o el reconeixement. SSH és per la comunicació entre l'AMM i els clients. L'HTTP servei és per fer la configuració de la reconeixement en ran de temps, en una lloc central. A més, per anar a cada servei, configurar-la a cada machine. Pots configurar-la en un servei d'AMM, i en ran de temps, quan l'AMM agraït de temps, per fer la reconeixement, la reconeixement s'ha d'enllocar de la reconeixement de la reconeixement de la reconeixement. L'AMM servei és per exportar el dinàster de reconeixement a les clients i el dinàster de reconeixement és per asseure que després de l'AMM servei tot el dinàster de reconeixement arribi al radiallable sobre la nèvora per recuperar el sistema. Com a final, el backup que tenim en la reconeixement. Aquí ens mostrarà el doble, però és l'execució del backup de DLM. El primer punt és per verificar el client. Si és alive, és el client del DLM. És registrat en database. Després, col·legueixem l'informació del client del DLM. Quan tenim la necessitat de l'espai, construïm la imatge del directe. Després, comencem el backup. Quan acabem, registrem el backup en database. Després d'això, col·legueixem el backup de DLM. Si algunes d'aquestes dades fail, si el DLM és probablement configurat, pot repartir l'error a alguns tools de monitoring, automàticament. I pot saber si alguna cosa va malament per poder-hi mirar i fer-ho sol. No ha passat. Si algunes d'aquestes dades fail, després de repartir l'error, DLM es repartirà a les primeres imatges per poder-hi recuperar el backup. Després d'aquestes projectes de backup, us mostraré també el procés de recuperació del DLM. El primer és el bot de nete, perquè, com he dit, el focus del DLM és poder-hi recuperar el sistema de nete. Després, el kernel i l'initard de files són llocs. Volem entrar a la moda de recuperació relaxant. Després, amb el setup de netes properes, amb l'HClient, es pot obtenir l'IP d'adres del servei que és controlat per DLM. No importa si tens una configuració de netes. DLM et dona l'IP d'adres que s'ha configurat per poder-hi recuperar el sistema. Després de recuperar-ho, quan acabis, es pot rebootar i obtenir el sistema del que va ser, amb l'últim backup. Però, si et vols, si et preveus un temps de cloning o migració, es pot ajustar la configuració de nete o el nom de la màxima, per restartar-lo i fer-ho regalar el servei. El DLM és basant en la recuperació relaxant i no tenim cap tipus. No tenim una interface gráfica. Aquestes són les operacions que pots fer amb l'interfacció de DLM. T'has de manejar les netes, els clients i els backups. I, al final, t'has d'integrair el joc de scheduler de DLM per planar uns backups de DLM. T'has d'ajudar i modificar o fer-li la configuració de DLM. I, amb simple comacitat, es posa sobre la configuració de totes les services i les configuracions. Ara, volem veure algunes de les features que el DLM té. Una d'aquestes és l'interfacció amb algunes standard sistemònitaris. Pots repartir per e-mail, utilitzar Xevix o Nageos i també amb HP OpenView. Pots repartir per aquests sistemes i pot ser extendit per suportar més sistemònitaris com aquests. Just necessitem d'ajudar una implementació en baix per poder repartir els errors per el que prefereu monitorar. Una altra interessant feature de DLM és que té una instal·lació atendida de relax i recupera les dependències i també fa una basic configuració de relax i recupera. La instal·lació és molt útil perquè si un servei s'ha d'organitzar per DLM cal instal·lar relax i recupera les dependències i també fa unes configuracions per poder fer-la i amb instal·lació atendida cal instal·lar-la a la client i llavors es repartirà i recuperar-la després. No cal cada configuració de relax i recupera instal·lar-la etcètera. A aquest moment suporta Vivian, Ubuntu, Sushi i Red Hat distribuïcions i variants CentOS, OpenSushi també perquè el primer concepte de DLM és l'eixi de utilitzar és una de les més grans fites per tenir un desastre recorregut en vermell. Ok? Another feature that is new in the last version is the integrated job scheduler. With this with a single command line you can you can plan backups for in a daily, weekly, monthly even yearly manner and also running planning backup for today at night because I don't want to be on the terminal typing it on waiting for the result and just with a single command you can maintain a scheduler of your backup jobs from DLM. Ok? And another new feature that we have in the new version is the export and import of disaster recovery match between DLM servers. Ok? Usually we will have a production data center with one DLM server maintaining all your disaster recovery environment but maybe you have a contingency site with another DLM server and you want that the contingency site will be ready in case of disaster to recover those disaster recovery match on the contingency site. Well, you just need to export the image from the source DLM server send it to the target server un important and after that you will be ready to recover that image from the target disaster recovery Linux manager server. Ok? Well, just some features a at the end of the presentation that it's a at FOSM page you can see more more features and more extended information of the presentation in the appendix ok? But we don't have time to say all the things today at this talk. Well now we will talk about the project history just a brief a brief explanation the project is 3 years old we started at 2013 at summer and as you can see 5 months later we had the first stable version of DLM this was thanks to because to relax and recover we used the framework and this helps us to develop it faster and on top a stable framework and a really good framework for us because it's well done it's very well done and after that one year later we integrated fully with relax and recover with the central configuration also from DLM server and then last summer we released version 2 of DLM with some major changes and now we have the 2.1 version almost ready to be released we are finishing some things at the documentation finishing the documentation of the last version in order to be able to release the new the last version today at the workshop we will use the develop branch of DLM that is quite stable to test it here with the new features and I hope that next week it will be released well, another thing it's an open source project and how to contribute to it you can use it and try it use it and send a full request to our YouTube page reporting issues if you are using it to us in order to improve and enhance the software and share your experience also with DLM could be a must if you are using it and you are happy with it please share your experience with more people in order to more and more people can use it well, that's all for the contribution you can see more information in the DLM.org site if you want and well, the roadmap I encourage you to come to the last talk of this depth room today in order to discuss about some future ideas of not just DLM also relax and recover and veros to talk about the next version and maybe some of your ideas could be helpful and we could implement them in the next version of the project ok? well, the introduction is almost done now I will start the workshop in few minutes I don't know if some body have the workshop prepared or have seen the workshop instructions I will give you 10 minutes more or less in order to if anyone could try to follow me with the workshop could be done with their laptops but anyway you can download the workshop from the or with hot page and try it at home at work and test it and play it because it's a background deployment that starts up some boxes to be installed relax and recover from DLM and test and play with it if you want at any time it will be available on GitHub and it's pretty easy to use it and follow the instructions in order to play with DLM and see how it works easily ok? no, let me have some time to prepare the workshop and if you have any problem of any questions I can... thank you thank you for coming ah, no? no, això que si escrius algo estaca aquí no es labora si pots fer-la més petita però és per allà ja, ja ho sé moltes gràcies estic prou si algú preveix alguna cosa si no, continuaré amb el workshop ok? el workshop d'envernament no és aquí però tenim tres machines client una és CentOS l'altre és Ubuntu i l'altre és d'Ivian van actar d'DLM clients i tenim també dos servers un Ubuntu server i un altre, d'Ivian hi havia... vull també una boxa més però la imatge d'envernament és molt gran i ens espereu una hora i així i jo et donaré una prova a testar-la aquí una hora no és una opció per aquest talk però, de segon ets a la instrucció d'atgithab amb el baram et pots encarmentar Testeu el que vulguis, per jugar amb el que vulguis, d'acord? La primera cosa que hem de fer és la instal·lació. La instal·lació no és cobrada en el workshop, perquè les instruccions són a la page de la DLM documentació. Tu pots instal·lar l'Hard Enterprise Linux o CentOS, d'Idian Ubuntu i also Slash 12, per la última versió. Havíem d'ajudar Super4 Ubuntu i Slash en aquesta última versió. I les instruccions no poden ser seguides aquí. A veure què és. DLM documentació, tu pots seguir les instruccions per instal·lar DLM. En aquest workshop, tenim el DLM deploy en els serveis d'DLM, prou per a tu, per intentar com funciona. La primera cosa que hem de fer, per fer-ho, és saber el nostre network. DLM ha de recuperar el nostre network, i hem de posar els nostres networks al DLM per poder menjar els serveis de recerca als clients d'DLM. Aquest són els comuns, però he de preparar els comuns aquí perquè és fàstic. En aquesta, hem de posar un nen network. No hi ha cap nen, només per referència. El network IP address, el network mask, el gateway, i el IP address d'DLM. Per configurar els serveis, després d'ajudar, aquí hi ha els altres networks d'DLM. Quan els networks són en DLM, hem de posar clients, instal·lar-los al servei d'DLM. A veure. Posem els clients a DLM amb diferents comuns. Aquest client en un futur que pot registrar al client. També, amb les seves comuns, pots instal·lar el seu client i les recerca d'extensió Samsung de la comandant, la mateixa comandant. La primera clàssica que volem ser la CentOS, volem posar l'upstream download de la última versió de relax i recuper, perquè la CentOS, en el repos, té versió 1.17. DLM, però, en l'última versió, té més progressió i soluciona molts bugs, i és millor anar a l'última versió sempre. Amb Ubuntu, volem especificar l'ús amb drets d'admin per fer la instal·lació. Per default, el root d'ús és utilitzat en la primera edició, però, en l'última, volem posar l'ús amb drets d'admin per instal·lar-lo i posar la clàssica a drets d'admin. D'acord? D'acord? D'acord? D'acord? No, volem posar la clàssica de la CentOS, amb drets d'admin, però volem ser aquí. A la primera versió de drets d'admin, volem posar l'adreç de mac, l'adreç de l'IP i moltes coses que no, volem agafar la clàssica si la clàssica és online i aviat. No volem posar la clàssica del cos de la clàssica, perquè no tenim d'inservenció d'inservenció. Però, si tens els clients registrats en dins, volem just posar l'adreç de l'IP i el IP d'adreç amb l'IDR NETMASC, i el telèfon va dir què es gaudirà i si és bé, cal confirmar-lo. I el client... No, és already in the alarm database i ara és començar la instal·lació en fase del client. Just confirm the SHHK, now the root password the sentos box because we didn't specify the user. Okay? And now we will set the default password of the DLM user. This is the user that DLM used to connect to each client in order to run the relax and recover backups. This password doesn't matter because after the installation it will be locked and only DLM server will be able to connect with this user to the relax and recover servers. Okay? Well, now it's installing all dependencies and configuring the sentos box. Let me show you. It's not yet installed. Okay? Now, as we can see, the client is in DLM database and if we want... no, here, it's installed and all the required environment settings and dependencies are installed also on this client and we will be ready to backup and recover it after that. Okay? Can I ask a question? Un package who creates a user or is it the server who does over SSH a connection and does the add user commands? Ah, it's the server that connects to the client and does it over the network with SSH, okay? On the server, all the commands are sent from SSH connection and all the steps are being done from the DLM server. Okay? Automatically. Well, as you can see now, we have the relax and recover install and as you can see, the only configuration that is needed here is managed by DLM because the configuration is on the server side. Okay? I will show you now. Okay? This is a standard relax and recover configuration that is in the server side and at runtime, when the DLM calls relax and recover to do the backup the configuration is downloaded and it will make the backup. You just need to maintain the configuration from one location not on each of the servers of your network. Okay? What? I mean, is this the DLM cluster in plain text? What? On the SSH? Yeah, it's the password of the recovery match. It's just the password of the DLM user in the server and it's just the password when you have the recovery match up and running in order to recover it you will, if you want to connect to our SSH to the recovery match when it has the network up you just need to set the root password is DLM but just on the recovery time without any system the plain password of the root user in the system itself is just in the recovery match, okay? It's not the password in the system it's just for booting and recovering the system it's just for that. Okay? Well, I will show you also that after the Client Edition all the DHCP configurations that you usually need to set up manually are done on DLM site and it's, after doing a backup of this Client it will be ready to boot over the network and recover it from DLM, okay? And also, as I said, it uses NFS for exporting the match and we will exporting the store location of the mounted DRMatch when it will be done and it takes care about all these configuration files and all these services for you in order to have a healthy and maintain the configurations of your needed disaster recovery services, okay? Well, let me add the other two clients the one, and the same, I put the hostname because we don't have DNS, just the IP address and the CADR and I said here that it will install it with the user background, okay? Just the confirmation, the background password, the DLM password, that we'll be looking after and one moment to wait it and for it and it will be installed also. So for a boot to the packages already in the default of stores? Yes. In Ubuntu and Debian, there is no relax and recover package yet in the repos. It will be soon, but we have pre-configurated variables in DLM to get the correct package from upstream download page of relax and recover in order to install it on these boxes. Since version 1.18, if I don't remember, okay? Good. It was a problem with dependencies with Choriso and that was not previously with older Debian releases and we forced it to use relax and recover version 1.18 in order to don't have any problems with dependencies, okay? It's not the last, but it's almost the last. At this time, it's the last that it's working by default. Okay? Well, I will add now the last client. Okay? Here I just add this client to database. It's on DLM, but it's not installed yet. Okay? No, we had the three clients in DLM database, but the Debian one, it's not installed yet. I will install it with the client command that is what a client by default now does if you specify that it needs, you will install it at the same time, okay? But just to know that there is another option, you can add the client to DLM if they are offline and after the systems are up and configured with the network, et cetera, you can install them from DLM after. Okay? Also do some steps already at the client side. For instance, if you like use some configuration management system that you... They are defined, I need to have these packages installed. Can you do it also in that way? So instead of, I think, the master server push everything or initiate all the commands, can you let the clients already do most of the work in advance? You can do it if you want with any configuration management system, you can prepare the clients to have all the configuration set. But if you don't have it, DLM, do it by yourself, okay? There are no problems with overwriting? No. If the packages are installed and configured, DLM, just try to install it already there, there is no problem. And also the DLM user, if something happens with SSH, the DLM user will be recreated because you must send the SSHK from the DLM server to the client in order to log the password of that, could be done many times if you want, okay? Well, now we have the three clients installed and properly configured on DLM with our networks, okay? And well, we will start now to plan some backups of the DLM server, okay? Let me adjust the times, okay? No, we added a backup job in one minute, each five minutes will be done and we will try to start backups again in 12.11, okay? It's just the scheduler and we can show you that the job is here and now it's running, okay? The backup is running. You can type DLM RAM backup with the client ID or the client name at any time if you want, but with the scheduler, by yourself in a planet manner if you want to keep a monthly, or backups of your clients, okay? Let me add another jobs here, okay? No, we planned another backup job without end date, it will continue backing up always until we drop that job from DLM and the last one will be a one-shot backup and well, we have the planned scheduler backups here, each five minutes, usually you will use daily, weekly or monthly basis for repeat, but here for testing and playing with it we set to five minutes. If any of these schedules found that there is the backup of the client is trying to backup again, it's running, it just detects it and stops trying to do that backup, okay? There will be no conflicts on that. You can have many, many jobs planned with the same client and if any one of them have a collision sometime, there's no problem. The first one is the correct backup and the next execution just skips the backup if found that there is already running, okay? And well, we just planned, we've seen that there are some jobs running here, okay? Here are the three client running backups now at this time. We don't have any backup yet, finished, okay? If we go to the client, it's running here, okay? Well, I missed to show one slide previously, but it's how all the jobs are configured at the RLM and well, already running, okay? And well, we have here the two backups have finished and we'll continue to be planned after that and in a few minutes we will see that there are more backups here in RLM. By default, RLM only keeps at this, the last two backups of your clients, okay? You can configure it to increase it, but you must have enough space on file system to save them. It's a good practice to each week or you have your backups weekly or monthly after that to send those backup files, just a recovery match to tape with a standard backup tool. But the two lasters by default will be ready to recover from the RLM server with no needs to go to the tape and recover the files in order to recover your systems, okay? Because first of all, before recovering any data, you need a operating system running, okay? And this is the concept and the main target of the RLM in order to have your operating systems ready to recover environment, okay? I was not here for the previous talk because probably that was about the rear. Can you very shortly explain like is the rear using our sync or it creates an image? Okay, well, let me show with... Okay, as you can see here, relax and recover supports different protocols to backup your systems. By default, use a ter backup. You can also configure it to send the backups directly to veros or other backup tools. We use NFS and default ter backup in order to store the images on the RLM, okay? These are some of the configurations that you need to do. It's while explaining relax and recover the quantization pages. But, by default, this should work. If you want to exclude some file systems or just backing up the root BG of your operating system because it's the main operating system itself and get... leave the backup data to another tools, you just need to extend the configuration on the RLM server in order to do it, okay? Let me show that at this time we have three mounted RLM images of each one of our clients and inside we'll see two folders, let me do this, okay? We'll have the backup file of ter, compress it and the kernel needs the files on the px folder to be able to boot from... over the network, okay? And it's already reconfigured in order to boot from the network and recover the system with the needs of... doing anything in the RLM. Okay, let me see if there are more backups. No, not yet. Okay, there are two running. Okay. Those who planned previously in the scheduler and now we will show you... I will show you how to configure the error... mail error reporting in order to... to send the... the errors to... by mailing that case because it's a busy way. I experienced some problems with internet connection previously to send outside from this network but you can set it to... configure it to send to the... root mail account of the server, okay? Let me... show that the configuration is already done because I prepared previously. Okay? This will be the configuration to send to an external mail server with my mail and my password is in that file. I don't want to show you my personal password but this works to... to send the mail. We will force to fail now a backup and show that the... it's sending to... to the... to the root localhost account now because I... I configure it... it for that. Okay. Run... Okay, it's... it's running yet. Okay, we... with some backups enabled by default it's enabled the last backup on the RLM server but if you experience any problem with the last backup and say to the RLM to enable a previously a good backup that was on... in disk and you can try to recover from that. Okay? Okay, it's still running from backups. I need to wait because the configuration was... no, I have the configuration already done. I can try to... RLM... RLM CentOS... The backup of the Debian client or... No, it's... CentOS backup, okay? Okay, I think I missed to configure something. RLM... Yes. Of course, I never... I never will receive any mail because... Okay? With a command line, but... I didn't write. Ubuntu, it's running. Still running. No, it's finished. RLM... Okay, I will... True phones in the backup status? Yes, it's a... we need to change it because it's a... if the backup is enabled or not, okay? It's the status of the backup, but... it's not correctly splined in the... in the output. No, I will try to... Stop. I will try to run a... Okay, CentOS... And it's... The machine is... I think the machine is stopped yet or if not, I will stop now. It's running. Let me... Stop it. The backup will file and I will receive it. All goes okay. I will receive a mail from... from RLM. No, it's... Just exporting the... It's a store, okay? It's the directory where the backup is mounted. Okay? It just... The... The just-regover image, it's great because we need to export it again after mounting. No. It's exported... The directory is exported to each client, okay? Just the... It's exported the... the image of each client to the client. It's not a... a good direct file system open to anything. We take some care about security and we only export the data image to the... to the client that needs. In the meantime, I will connect to the... Okay, we have mounted this... You want to backup? Okay. Okay, as we can see here the backup I have enabled is... this one, okay? And now we will this other one. Oh, maybe the Ubuntu one is... It's full. No, I think it's... because the previous backup and now it's running Ubuntu. I configured the... the job scheduler with... to continue planning and planning and planning jobs. And the last... the previous one was dropped and the new one it's... it's there, okay? And... let me check it. I don't know yet but... or maybe I missed something in the... No, it's... Okay, I... Okay. I copied the wrong backup ID. Okay, now, as we can see the backup I've enabled it's the... the previous one. Okay. Now, enabled. And the other one it's disabled, okay? Okay, thank you. Let me show briefly because I have no time to... to... I... we can see here we... the nine. Okay, the alert that there was a problem with the backup we received the... the alert by mail. And this will be the same for for other monitoring tools like Nages, Xevix, or... Anywhere, if... that are supported by DRLAM, okay? DRLAM... I will show you how to export an image and... and recover it from the other server, okay? Okay. Okay. Now we are sending the exported backup to the... destination DRLAM server, the devian one. As you can see I already configured the networks and... and the client to be imported in the... in this... in this server. Okay. It's understanding something. When you turn off the... you go into image that you already backup I was expecting to be launched automatically by the D1 of the DRLAM. Sorry. I mean, when you turn off the... you go into image I was expecting the D1 of your tool to automatically launch the backup to deploy it. No, you can enable or disable the... it's for the backups already done, okay? You can... disable the last backup and enable a previous one, but it's not the... it will not start again a new backup. You can start the backup with another command, the RAM backup command, okay? In a situation you like every day the backup, is it incremental incremental? No, each time it's a full backup. It's a full image, okay? We will... we hope that we will support incremental, something like that in the future, but at this time it's full backup with... all the... it's... each time with each recovery match, okay? At the time of that backup. You mentioned there was integration with some backup tools. What kind of integration? Relax and Recovery has integration with some backup tools like Veros and others. DLM just takes care of the management of Relax and Recovery deployment and when after... after you have the unit matches in DLM you can use Veros or any backup tool to send to... those matches to the tape, but having a... ready match to recover from DLM server, okay, over the network. So if there is no like integration that the client would use Veros or something else... No, we stored the backup on the DLM match because it... no, we are exporting one sending to another DLM server in an easy way and after importing it I will show you to recover from the new one, okay? It's... okay? It's transferred. No, I'm importing the... importing the... the received backup from the other DLM server. I will stop the previous one to be sure that... here we have the... the same networks between the servers and it's not a typical use case because we usually send from one side to another side with different networks and different network configurations, but in order to the PXS services doesn't have conflict we have... I put off. No, as we can see, the client the backup is also here and ready to recover, okay? No, let's... let me show you how to... setting... no, not yet. Hold off. Okay. We need to disable the NAT N80 device because... virtual box only boots from the first available adapter from PXE and set the network to boot in the... machine and well... if... if it's working... okay? Let me... We have the DLM boot menu in order to... we will start loading the kernel and enter the image from the DLM server Okay. I hope that I could recover in time. Okay. We don't have any client. Okay. We are gathering the network from the DLM server. Okay. We need to specify the server and the... idea of the... of the client. And well, it's... it's pure relax and recover now. It's just... integrating the layout and recovering now the backup image from the new DLM server and... I think it will be ready to... to show you that the... system was recovered. It's okay. Okay. When you... when you connected to these consoles to the recovery version, it didn't ask you for the password that was in the... if you have the network available in the recovery match here, you can connect over the network and run rear recover. Okay. It will ask for the password for SSH. No. We don't need to change anything in the system because it's the same machine with the same network configuration. Power off. System. Hard drive again. Adjust the network. And it will start again with... the system recovered. Okay. That's all for the recovery. These are the slides to... of what we don't previously, the recover slides because I don't have time to explain it because we are... the time is over. And some best practice that we... we can use with DLM keep the arch there when the DLM is restored at the file system with enough space. Adjust the kernel loop limit to the amount of lines you want to manage. Please use it always in a dedicated server with a proper data and time configured because the server if not doesn't work. Avoid using it in networks with another page of services on it because they could experience problems. And the minimum amount of RAM for DLM should be 4 gigabytes for a production site with some clients managing some disaster recovery or clients for disaster recovery. Okay. And well, I don't know that if we have time for more questions or answers but anyway, we will be here all the whole day and later in the last talk of this depth room and well, hope that this workshop and the talk helped you to know more about disaster recovery management and encourage you to use it and try it with the workshop at home or at work and use it, please. And well, thank you very much for coming. We have time for any question? Yeah. At what scale have you seen this being run? Like how many machines have we... We have a customer with... Ah, yes. Yes, we have a customer using DLM in production and with 4 gigabytes of RAM on DLM server it's managing over 100 clients easily from... It always depends on network throughput your CPU power and your memory to be able to manage more and more clients from one DLM server, ok? We have mentioned the 4 distributions that are supported at the moment and a couple of distributions that look like them but a lot of servers start like... For example, after a couple of years of changes everywhere it looks more like an alabit system how would that work? Would it still be... I don't understand your question, sorry? Well, the distribution that it will serve it's no longer... Well, it started as a... a random distribution but had a lot of changes configuration files are now in other places a lot of software installed from SOAR instead of packages it's no more real reted for example, would that work? This is... the... the... to be able to recover a system with... build software from SOAR and with different configurations the better option is to use relax and recover because it recovers all the data and files whenever they were before the... at backup time, ok? There is no re-installation of packages recovers, regreta-li out of your system and recovers it to the... to the system, ok? It's a... also for migrations between... between different supervisors as you are taking the backup inside the virtual machine it doesn't matter if you have a VMware or KVM you just back it or physical server it's... it could be done because the... you are backing up the operating system inside the VM and you can recover it wherever you want, ok? Ok, time is over ok, please? Given that it's exclusively managing the PXE environment it doesn't sit well out of the box with bare metal deployments so are there any sort of practices for initial deployment of a packet to deploy a base in-stool backup that you've taken to another client, say, that sort of thing? Yes, you can... you can take the backup of one... of one of your servers and use it as a template of... of other servers to... to perform a bare metal install from... from the RLM you just need to export the image and import to a new client in the same server you could be done also but later you need to adjust it, ok? Sorry, time is over we will be here all the whole day if you have more questions or something we will be pleasure to ask it, ok? To answer it Thank you