Xen stepbystep Tomas Karpati Preface
Contents
1. Xen CONFIG XEN PRIVILEGED GUEST is not set CONFIG XEN PHYSDEV_ ACCESS is not set CONFIG _XEN_SCRUB_PAGES y CONFIG XEN NETDEV_FRONTEND y CONFIG XEN _BLKDEV_FRONTEND y CONFIG NO IDLE HZ y CONFIG FOREIGN PAGES is not set CONFIG NETDEVICES y 18 21 Xen step by step Tomas Karpati By default the only priviledged domain that has access to all the physical devices of the machine is domainO Running a second priviledged domain will behave almost identical to an unpriviledged domain unless we make some changes to the domains configuration First if we want to asign a physical device to a domain that requires a soundcard we have to hide this device to be recognized by domainO To achieve this we modify the grub configuration as follows title Xen 2 0 XenoLinux 2 4 27 kernel boot xen gz dom0_mem 65536 physdev_dom0_hide 00 17 5 module boot vmlinuz 2 4 27 xen0 root dev hdal console tty0 title Xen 2 0 XenoLinux 2 6 9 kernel boot xen gz dom0_mem 65536 physdev_dom0_hide 00 17 5 module boot vmlinuz 2 6 9 xen0 root dev hdal console tty0 To know the address of the device you want to hide you can take a look at the pci process using cat proc pci In my machine I get the following information Bus 0 device 17 function 5 Multimedia audio controller VIA Technologies Inc VT8233 A 8235 AC97 Audio Controller rev 64 IRQ 10 I O at 0xc000 OxcOff The values of the device to be written in t2. id vifl mac aa 00 00 00 22 22 device vbd uname file opt mailserver img dev hdal mode w device vbd uname file opt ms_swap img dev hda2 mode w Files ending with xsp are considered as the native configuration files in Zen The first line of the configuration tell xm that we are defining a virtual machine The second line define the name that will be used to call our server The third line define the amount of memory that Xen will assign to this machine We assigned to mailserver 128MB of RAM Lines 4 11 represent the equivalent of the grub definition of the booting parameters We tell xm to use the kernel found in boot vmlinuz 2 4 27 xenU If you had checked the content of boot you may have seen two 11 21 Xen step by step Tomas Karpati versions of the Xen kernels vmlinuz 2 4 27 xenO and vmlinuz 2 4 27 xenU The first one correspond to the kernel used for running priviledged domains while the second is for unpriviledged domains We will talk about both types later By now I only will tell you that we will run an unpriviledged domain We also define here the network interface with its address and the root directory We can define anything as our root directory for example dev hdb4 dev sda2 dev srO etc as long as in the next lines we tell xm where to find that device as we do in lines 13 14 We designed dev hdal to a virtual file that contains the domain linux distribution opt mailserve
3. engine running in the base system which permits running unmodified operating systems on it This model is based on full virtualization In contrast Xen is based on a slight modification of the base operating system to achieve better performance and need that the virtualized systems run also a modified operating systems However the user s applications run unmodified on those systems This approach has been called paravirtualization I think of a machine running Xen as the hardware equivalent of partitioning a hardisk When we make a disk partition we virtualy divide one physical disk into many disks If we look at a partitioned disk he have the impression of having many disks each independent of the other and even formated with different filesystems that will contain different operating system files In the same way Xen divide the computer into many virtual machines giving the impression of many computers running on it If we run a network scanning program we can find that each virtual server is recognized as a different and independent machine One of the most powerfull characteristics of Xen is compatmentalization Xen isolates each of the virtual machines running on it and give them the posibility to control their own real resources besides of the virtualized ones For example if you have two network interfaces you can asign one to Xen that will be used to connect all the virtual machines with themselves and those with other real m
4. mailserver 168 168 3 smoothwall In etc resolv conf copy the same configuration values than in domaino In etc network interfaces you specify the network interface passed on by domaino Our system recognize this interface as eth0 and the driver loaded is the same driver that we loaded in domaino auto lo eth0 iface lo inet loopback iface eth0 inet static address 192 168 168 2 netmask 255 255 255 0 10 21 Xen step by step Tomas Karpati The next file to modify is etc apt sources list Again we copy the same values as used in domaino The last file we will modify is etc fstab This file will reflect the mounting points of the system as passed by domainO we will define it in the next section dev hdal ext2 errors remount ro dev hda2 none swap sw proc proc proc defaults Logout from the chrooted directory In order to boot this distribution without problems we have to copy to the system the kernel modules for the xenized kernels cp rp lib modules mnt lib modules umount mnt Now we are ready with our first independent virtual machine To run it we have to create a configuration file We will call this configuration file mailserver xsp and save it on etc xen The content of this file looks as follows vm name mailserver memory 128 image linux kernel boot vmlinuz 2 4 27 xenU ip mailserver eth0 192 168 168 2 root dev hdal args ro fastboot 2 device vif
5. system The first step is to create a fresh pristine linux distribution We will base our pristine distribution on Debian Sarge You can use the technique you preferre but we recommend the use of debootstrap You can install debootstrap by downloading it from http people debian org blade install debootstrap To install debian for our system you need a partition of at least 1 5GB Format this partition be carefull to check if there is some important data on this partition before proceding using mke2fs Now mount the formatted partition as following mount dev hdaX mnt t ext2 where hdaX is the name of the formatted partition substitute it to the correct values If you have not installed linux on your test computer you can run linux from your cdrom drive using any live distribution like Knoppix install debootstrap locally in your home directory that is saved into your C drive as an image file or on other media When installing debootstrap you are asked for a pass prase Write the following phrase as the password Yes I know that this data may be harmful Be carefull to write all the words spaces and signs as written here any change you made will return an incorrect pass phrase error After that extract and install debootstrap Creating a pristine debian distribution Now we will procede to install a base Debian system on mnt debootstrap arch i386 sarge mnt http ftp lt countrycode gt debian
6. topology change detected propagating 17 21 Xen step by step Tomas Karpati This bridge network device will be recognized in other virtual machine and will be assigned the network interface ethO The kernel dmesg log for this interface looks like this Universal TUN TAP device driver 1 5 C 1999 2002 Maxim Krasnyansky XEN Initialising virtual ethernet driver XEN Netfront recovered tx 0 rxfree 0 NET4 Linux TCP IP 1 0 for NET4 0 IP Protocols ICMP UDP TCP IP routing cache hash table of 512 buckets 4Kbytes TCP Hash tables configured established 4096 bind 8192 Sending DHCP requests OK IP Config Got DHCP answer from 0 0 0 0 my address is 192 168 168 252 IP Config Complete device eth0 addr 192 168 168 252 mask 255 255 255 0 gw 192 168 168 230 host domain2 domain nis domain none bootserver 0 0 0 0 rootserver 0 0 0 0 rootpath ip_conntrack version 2 1 512 buckets 4096 max 288 bytes per conntrack In this case vif1 0 is the backend of the network interface which is attached to ethO in domainoO while ethO running in domain2 represent the forntend interface The kernel configuration that define a priviledged domain are Xen CONFIG XEN PRIVILEGED _GUEST y CONFIG XEN PHYSDEV_ACCESS y CONFIG XEN SCRUB_PAGES y CONFIG XEN NETDEV_FRONTEND y CONFIG XEN BLKDEV_FRONTEND y CONFIG NO IDLE HZ y CONFIG FOREIGN PAGES y CONFIG NETDEVICES is not set while for an unpriviledged domain are
7. Xen step by step Tomas Karpati xen step by step by Tomas Karpati Preface Why Xen Through the last years I have been playing around with many virtual engines There are many of them each with its own advantages and disadvantages There are very good virtualization software developed as commercial products and as open source I have tested VMWare in its very initial versions dosemu Bochs User Mode Linux UML and Xen To be faire I think that all of them are good However I don t think that all them are suitable for same projects Before a month and half I had been involved in a new project that demmanded a very specific configuration with specialized servers The most logical solution was to setup three or four machines each running safely isolated in an intranet and protected with a firewall The problem was that the budget was scare but the bandwich and demmanding usage of the system did not justify the high costs it may carry In this context I begun to think on the use of one machine running many virtual machines each independent from the others with their own resources managed independently At the beginning I planned to use UML as it was the one I thought may comply with most of the requirements However its use falled down because I needed that some of the virtual servers had direct access to some hardware components while those components where not necessary for the other servers I searched in the UML discussion gr
8. achines in your intranet The other interface could be asigned to another virtual machine that may act as a router that securely connect to the internet your whole intranet network Another example of isolation is the use of two graphic cards two usb mouses and two usb keyboard asigned to two independent virtual machines each running X with its own set of harware When running Xen we actualy run at least two kernels The first one is xen gz which is the kernel implementation of the virtualization engine The other one is the priviledged kernel called umlinuz x x x xenO This is the kernel that will control most of the hardware generating virtual frontends of them for the other virtual machines This kernel will run the linux distribution called domainO Under this domain Xen 2 21 Xen step by step Tomas Karpati has to run xend the Xen control daemon xend is the responsible of creating and destroying domains It also is the responsible of managing the resources of those domains xend is writen in the python scripting language and its communication signals are managed through the Twisted network framework also based on python Those signals are transmitted using the HTTP protocol The HTTP interface is present in all the most commonly used programming languages Python Perl Java C C Rubi etc letting programmers to use their preferred language for interfacing with Xen Preparing Xen to run For running Xen we have to prepare our
9. cial refrigeration meassures without a hole in your pocket and without your wife reclaiming your computers are invading your home Customizing Xen Until now we have been running Xen in a very basic way Xen could be used with more specialized and cumtomized configurations Lets say you want to create a Xen based system that will include a router firewall virtual machine based on Smoothwall a LAMP system Linux Apache MySQL and PHP4 a mail gateway Postfix Amavis with antivirus ClamAV and antispam SpamAssassin protection and an internet radio server Icecast Such a system can not be ran successfully on the Xen configuration described before This configuration requires some of the systems to have direct access to some hardware However all the hardware is only available to domainO In the following diagram we show the conguration we are trying to achieve 14 21 Xen step by step Tomas Karpati Ss Hub intranet Virtual Ethernet Virtual Ethernet card vif0 1 card eth0 Real Ethernet card eth0 domainO A ri Virtual Ethernet Virtual Ethernet Gee i card eth0 card eth0 green interface Real Ethernet card eth1 red interface Smoothwall iRadio The computer used for running this Xen example has the following configuration AMD Athlon 1900 1600MHz processor 1GB SDRAM Asus AT7 motherboard w Realtek 8139 network card and Via VT8233A AC97 Audio Con
10. dropped 0 overruns 0 frame 0 TX packets 13504 errors 0 dropped 0 overruns 0 carrier 0 collisions 0 txqueuelen 1000 RX bytes 8909969 8 4 Mb TX bytes 1954311 1 8 Mb Interrupt 11 Base address 0xa800 Link encap Ethernet HWaddr 00 50 8D A6 10 19 inet addr 192 168 168 254 Bcast 192 168 168 255 Mask 255 255 255 255 UP BROADCAST RUNNING MULTICAST MTU 1500 Metric 1 RX packets 94 errors 0 dropped 0 overruns 0 frame 0 TX packets 20 errors 0 dropped 0 overruns 0 carrier 0 collisions 0 txqueuelen 0 RX bytes 16360 15 9 KiB TX bytes 1313 1 2 KiB After xend is running we are now able to run the virtual machines We can communicate with xend through the command xm This command passes the instructions we have give to it as parameters 9 21 Xen step by step Tomas Karpati xm create f etc xen xmexamplel will create the virtual machine as described in the configuration file etc xen xmexamplel However to run a new virtrual machine we have to configure a new pristine distribution We will copy our previously generated template into the file opt domainl cp opt vserver img opt domainl img cp opt swap img opt ms_swap img mount opt domainl img mnt o loop chroot mnt bin bash We will configure now domain in a similar way we configured domainO We also created a swap file for this domain echo mailserver gt etc hostname In etc hosts we write the following 0 0 1 localhost 168 168 1 domain0 168 168 2
11. er for security reasons it is recommended to logging as this unpriviledged user and not as root Use the command adduser and follow the instructions If you are a curious you can now check the system which processes are running which devices are recognized check the log files etc You will note that it runs like any regular distribution Take a look at the memory statistics cat proc memstat you will note that the reported memory is not the real memory of the system It is the memory we assigned to domainO when grub was configured Everithing looks wonderfull but we are interested to run more virtrual machines At this point we cannot run any other virtual machine We have before to run the Xen daemon etc init rd xend start If you check now the network interfaces ifconfig you will find that there is a new network interfaces added domain0 gt sbin ifconfig lo Link encap Local Loopback inet addr 127 0 0 1 Mask 255 0 0 0 inet6 addr 1 128 Scope Host UP LOOPBACK RUNNING MTU 16436 Metric 1 RX packets 1002 errors 0 dropped 0 overruns 0 frame 0 TX packets 1002 errors 0 dropped 0 overruns 0 carrier 0 collisions 0 txqueuelen 0 RX bytes 72584 70 8 Kb TX bytes 72584 70 8 Kb Link encap Ethernet HWaddr 00 50 8D A6 10 19 inet addr 192 168 168 254 Bcast 192 168 168 255 Mask 255 255 255 0 inet6 addr fe80 250 8dff fea6 1019 64 Scope Link UP BROADCAST NOTRAILERS RUNNING MULTICAST MTU 1500 Metric 1 RX packets 13283 errors 0
12. he grub configuration should be in the form bus dev func The values shown in the pci info are in decimal numbers You can define it using decimal or octal values If we define in decimal values it will be 0 17 5 while if we define it as octal it will be 0xO 0x11 0x5 After changing the grub configuration we have also change the configuration for the virtual machine 19 21 Xen step by step Tomas Karpati vm name asterisk memory 80 image linux kernel boot vmlinuz 2 4 27 xen0 ip iradio eth0 192 168 168 4 root dev hda2 args ro fastboot 2 device vif id vifl mac aa 00 00 00 00 11 device vbd uname file opt iradio img dev hda2 mode w device pci bus 0 dev 17 func 5 In the configuration bellow we defined the pci device that should be passed to that virtual machine Another differece to the configurations we made at now is the use of the vmlinuz 2 4 27 xenO kernel This kernel defines the virtual machine as a priviledged domain This example shows how Xen use isolation to run customized virtual machines which is one of the most important differeces between Xen and other virtualization engines 20 21
13. nd never restart never Xen step by step Tomas Karpati This file is self explained and may be more clearly understood if compared in parallel with its XSP counterpart After we have our disk image and its configuration ready we will run our new virtual machine Try this xm create F etc xen mailserver xsp c or this xm create f etc xen mailserver conf c the F option tells xm we are using the native configuration format while the f option tells xm we are using an alternative format The c option tells xm to assign the actual console to the new virtual machine You will see now an almost normal linux kernel booting that will get you to the login command prompt Type root and you are in The first thing to do is to add a root password and to create a new unpriviledged user as done bellow in domaino Now you can play a little with the new system and install the server applications you want it to run I recommend you to start with a mail gateway with antivirus and antispam protection an open LDAP server and an NFS server The same steps should be used to generate domain2 installing on it a web server like apache with perl and or php4 a database server like MySQL or Postgress and so on CONGRATULATIONS You are now running your own super ultra fantastic virtual server system And you have all the servers you ever dreamed without consuming all that power a network of servers consume without the need of spe
14. omain you want to create or you can set the parameters for the domain on the xm command line Kernel image file kernel boot vmlinuz 2 4 27 xenU Initial memory allocation in megabytes for the new domain memory 128 A name for your domain All domains must have different names name mailserver Define network interfaces ip 192 168 168 2 Optionally define mac and or bridge for the network interfaces Random MACs are assigned if not given vif mac aa 00 00 00 22 22 bridge xen br0 Define the disk devices you want the domain to have access to what you want them accessible as Each disk entry is of the form phy UNAME DEV MODE where UNAME is the device DEV is the device name the domain will see and MODE is r for read only w for read write disk phy hdal hdal r disk file opt mailserver img hdal w file opt ms_swap img hda2 w Set the kernel command line for the new domain You only need to define the IP parameters and hostname if the domain s IP config doesn t e g in ifcfg eth0O or via DHCP You can use extra to set the runlevel and custom environment variables used by custom re scripts e g VMID usr root dev hdal ro Sets runlevel 2 extra 2 Set according to whether you want the domain restarted when it exits The default is onreboot which restarts the domain when it shuts down with exit code reboot Other values are always a
15. onfiguration We begin with domainO On the command line write 4 21 Xen step by step Tomas Karpati chroot mnt bin bash We have to change the following files echo vserver gt etc hostname In etc hosts we write the following 0 0 1 localhost 168 168 1 domain0 168 168 2 mailserver 168 168 3 smoothwall In etc resolv conf we add the address of our DNS server ask your internet service provider for the address of their DNS servers In my case it looks like nameserver 194 90 1 5 nameserver 192 115 106 10 In etc network interfaces you must specify the network interface existent in your system and the method for obtaining their addresses In our example we present two ethernet interfaces one configured to obtaine its address from a dhcp server the second with a static address auto lo eth0 ethl iface lo inet loopback iface eth0 inet dhcp hostname vserver iface ethl inet static address 10 1 0 2 netmask 255 255 255 0 The next file to modify is etc apt sources list We add here the address of the Debian mirror we used with debootstrap 5 21 Xen step by step Tomas Karpati deb http ftp lt countrycode gt debian org debian sarge main contrib non free The last file we will modify is etc fstab This file must reflect the mounting points of our system dev hdal ext2 errors remount ro dev hda2 none swap sw proc proc proc defaults Substitute dev hdal and dev hda2 for
16. org debian 3 21 Xen step by step Tomas Karpati where lt countrycode gt is the two letters code of your country or in defect of the nearest country having a Debian mirror The installation of Debian will take a while thus enjoy your time reading this book or playing debootstrap will make the work for you After debootstrap finished without errors logout from the chrooted archive logout At this point we have created a distribution that we will use as our template Before configuring it for using as our domainO we can generate a filesystem on a file and copy the distribution to it so we can use it for the creation of our virtual machines To do that we procede as follows dd if dev zero of opt vserver img bs 1024k count 1 5MB mke2fs opt vserver img dd if dev zero of opt swap img bs 1024k count 256MB mkswap opt swap img Now change to root and procede with the following su root mount opt vserver img cdrom o loop ed mnt cp rp boot bin dev etc lib root sbin usr var cdrom cd cdrom mkdir cdrom floppy home mnt opt proc tmp umount cdrom Now we have two copies of the filesystem one on a hardisk partition and the second on a virtual filesystem I recommend you to use gzip or bzip2 to compress the vserver img and record it into a cdrom so you can always have a copy of it in case you want to create a new virtual server on another machine Configuring domainO We are now ready to procede with the system c
17. oups for someone who had tried such setup and found a comment that said it was possible but needed some hack on the kernel As I had not as much time nor the expertice to do that I begin my search for an alternative I rapidly found the link to the Xen homepage I heard of Xen before but never tried it When reading the user manual and the latests threads on the discussion group I found that Xen was the perfect match for my project So I give it a try The beginning was a little troublesome but relatively fastly in about a month and half I had all my project completted I have to thank the developers of Xen for the well done piece of gold and the people at the discussion group which are very active and helpful responding to the questions posted even by newbees The only problem I 1 21 Xen step by step Tomas Karpati saw was the lack of documentation and examples for those who want to try Xen but without having to get into the intrincated source code So I dedicate this guide for the newbees wanting to play with Xen This guide is dedicated to the Xen newbee and to those who want to get into a production ready system without the pains of learning the intrincated code and obscure theories behind Xen However the reader has to have some degree of experience in working and maintainig linux to successfully get Xen running An Introduction to Xen The model of virtualization used in most of projects are based on creating a virtualization
18. r img In the same way we assigned dev hda2 to the swap file take now a look at the etc fstab we created for this distribution We may also define here a real harddisk partition like in this example vm name mailserver memory 128 image linux kernel boot vmlinuz 2 4 27 xenU ip mailserver eth0 192 168 168 2 root dev hdal args ro fastboot 2 device vif id vifl mac aa 00 00 00 22 22 device vbd uname phy hda4 dev hdal mode w where we assign a real physical device dev hda4 phy hda4 that will be recognized by the target virtual operating system as dev hdal WARNING Be carefull to not assign block devices in use by other domains to any new domain because this may corrupt the filesystem and will cause troubles to the operating systems In line 12 we define the virtual network device to be passed to the virtual kernel The mac argument defines the mac address of the virtual network interface This value has to be different to that of the other interfaces defined in other virtual machines or from real machines in your network There is also another way to write configuration files It is interpreted by a python script that will then pass the parameters to xm An example showing the same parameters as defined bellow is 12 21 Xen step by step Tomas Karpati This script sets the parameters used when a domain is created using xm create You use a separate script for each d
19. red domaino as the only priviledged domain However we can also configure other virtual machines as priviledged domains The other way to run a virtual machine is configuring it as an unpriviledged domain Unpriviledged domains are characterized by having no direct access to the physical devices of the system Their access are made through virtual interfaces managed by domainO Examples of those interfaces are network and block devices As we stated before most of the physical devices are controller by domainO This domain creates a backend port for each device which can communicates with other domains having configuration with a frontend driver This communication is controlled by xend The reason for this model is that any physical device can be controlled only by one kernel driver I we permit many kernels to access and controll the same device misbehaving software may crash or lock all the system and other machines may stop working or responding The recognition of those virtual drivers by the virtual machine kernel is achieved by modifications introduced into the kernel For example when initialising xend in domainO a virtual network interface is generated using bridge utils and attached to the real network device on the system On starting xend we will see in the console the following message device vifl 0 entered promiscuous mode xen br0 port 2 vifl 0 entering learning state xen br0 port 2 vifl 0 entering forwarding state xen br0
20. talled by default in Debian so we install it into Debian apt get install grub this will install grub on the system Now we have to create or modify the file boot grub menu lst title Xen 2 0 XenoLinux 2 4 27 kernel boot xen gz dom0_mem 65536 module boot vmlinuz 2 4 27 xen0 root dev hdal console tty0 title Xen 2 0 XenoLinux 2 6 9 kernel boot xen gz dom0_mem 65536 module boot vmlinuz 2 6 9 xen0 root dev hdal console tty0 This file says grub the sequence of actions and the parameters it has to pass to the kernel on booting For those used to boot with lilo it is the equivalent of lilo conf For the moment Lilo is not supported by Xen In this file we have two booting options botting Xen with a 2 4 27 kernel or booting with a 2 6 9 kernel The first line defines the title of the boot option The second line is identical in both options It says the bootloader to use the file boot xen gz as the booting kernel and to assign to it a total of 64MB of RAM No matter if you have 1000MB RAM you assigned to domainO only 64MB The restant memory may be used later when assigning memory to other virtual servers xen gz is the control layer of Xen and is 7 21 Xen step by step Tomas Karpati first runned The third line is the name of the kernel that will run the virtual machine domainO This is passed by grub to the xen gz kernel with the descriptor module and the parameters for the location of the root directory and the console
21. the Xen utilities Then make will patch your vanilla kernel as needed and will run the following make ARCH xen oldconfig At this point you may be required to answer some questions related to the kernel 16 21 Xen step by step Tomas Karpati configuration For those parameters related to Xen we recommend to accept the defaults For the other parameters it is up to you if you want to make any addition or deletion Now make will procede with the compilation of the linux x y z xenO kernel and modules and later it will repeat the same steps to compile the linux x y z xenU kernel tree At the end of this process we will have new kernels and modules that will be automatically installed into domainO To have our distributions booting without problems we have to copy the modules into them To do that mount the virtrual images we created and copy the modules from usr src xen install install lib modules into mnt lib modules Now we have to reconfigure some of the system configurations but before lets talk a little about priviledged and unpriviledged domains Priviledged vs Unpriviledged domains Virtual machines running under Xen are runned in two ways as priviledged domains or as unpriviledged domains The most common example of a priviledged domain is domainO The term priviledged specify that we are running a domain that has direct access priviledges over part or all the hardware In the examples we have discused previously we have configu
22. the correct values for your system As we explained before DomainO is the responsible of running the Xend server So we have to install the Xen binaries on it Before procedding we have to install some software libraries and applications which are pre requisite for Xen The packages we need are iproute2 bridge utils python Twisted libcurl zlib If you want to build Xen from source you will also need gcc v3 3 x binutils GNU make libcurl devel zlib dev python dev LATEX transfig and tgif You can install them using apt get or dselect I recommend the use of dselect as it can show you the version of the packages their dependencies and the recommended package associated with them 6 21 Xen step by step Tomas Karpati After completing the installation of the required packages we procede to download and install Xen Download the last binary distribution of Xen from the Xen web site www cl cam ac uk Research SRG netos xen downloads html Unpack the package and install cd usr sre untar xzf xen install tgz cd xen 2 0 1 install install sh This will check for the needed dependencies and will install Xen The installed files are the Xen kernels in boot the Xen module in lib modules the Xen tools in usr some python libraries in lib python and in var xen and the Xen scripts that run the Xen virtualization engine in etc In order to run Xen it is needed to boot using grub However grub is not ins
23. to be used for logging in In this case the root directory is dev hdal change it to reflect your system configuration and the console will be dev tty0 If you have installed lilo and don t want to change bootloader you can create a floppy booting disk or even a booting cdrom see the grub manual cd usr share grub i386 pc dd if stagel of dev fd0 bs 512 count 1 dd if stage2 of dev fd0 bs 512 seek 1 To install grub into the master boot record run the following grub install dev hda where dev hda is the master disk WARNING if you use the command bellow without adding the corresponding configuration definitions for the rest of yor system the system will not boot your original operating systems on this machine Please referre to the grub manual We are ready with our installation logout from the chrooted directory and umount the mnt directory Booting Xen for the first time Now we are ready to reboot our system for the first time You can try it by rebooting your computer insert the floppy if you decided to make one If it boots successfully CONGRATULATIONS you are running your first Xen Virtual machine and you are running domainO You have now to login as root Being a fresh install no password will be required This is the moment to create the root password for domain0O Type passw 8 21 Xen step by step Tomas Karpati and follow the instructions You may also want to create a new non priviledged us
24. troller both on board An additional 3Com 905b typhoon ethernet card We have to assign One of the ethernet cards to domainO the other to Smoothwall 15 21 Xen step by step Tomas Karpati and the soundcard to iRadio server However the module drivers corresponding to those hardware are not included in the binary distribution of Xen So we have to customize Xen for our needs Compiling Xen Our first step in the customization of Xen is to download the source code for Xen and a vanilla linux kernel Etp ftp kernel org We will make all this work on domaino Decompress the Xen source into usr src and the linux kernel source into the usr src xen 2 0 1 directory Change the name of the kernel source as following cd usr src xen 2 0 1 mv linux x y z pristine linux x y z Now if you have a configuration file that works for your computer copy that to the boot directory the Xen Makefile will check if the configuration files exist in the boot directory and if true will use those configuration files for compiling the kernels cp usr linux myconfig boot config x y z xen0 Another way to customize your kernel is to get into the vanilla kernel source and run make mrproper make menuconfig after done save the config file and copy it to the boot directory as explained before Now run make and wait for a while First make will check if you has all the required dependencies needed for compilation Next it will compile
Download Pdf Manuals
Related Search