Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.157

1.144     maxv        1: [[!meta title="Xen HowTo"]]
                      2: 
1.157   ! maxv        3: Xen is a Type 1 hypervisor which supports running multiple guest operating
        !             4: systems on a single physical machine. One uses the Xen kernel to control the
        !             5: CPU, memory and console, a dom0 operating system which mediates access to
        !             6: other hardware (e.g., disks, network, USB), and one or more domU operating
        !             7: systems which operate in an unprivileged virtualized environment. IO requests
        !             8: from the domU systems are forwarded by the Xen hypervisor to the dom0 to be
        !             9: fulfilled.
1.12      gdt        10: 
1.149     maxv       11: Xen supports different styles of guest:
                     12: 
                     13: [[!table data="""
                     14: Style of guest |Supported by NetBSD
1.150     maxv       15: PV             |Yes (dom0, domU)
                     16: HVM            |Yes (domU)
1.149     maxv       17: PVHVM          |No
                     18: PVH            |No
                     19: """]]
                     20: 
                     21: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
                     22: hardware directly, but instead makes hypercalls to the hypervisor; PV
                     23: guests must be specifically coded for Xen. In HVM mode, no guest
                     24: modification is required; however, hardware support is required, such
                     25: as VT-x on Intel CPUs and SVM on AMD CPUs.
                     26: 
1.27      jnemeth    27: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
1.12      gdt        28: The dom0 can start one or more domUs.  (Booting is explained in detail
                     29: in the dom0 section.)
                     30: 
                     31: This HOWTO presumes a basic familiarity with the Xen system
1.123     gdt        32: architecture, with installing NetBSD on i386/amd64 hardware, and with
                     33: installing software from pkgsrc.  See also the [Xen
                     34: website](http://www.xenproject.org/).
1.1       mspo       35: 
1.150     maxv       36: This HOWTO attempts to address both the case of running a NetBSD dom0
                     37: on hardware and running domUs under it (NetBSD and other), and also
                     38: running NetBSD as a domU in a VPS.
                     39: 
1.157   ! maxv       40: Versions and Support
        !            41: ====================
1.15      gdt        42: 
1.111     wiz        43: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
1.15      gdt        44: xenkernel and xentools.  We will refer only to the kernel versions,
                     45: but note that both packages must be installed together and must have
                     46: matching versions.
                     47: 
1.145     maxv       48: Versions available in pkgsrc:
1.85      gdt        49: 
1.145     maxv       50: [[!table data="""
                     51: Xen Version    |Package Name   |Xen CPU Support        |EOL'ed By Upstream
                     52: 4.2            |xenkernel42    |32bit, 64bit           |Yes
                     53: 4.5            |xenkernel45    |64bit                  |Yes
                     54: 4.6            |xenkernel46    |64bit                  |Partially
                     55: 4.8            |xenkernel48    |64bit                  |No
                     56: 4.11           |xenkernel411   |64bit                  |No
                     57: """]]
1.113     gdt        58: 
1.96      gdt        59: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
                     60: 
1.157   ! maxv       61: Multiprocessor (SMP) support in NetBSD differs depending on the domain:
1.145     maxv       62: 
1.157   ! maxv       63: [[!table data="""
        !            64: Domain         |Supports SMP
        !            65: dom0           |No
        !            66: domU           |Yes
        !            67: """]]
1.15      gdt        68: 
1.147     maxv       69: Note: NetBSD support is called XEN3. However, it does support Xen 4,
                     70: because the hypercall interface has remained identical.
                     71: 
1.18      gdt        72: Architecture
                     73: ------------
                     74: 
1.157   ! maxv       75: Xen itself runs on x86_64 hardware.
1.155     maxv       76: 
                     77: The dom0 system, plus each domU, can be either i386PAE or amd64.
                     78: i386 without PAE is not supported.
                     79: 
1.157   ! maxv       80: The standard approach is to use NetBSD/amd64 for the dom0.
        !            81: 
        !            82: To use an i386PAE dom0, one must build or obtain a 64bit Xen kernel and
        !            83: install it on the system.
        !            84: 
        !            85: For domUs, i386PAE is considered as
        !            86: [faster](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html)
        !            87: than amd64.
1.29      gdt        88: 
1.15      gdt        89: NetBSD as a dom0
                     90: ================
                     91: 
                     92: NetBSD can be used as a dom0 and works very well.  The following
                     93: sections address installation, updating NetBSD, and updating Xen.
1.157   ! maxv       94: 
1.19      gdt        95: Note that it doesn't make sense to talk about installing a dom0 OS
                     96: without also installing Xen itself.  We first address installing
                     97: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
                     98: NetBSD install to a dom0 install by just changing the kernel and boot
                     99: configuration.
1.15      gdt       100: 
1.142     gdt       101: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
                    102: 512M it worked reliably.  This does not make sense, but if you see
                    103: "not ELF" after Xen boots, try increasing dom0 RAM.
                    104: 
1.15      gdt       105: Styles of dom0 operation
                    106: ------------------------
                    107: 
                    108: There are two basic ways to use Xen.  The traditional method is for
                    109: the dom0 to do absolutely nothing other than providing support to some
                    110: number of domUs.  Such a system was probably installed for the sole
                    111: purpose of hosting domUs, and sits in a server room on a UPS.
                    112: 
                    113: The other way is to put Xen under a normal-usage computer, so that the
                    114: dom0 is what the computer would have been without Xen, perhaps a
1.157   ! maxv      115: desktop or laptop.  Then, one can run domUs at will.
1.15      gdt       116: 
1.19      gdt       117: Installation of NetBSD
                    118: ----------------------
1.13      gdt       119: 
1.157   ! maxv      120: [Install NetBSD/amd64](/guide/inst/)
1.19      gdt       121: just as you would if you were not using Xen.
1.1       mspo      122: 
1.19      gdt       123: Installation of Xen
                    124: -------------------
1.1       mspo      125: 
1.155     maxv      126: We will consider that you chose to use Xen 4.8, with NetBSD/amd64 as
                    127: dom0. In the dom0, install xenkernel48 and xentools48 from pkgsrc.
                    128: Ensure that your packages are recent.
                    129: 
                    130: Once this is done, install the Xen kernel itself:
                    131: 
                    132: [[!template id=programlisting text="""
                    133: # cp /usr/pkg/xen48-kernel/xen.gz /
                    134: """]]
                    135: 
1.157   ! maxv      136: Then, place a NetBSD XEN3_DOM0 kernel in the `/` directory, copied from
        !           137: `releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz` of a NetBSD build.
1.20      gdt       138: 
1.155     maxv      139: Add a line to /boot.cfg to boot Xen:
1.20      gdt       140: 
1.157   ! maxv      141: [[!template id=filecontent name="/boot.cfg" text="""
1.152     maxv      142: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
                    143: """]]
1.20      gdt       144: 
1.155     maxv      145: which specifies that the dom0 should have 512MB of ram, leaving the rest
                    146: to be allocated for domUs.  To use a serial console, use
1.77      gdt       147: 
1.157   ! maxv      148: [[!template id=filecontent name="/boot.cfg" text="""
1.152     maxv      149: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
                    150: """]]
1.77      gdt       151: 
                    152: which will use the first serial port for Xen (which counts starting
1.139     gson      153: from 1, unlike NetBSD which counts starting from 0), forcing
                    154: speed/parity.  Because the NetBSD command line lacks a
                    155: "console=pc" argument, it will use the default "xencons" console device,
                    156: which directs the console I/O through Xen to the same console device Xen
                    157: itself uses (in this case, the serial port).
                    158: 
1.157   ! maxv      159: In an attempt to add performance, one can also add `dom0_max_vcpus=1 dom0_vcpus_pin`,
1.37      gdt       160: to force only one vcpu to be provided (since NetBSD dom0 can't use
1.155     maxv      161: more) and to pin that vcpu to a physical CPU.
1.20      gdt       162: 
1.93      gdt       163: Xen has [many boot
                    164: options](http://xenbits.xenproject.org/docs/4.5-testing/misc/xen-command-line.html),
1.111     wiz       165: and other than dom0 memory and max_vcpus, they are generally not
1.93      gdt       166: necessary.
                    167: 
1.76      gdt       168: Now, reboot so that you are running a DOM0 kernel under Xen, rather
                    169: than GENERIC without Xen.
                    170: 
1.21      gdt       171: Configuring Xen
                    172: ---------------
                    173: 
1.76      gdt       174: Now, you have a system that will boot Xen and the dom0 kernel, but not
                    175: do anything else special.  Make sure that you have rebooted into Xen.
                    176: There will be no domUs, and none can be started because you still have
1.102     gdt       177: to configure the dom0 daemons.
1.21      gdt       178: 
1.157   ! maxv      179: Since Xen 4.2, the tool which should be used is `xl`.
        !           180: 
        !           181: For 4.1 and up, you should enable `xencommons`:
1.31      gdt       182: 
1.157   ! maxv      183: [[!template id=filecontent name="/etc/rc.conf" text="""
1.152     maxv      184: xencommons=YES
                    185: """]]
1.31      gdt       186: 
1.157   ! maxv      187: Not enabling xencommons will result in a hang; it is necessary to hit `^C` on
        !           188: the console to let the machine finish booting:
        !           189: 
1.31      gdt       190: TODO: Recommend for/against xen-watchdog.
1.27      jnemeth   191: 
1.53      gdt       192: After you have configured the daemons and either started them (in the
1.157   ! maxv      193: order given) or rebooted, use `xl` to inspect Xen's boot messages,
        !           194: available resources, and running domains.  For example:
1.34      gdt       195: 
1.153     maxv      196: [[!template id=programlisting text="""
                    197: # xl dmesg
                    198: ... xen's boot info ...
                    199: # xl info
                    200: ... available memory, etc ...
                    201: # xl list
                    202: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
                    203: Domain-0           0       64    0  r----     58.1
                    204: """]]
                    205: 
                    206: Xen logs will be in /var/log/xen.
1.33      gdt       207: 
1.88      gdt       208: ### Issues with xencommons
                    209: 
1.157   ! maxv      210: `xencommons` starts `xenstored`, which stores data on behalf of dom0 and
1.88      gdt       211: domUs.  It does not currently work to stop and start xenstored.
                    212: Certainly all domUs should be shutdown first, following the sort order
                    213: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
                    214: and is not notified when xenstored exits, leading to not recreating
                    215: the state when the new xenstored starts.  Until there's a mechanism to
                    216: make this work, one should not expect to be able to restart xenstored
                    217: (and thus xencommons).  There is currently no reason to expect that
                    218: this will get fixed any time soon.
                    219: 
1.41      gdt       220: anita (for testing NetBSD)
                    221: --------------------------
                    222: 
1.157   ! maxv      223: With the setup so far (assuming 4.8/xl), one should be able to run
1.82      gdt       224: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
                    225: root, because anita must create a domU):
                    226: 
1.153     maxv      227: [[!template id=programlisting text="""
                    228: anita --vmm=xl test file:///usr/obj/i386/
                    229: """]]
1.82      gdt       230: 
1.40      gdt       231: Xen-specific NetBSD issues
                    232: --------------------------
                    233: 
                    234: There are (at least) two additional things different about NetBSD as a
                    235: dom0 kernel compared to hardware.
                    236: 
1.111     wiz       237: One is that the module ABI is different because some of the #defines
1.109     gdt       238: change, so one must build modules for Xen.  As of netbsd-7, the build
1.157   ! maxv      239: system does this automatically.
1.40      gdt       240: 
                    241: The other difference is that XEN3_DOM0 does not have exactly the same
                    242: options as GENERIC.  While it is debatable whether or not this is a
                    243: bug, users should be aware of this and can simply add missing config
                    244: items if desired.
                    245: 
1.15      gdt       246: Updating NetBSD in a dom0
                    247: -------------------------
                    248: 
                    249: This is just like updating NetBSD on bare hardware, assuming the new
                    250: version supports the version of Xen you are running.  Generally, one
                    251: replaces the kernel and reboots, and then overlays userland binaries
1.157   ! maxv      252: and adjusts `/etc`.
1.15      gdt       253: 
                    254: Note that one must update both the non-Xen kernel typically used for
                    255: rescue purposes and the DOM0 kernel used with Xen.
                    256: 
1.55      gdt       257: Converting from grub to /boot
                    258: -----------------------------
                    259: 
1.157   ! maxv      260: These instructions were used to convert a system from
1.55      gdt       261: grub to /boot.  The system was originally installed in February of
                    262: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
                    263: over time.  Before these commands, it was running NetBSD 6 i386, Xen
                    264: 4.1 and grub, much like the message linked earlier in the grub
                    265: section.
                    266: 
1.157   ! maxv      267: [[!template id=programlisting text="""
        !           268: # Install MBR bootblocks on both disks.
        !           269: fdisk -i /dev/rwd0d
        !           270: fdisk -i /dev/rwd1d
        !           271: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
        !           272: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
        !           273: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
        !           274: # Install secondary boot loader
        !           275: cp -p /usr/mdec/boot /
        !           276: # Create boot.cfg following earlier guidance:
        !           277: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
        !           278: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
        !           279: menu=GENERIC:boot
        !           280: menu=GENERIC single-user:boot -s
        !           281: menu=GENERIC.ok:boot netbsd.ok
        !           282: menu=GENERIC.ok single-user:boot netbsd.ok -s
        !           283: menu=Drop to boot prompt:prompt
        !           284: default=1
        !           285: timeout=30
        !           286: """]]
1.22      gdt       287: 
1.102     gdt       288: Upgrading Xen versions
1.15      gdt       289: ---------------------
                    290: 
1.110     gdt       291: Minor version upgrades are trivial.  Just rebuild/replace the
1.157   ! maxv      292: xenkernel version and copy the new xen.gz to `/` (where `/boot.cfg`
1.110     gdt       293: references it), and reboot.
                    294: 
1.28      gdt       295: Unprivileged domains (domU)
                    296: ===========================
                    297: 
                    298: This section describes general concepts about domUs.  It does not
1.33      gdt       299: address specific domU operating systems or how to install them.  The
1.157   ! maxv      300: config files for domUs are typically in `/usr/pkg/etc/xen`, and are
1.60      wiki      301: typically named so that the file name, domU name and the domU's host
1.33      gdt       302: name match.
                    303: 
1.111     wiz       304: The domU is provided with CPU and memory by Xen, configured by the
1.33      gdt       305: dom0.  The domU is provided with disk and network by the dom0,
                    306: mediated by Xen, and configured in the dom0.
                    307: 
                    308: Entropy in domUs can be an issue; physical disks and network are on
                    309: the dom0.  NetBSD's /dev/random system works, but is often challenged.
                    310: 
1.48      gdt       311: Config files
                    312: ------------
                    313: 
                    314: There is no good order to present config files and the concepts
                    315: surrounding what is being configured.  We first show an example config
                    316: file, and then in the various sections give details.
                    317: 
1.157   ! maxv      318: See /usr/pkg/share/examples/xen/xmexample*,
1.48      gdt       319: for a large number of well-commented examples, mostly for running
                    320: GNU/Linux.
                    321: 
1.157   ! maxv      322: The following is an example minimal domain configuration file. The domU
        !           323: serves as a network file server.
        !           324: 
        !           325: [[!template id=filecontent name="/usr/pkg/etc/xen/foo" text="""
        !           326: name = "domU-id"
        !           327: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
        !           328: memory = 1024
        !           329: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
        !           330: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
        !           331:          'file:/n0/xen/foo-wd1,0x1,w' ]
        !           332: """]]
1.48      gdt       333: 
1.157   ! maxv      334: The domain will have name given in the `name` setting.  The kernel has the
1.48      gdt       335: host/domU name in it, so that on the dom0 one can update the various
1.157   ! maxv      336: domUs independently.  The `vif` line causes an interface to be provided,
1.48      gdt       337: with a specific mac address (do not reuse MAC addresses!), in bridge
                    338: mode.  Two disks are provided, and they are both writable; the bits
                    339: are stored in files and Xen attaches them to a vnd(4) device in the
1.111     wiz       340: dom0 on domain creation.  The system treats xbd0 as the boot device
1.48      gdt       341: without needing explicit configuration.
                    342: 
1.157   ! maxv      343: By default, `xl` looks for domain config files in `/usr/pkg/etc/xen`.  Note
        !           344: that "xl create" takes the name of a config file, while other commands
        !           345: take the name of a domain.
        !           346: 
        !           347: Examples of commands:
1.48      gdt       348: 
1.157   ! maxv      349: [[!template id=programlisting text="""
        !           350: xl create foo
        !           351: xl console foo
        !           352: xl create -c foo
        !           353: xl shutdown foo
        !           354: xl list
        !           355: """]]
        !           356: 
        !           357: Typing `^]` will exit the console session.  Shutting down a domain is
1.48      gdt       358: equivalent to pushing the power button; a NetBSD domU will receive a
                    359: power-press event and do a clean shutdown.  Shutting down the dom0
                    360: will trigger controlled shutdowns of all configured domUs.
                    361: 
                    362: domU kernels
                    363: ------------
                    364: 
                    365: On a physical computer, the BIOS reads sector 0, and a chain of boot
                    366: loaders finds and loads a kernel.  Normally this comes from the root
1.111     wiz       367: file system.  With Xen domUs, the process is totally different.  The
1.48      gdt       368: normal path is for the domU kernel to be a file in the dom0's
1.111     wiz       369: file system.  At the request of the dom0, Xen loads that kernel into a
1.157   ! maxv      370: new domU instance and starts execution. domU kernels can be anyplace.
1.48      gdt       371: 
1.59      gdt       372: Note that loading the domU kernel from the dom0 implies that boot
                    373: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
1.48      gdt       374: See the VPS section near the end for discussion of alternate ways to
                    375: obtain domU kernels.
                    376: 
1.33      gdt       377: CPU and memory
                    378: --------------
                    379: 
1.48      gdt       380: A domain is provided with some number of vcpus, less than the number
1.157   ! maxv      381: of CPUs seen by the hypervisor. For a domU, it is controlled
1.48      gdt       382: from the config file by the "vcpus = N" directive.
                    383: 
                    384: A domain is provided with memory; this is controlled in the config
                    385: file by "memory = N" (in megabytes).  In the straightforward case, the
                    386: sum of the the memory allocated to the dom0 and all domUs must be less
1.33      gdt       387: than the available memory.
                    388: 
                    389: Xen also provides a "balloon" driver, which can be used to let domains
1.157   ! maxv      390: use more memory temporarily.
1.28      gdt       391: 
                    392: Virtual disks
                    393: -------------
                    394: 
1.33      gdt       395: With the file/vnd style, typically one creates a directory,
                    396: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
                    397: domUs.  Then, for each domU disk, one writes zeros to a file that then
                    398: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
                    399: for the first virtual disk for the domU called foo.  Writing zeros to
                    400: the file serves two purposes.  One is that preallocating the contents
                    401: improves performance.  The other is that vnd on sparse files has
                    402: failed to work.  TODO: give working/notworking NetBSD versions for
1.127     gdt       403: sparse vnd and gnats reference.  Note that the use of file/vnd for Xen
                    404: is not really different than creating a file-backed virtual disk for
                    405: some other purpose, except that xentools handles the vnconfig
                    406: commands.  To create an empty 4G virtual disk, simply do
1.39      gdt       407: 
                    408:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
1.33      gdt       409: 
1.89      gdt       410: Do not use qemu-img-xen, because this will create sparse file.  There
                    411: have been recent (2015) reports of sparse vnd(4) devices causing
                    412: lockups, but there is apparently no PR.
                    413: 
1.33      gdt       414: With the lvm style, one creates logical devices.  They are then used
1.48      gdt       415: similarly to vnds.  TODO: Add an example with lvm.
                    416: 
                    417: In domU config files, the disks are defined as a sequence of 3-tuples.
                    418: The first element is "method:/path/to/disk".  Common methods are
                    419: "file:" for file-backed vnd. and "phy:" for something that is already
                    420: a (TODO: character or block) device.
                    421: 
                    422: The second element is an artifact of how virtual disks are passed to
                    423: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
                    424: are given a device name to associate with the disk, and values like
                    425: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
                    426: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
                    427: second argument.  The name given is converted to a major/minor by
1.49      gdt       428: calling stat(2) on the name in /dev and this is passed to the domU.
                    429: In the general case, the dom0 and domU can be different operating
1.48      gdt       430: systems, and it is an unwarranted assumption that they have consistent
                    431: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
                    432: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
1.49      gdt       433: for the second works fine and avoids this issue.  For a GNU/Linux
                    434: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
                    435: /dev/hda1.
1.48      gdt       436: 
                    437: The third element is "w" for writable disks, and "r" for read-only
                    438: disks.
1.28      gdt       439: 
1.127     gdt       440: Note that NetBSD by default creates only vnd[0123].  If you need more
                    441: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
                    442: dom0.
                    443: 
                    444: Note that NetBSD by default creates only xbd[0123].  If you need more
                    445: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
                    446: 
1.28      gdt       447: Virtual Networking
                    448: ------------------
                    449: 
1.111     wiz       450: Xen provides virtual Ethernets, each of which connects the dom0 and a
1.46      gdt       451: domU.  For each virtual network, there is an interface "xvifN.M" in
                    452: the dom0, and in domU index N, a matching interface xennetM (NetBSD
                    453: name).  The interfaces behave as if there is an Ethernet with two
1.111     wiz       454: adapters connected.  From this primitive, one can construct various
1.46      gdt       455: configurations.  We focus on two common and useful cases for which
                    456: there are existing scripts: bridging and NAT.
1.28      gdt       457: 
1.48      gdt       458: With bridging (in the example above), the domU perceives itself to be
                    459: on the same network as the dom0.  For server virtualization, this is
                    460: usually best.  Bridging is accomplished by creating a bridge(4) device
                    461: and adding the dom0's physical interface and the various xvifN.0
                    462: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
                    463: config file.  The bridge must be set up already in the dom0; an
                    464: example /etc/ifconfig.bridge0 is:
1.46      gdt       465: 
1.157   ! maxv      466: [[!template id=filecontent name="/etc/ifconfig.bridge0" text="""
        !           467: create
        !           468: up
        !           469: !brconfig bridge0 add wm0
        !           470: """]]
1.28      gdt       471: 
                    472: With NAT, the domU perceives itself to be behind a NAT running on the
                    473: dom0.  This is often appropriate when running Xen on a workstation.
1.48      gdt       474: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28      gdt       475: 
1.49      gdt       476: The MAC address specified is the one used for the interface in the new
1.53      gdt       477: domain.  The interface in dom0 will use this address XOR'd with
1.49      gdt       478: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
                    479: 
1.33      gdt       480: Sizing domains
                    481: --------------
                    482: 
                    483: Modern x86 hardware has vast amounts of resources.  However, many
                    484: virtual servers can function just fine on far less.  A system with
1.142     gdt       485: 512M of RAM and a 4G disk can be a reasonable choice.  Note that it is
1.33      gdt       486: far easier to adjust virtual resources than physical ones.  For
                    487: memory, it's just a config file edit and a reboot.  For disk, one can
                    488: create a new file and vnconfig it (or lvm), and then dump/restore,
                    489: just like updating physical disks, but without having to be there and
                    490: without those pesky connectors.
                    491: 
1.48      gdt       492: Starting domains automatically
                    493: ------------------------------
1.28      gdt       494: 
1.48      gdt       495: To start domains foo at bar at boot and shut them down cleanly on dom0
                    496: shutdown, in rc.conf add:
1.28      gdt       497: 
1.48      gdt       498:         xendomains="foo bar"
1.28      gdt       499: 
1.111     wiz       500: Note that earlier versions of the xentools41 xendomains rc.d script
                    501: used xl, when one should use xm with 4.1.
1.28      gdt       502: 
                    503: Creating specific unprivileged domains (domU)
                    504: =============================================
1.14      gdt       505: 
                    506: Creating domUs is almost entirely independent of operating system.  We
1.49      gdt       507: have already presented the basics of config files.  Note that you must
                    508: have already completed the dom0 setup so that "xl list" (or "xm list")
                    509: works.
1.14      gdt       510: 
                    511: Creating an unprivileged NetBSD domain (domU)
                    512: ---------------------------------------------
1.1       mspo      513: 
1.49      gdt       514: See the earlier config file, and adjust memory.  Decide on how much
                    515: storage you will provide, and prepare it (file or lvm).
                    516: 
1.111     wiz       517: While the kernel will be obtained from the dom0 file system, the same
1.49      gdt       518: file should be present in the domU as /netbsd so that tools like
                    519: savecore(8) can work.   (This is helpful but not necessary.)
                    520: 
                    521: The kernel must be specifically for Xen and for use as a domU.  The
                    522: i386 and amd64 provide the following kernels:
                    523: 
                    524:         i386 XEN3PAE_DOMU
1.95      gdt       525:         amd64 XEN3_DOMU
1.5       mspo      526: 
1.49      gdt       527: This will boot NetBSD, but this is not that useful if the disk is
                    528: empty.  One approach is to unpack sets onto the disk outside of xen
                    529: (by mounting it, just as you would prepare a physical disk for a
                    530: system you can't run the installer on).
                    531: 
                    532: A second approach is to run an INSTALL kernel, which has a miniroot
                    533: and can load sets from the network.  To do this, copy the INSTALL
                    534: kernel to / and change the kernel line in the config file to:
1.5       mspo      535: 
1.49      gdt       536:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5       mspo      537: 
1.49      gdt       538: Then, start the domain as "xl create -c configname".
1.1       mspo      539: 
1.49      gdt       540: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
                    541: line should be used in the config file.
1.1       mspo      542: 
1.3       mspo      543:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1       mspo      544: 
                    545: After booting the domain, the option to install via CDROM may be
1.49      gdt       546: selected.  The CDROM device should be changed to `xbd1d`.
1.1       mspo      547: 
1.49      gdt       548: Once done installing, "halt -p" the new domain (don't reboot or halt,
                    549: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
                    550: config file), switch the config file back to the XEN3_DOMU kernel,
                    551: and start the new domain again. Now it should be able to use "root on
                    552: xbd0a" and you should have a, functional NetBSD domU.
1.1       mspo      553: 
1.49      gdt       554: TODO: check if this is still accurate.
1.1       mspo      555: When the new domain is booting you'll see some warnings about *wscons*
                    556: and the pseudo-terminals. These can be fixed by editing the files
1.5       mspo      557: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
                    558: `/etc/ttys`, except *console*, like this:
1.1       mspo      559: 
1.3       mspo      560:     console "/usr/libexec/getty Pc"         vt100   on secure
                    561:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
                    562:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
                    563:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
                    564:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
1.1       mspo      565: 
1.5       mspo      566: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1       mspo      567: 
                    568: It is also desirable to add
                    569: 
1.49      gdt       570:         powerd=YES
1.1       mspo      571: 
1.5       mspo      572: in rc.conf. This way, the domain will be properly shut down if
1.53      gdt       573: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
1.1       mspo      574: 
1.92      gdt       575: It is not strictly necessary to have a kernel (as /netbsd) in the domU
1.111     wiz       576: file system.  However, various programs (e.g. netstat) will use that
1.92      gdt       577: kernel to look up symbols to read from kernel virtual memory.  If
                    578: /netbsd is not the running kernel, those lookups will fail.  (This is
                    579: not really a Xen-specific issue, but because the domU kernel is
                    580: obtained from the dom0, it is far more likely to be out of sync or
                    581: missing with Xen.)
                    582: 
1.14      gdt       583: Creating an unprivileged Linux domain (domU)
1.5       mspo      584: --------------------------------------------
1.1       mspo      585: 
                    586: Creating unprivileged Linux domains isn't much different from
                    587: unprivileged NetBSD domains, but there are some details to know.
                    588: 
                    589: First, the second parameter passed to the disk declaration (the '0x1' in
                    590: the example below)
                    591: 
1.3       mspo      592:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1       mspo      593: 
                    594: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49      gdt       595: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
                    596: So, hda1 which has major 3 and minor 1 on a Linux system will have
                    597: device number 0x301.  Alternatively, devices names can be used (hda,
                    598: hdb, ...)  as xentools has a table to map these names to devices
                    599: numbers.  To export a partition to a Linux guest we can use:
1.1       mspo      600: 
1.49      gdt       601:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
                    602:         root = "/dev/hda1 ro"
1.1       mspo      603: 
                    604: and it will appear as /dev/hda on the Linux system, and be used as root
                    605: partition.
                    606: 
1.49      gdt       607: To install the Linux system on the partition to be exported to the
                    608: guest domain, the following method can be used: install
                    609: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
                    610: that will be the root partition of your Linux domain, and mount it.
                    611: Then copy the files from a working Linux system, make adjustments in
                    612: `/etc` (fstab, network config).  It should also be possible to extract
                    613: binary packages such as .rpm or .deb directly to the mounted partition
                    614: using the appropriate tool, possibly running under NetBSD's Linux
1.111     wiz       615: emulation.  Once the file system has been populated, umount it.  If
                    616: desirable, the file system can be converted to ext3 using tune2fs -j.
1.49      gdt       617: It should now be possible to boot the Linux guest domain, using one of
                    618: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1       mspo      619: 
1.111     wiz       620: To get the Linux console right, you need to add:
1.1       mspo      621: 
1.3       mspo      622:     extra = "xencons=tty1"
1.1       mspo      623: 
1.111     wiz       624: to your configuration since not all Linux distributions auto-attach a
1.1       mspo      625: tty to the xen console.
                    626: 
1.14      gdt       627: Creating an unprivileged Solaris domain (domU)
1.5       mspo      628: ----------------------------------------------
1.1       mspo      629: 
1.50      gdt       630: See possibly outdated
                    631: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5       mspo      632: 
1.1       mspo      633: 
1.52      gdt       634: PCI passthrough: Using PCI devices in guest domains
                    635: ---------------------------------------------------
1.1       mspo      636: 
1.53      gdt       637: The dom0 can give other domains access to selected PCI
1.52      gdt       638: devices. This can allow, for example, a non-privileged domain to have
                    639: access to a physical network interface or disk controller.  However,
                    640: keep in mind that giving a domain access to a PCI device most likely
                    641: will give the domain read/write access to the whole physical memory,
                    642: as PCs don't have an IOMMU to restrict memory access to DMA-capable
1.53      gdt       643: device.  Also, it's not possible to export ISA devices to non-dom0
1.52      gdt       644: domains, which means that the primary VGA adapter can't be exported.
                    645: A guest domain trying to access the VGA registers will panic.
                    646: 
1.53      gdt       647: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1.52      gdt       648: not been ported to later versions at this time.
                    649: 
                    650: For a PCI device to be exported to a domU, is has to be attached to
                    651: the "pciback" driver in dom0.  Devices passed to the dom0 via the
                    652: pciback.hide boot parameter will attach to "pciback" instead of the
                    653: usual driver.  The list of devices is specified as "(bus:dev.func)",
1.5       mspo      654: where bus and dev are 2-digit hexadecimal numbers, and func a
                    655: single-digit number:
1.1       mspo      656: 
1.52      gdt       657:         pciback.hide=(00:0a.0)(00:06.0)
1.1       mspo      658: 
1.52      gdt       659: pciback devices should show up in the dom0's boot messages, and the
1.5       mspo      660: devices should be listed in the `/kern/xen/pci` directory.
1.1       mspo      661: 
1.52      gdt       662: PCI devices to be exported to a domU are listed in the "pci" array of
                    663: the domU's config file, with the format "0000:bus:dev.func".
1.1       mspo      664: 
1.52      gdt       665:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1       mspo      666: 
1.52      gdt       667: In the domU an "xpci" device will show up, to which one or more pci
1.111     wiz       668: buses will attach.  Then the PCI drivers will attach to PCI buses as
1.52      gdt       669: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
                    670: or any PCI drivers built in by default; you have to build your own
                    671: kernel to use PCI devices in a domU.  Here's a kernel config example;
                    672: note that only the "xpci" lines are unusual.
                    673: 
                    674:         include         "arch/i386/conf/XEN3_DOMU"
                    675: 
1.111     wiz       676:         # Add support for PCI buses to the XEN3_DOMU kernel
1.52      gdt       677:         xpci* at xenbus ?
                    678:         pci* at xpci ?
                    679: 
                    680:         # PCI USB controllers
                    681:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
                    682: 
                    683:         # USB bus support
                    684:         usb*    at uhci?
                    685: 
                    686:         # USB Hubs
                    687:         uhub*   at usb?
                    688:         uhub*   at uhub? port ? configuration ? interface ?
                    689: 
                    690:         # USB Mass Storage
                    691:         umass*  at uhub? port ? configuration ? interface ?
                    692:         wd*     at umass?
                    693:         # SCSI controllers
                    694:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
                    695: 
                    696:         # SCSI bus support (for both ahc and umass)
                    697:         scsibus* at scsi?
                    698: 
                    699:         # SCSI devices
                    700:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
                    701:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
1.1       mspo      702: 
                    703: 
1.28      gdt       704: NetBSD as a domU in a VPS
                    705: =========================
                    706: 
                    707: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
                    708: hardware.  This section explains how to deal with Xen in a domU as a
                    709: virtual private server where you do not control or have access to the
1.70      gdt       710: dom0.  This is not intended to be an exhaustive list of VPS providers;
                    711: only a few are mentioned that specifically support NetBSD.
1.28      gdt       712: 
1.52      gdt       713: VPS operators provide varying degrees of access and mechanisms for
                    714: configuration.  The big issue is usually how one controls which kernel
1.111     wiz       715: is booted, because the kernel is nominally in the dom0 file system (to
                    716: which VPS users do not normally have access).  A second issue is how
1.70      gdt       717: to install NetBSD.
1.52      gdt       718: A VPS user may want to compile a kernel for security updates, to run
                    719: npf, run IPsec, or any other reason why someone would want to change
                    720: their kernel.
                    721: 
1.111     wiz       722: One approach is to have an administrative interface to upload a kernel,
1.68      gdt       723: or to select from a prepopulated list.  Other approaches are pygrub
1.59      gdt       724: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
1.111     wiz       725: kernel from the domU file system.  This is closer to a regular physical
1.59      gdt       726: computer, where someone who controls a machine can replace the kernel.
1.52      gdt       727: 
1.74      gdt       728: A second issue is multiple CPUs.  With NetBSD 6, domUs support
                    729: multiple vcpus, and it is typical for VPS providers to enable multiple
                    730: CPUs for NetBSD domUs.
                    731: 
1.68      gdt       732: pygrub
1.59      gdt       733: -------
1.52      gdt       734: 
1.111     wiz       735: pygrub runs in the dom0 and looks into the domU file system.  This
                    736: implies that the domU must have a kernel in a file system in a format
1.68      gdt       737: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
                    738: interest.
1.52      gdt       739: 
1.59      gdt       740: pvgrub
                    741: ------
                    742: 
                    743: pvgrub is a version of grub that uses PV operations instead of BIOS
                    744: calls.  It is booted from the dom0 as the domU kernel, and then reads
1.111     wiz       745: /grub/menu.lst and loads a kernel from the domU file system.
1.59      gdt       746: 
1.70      gdt       747: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
1.71      gdt       748: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
                    749: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
1.70      gdt       750: page](http://www.panix.com/v-colo/grub.html), which describes only
1.74      gdt       751: Linux but should be updated to cover NetBSD :-).
1.70      gdt       752: 
                    753: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
                    754: their own kernel.  See then [prgmr.com NetBSD
1.74      gdt       755: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
                    756: (which is in need of updating).
1.59      gdt       757: 
1.70      gdt       758: It appears that [grub's FFS
                    759: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
                    760: does not support all aspects of modern FFS, but there are also reports
1.72      gdt       761: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
1.70      gdt       762: partition for the kernel with the intent that grub can understand it,
                    763: which leads to /netbsd not being the actual kernel.  One must remember
1.111     wiz       764: to update the special boot partition.
1.59      gdt       765: 
                    766: Amazon
                    767: ------
                    768: 
1.143     wiki      769: See the [Amazon EC2 page](/amazon_ec2/).
1.44      gdt       770: 
1.93      gdt       771: Random pointers
                    772: ===============
                    773: 
1.117     gdt       774: This section contains links from elsewhere not yet integrated into the
                    775: HOWTO, and other guides.
1.93      gdt       776: 
                    777: * http://www.lumbercartel.ca/library/xen/
                    778: * http://pbraun.nethence.com/doc/sysutils/xen_netbsd_dom0.html
1.117     gdt       779: * https://gmplib.org/~tege/xen.html

CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb