Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.52

1.5       mspo        1: Introduction
1.13      gdt         2: ============
1.1       mspo        3: 
                      4: [![[Xen
1.7       mspo        5: screenshot]](http://www.netbsd.org/gallery/in-Action/hubertf-xens.png)](../../gallery/in-Action/hubertf-xen.png)
1.1       mspo        6: 
1.12      gdt         7: Xen is a virtual machine monitor or hypervisor for x86 hardware
                      8: (i686-class or higher), which supports running multiple guest
                      9: operating systems on a single physical machine.  With Xen, one uses
                     10: the Xen kernel to control the CPU, memory and console, a dom0
                     11: operating system which mediates access to other hardware (e.g., disks,
                     12: network, USB), and one or more domU operating systems which operate in
                     13: an unprivileged virtualized environment.  IO requests from the domU
                     14: systems are forwarded by the hypervisor (Xen) to the dom0 to be
                     15: fulfilled.
                     16: 
                     17: Xen supports two styles of guests.  The original is Para-Virtualized
                     18: (PV) which means that the guest OS does not attempt to access hardware
                     19: directly, but instead makes hypercalls to the hypervisor.  This is
                     20: analogous to a user-space program making system calls.  (The dom0
                     21: operating system uses PV calls for some functions, such as updating
                     22: memory mapping page tables, but has direct hardware access for disk
                     23: and network.)   PV guests must be specifically coded for Xen.
                     24: 
                     25: The more recent style is HVM, which means that the guest does not have
                     26: code for Xen and need not be aware that it is running under Xen.
                     27: Attempts to access hardware registers are trapped and emulated.  This
                     28: style is less efficient but can run unmodified guests.
                     29: 
1.29      gdt        30: Generally any amd64 machine will work with Xen and PV guests.  In
                     31: theory i386 computers without amd64 support can be used for Xen <=
                     32: 4.2, but we have no recent reports of this working (this is a hint).
                     33: For HVM guests, the VT or VMX cpu feature (Intel) or SVM/HVM/VT
                     34: (amd64) is needed; "cpuctl identify 0" will show this.  TODO: Clean up
                     35: and check the above features.
1.19      gdt        36: 
1.27      jnemeth    37: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
1.12      gdt        38: The dom0 can start one or more domUs.  (Booting is explained in detail
                     39: in the dom0 section.)
                     40: 
                     41: NetBSD supports Xen in that it can serve as dom0, be used as a domU,
                     42: and that Xen kernels and tools are available in pkgsrc.  This HOWTO
                     43: attempts to address both the case of running a NetBSD dom0 on hardware
1.24      gdt        44: and running domUs under it (NetBSD and other), and also running NetBSD
                     45: as a domU in a VPS.
1.12      gdt        46: 
1.20      gdt        47: Some versions of Xen support "PCI passthrough", which means that
                     48: specific PCI devices can be made available to a specific domU instead
                     49: of the dom0.  This can be useful to let a domU run X11, or access some
                     50: network interface or other peripheral.
                     51: 
1.12      gdt        52: Prerequisites
1.13      gdt        53: -------------
1.12      gdt        54: 
                     55: Installing NetBSD/Xen is not extremely difficult, but it is more
                     56: complex than a normal installation of NetBSD.
1.15      gdt        57: In general, this HOWTO is occasionally overly restrictive about how
                     58: things must be done, guiding the reader to stay on the established
                     59: path when there are no known good reasons to stray.
1.12      gdt        60: 
                     61: This HOWTO presumes a basic familiarity with the Xen system
1.16      gdt        62: architecture.  This HOWTO presumes familiarity with installing NetBSD
                     63: on i386/amd64 hardware and installing software from pkgsrc.
1.27      jnemeth    64: See also the [Xen website](http://www.xenproject.org/).
1.1       mspo       65: 
1.19      gdt        66: History
                     67: -------
                     68: 
                     69: NetBSD used to support Xen2; this has been removed.
                     70: 
                     71: Before NetBSD's native bootloader could support Xen, the use of
                     72: grub was recommended.  If necessary, see the
1.27      jnemeth    73: [old grub information](/ports/xen/howto-grub/).
1.19      gdt        74: 
1.15      gdt        75: Versions of Xen and NetBSD
                     76: ==========================
                     77: 
1.27      jnemeth    78: Most of the installation concepts and instructions are independent
                     79: of Xen version and NetBSD version.  This section gives advice on
                     80: which version to choose.  Versions not in pkgsrc and older unsupported
                     81: versions of NetBSD are intentionally ignored.
1.15      gdt        82: 
                     83: Xen
                     84: ---
                     85: 
                     86: In NetBSD, xen is provided in pkgsrc, via matching pairs of packages
                     87: xenkernel and xentools.  We will refer only to the kernel versions,
                     88: but note that both packages must be installed together and must have
                     89: matching versions.
                     90: 
                     91: xenkernel3 and xenkernel33 provide Xen 3.1 and 3.3.  These no longer
1.20      gdt        92: receive security patches and should not be used.  Xen 3.1 supports PCI
1.29      gdt        93: passthrough.  Xen 3.1 supports non-PAE on i386.
1.15      gdt        94: 
                     95: xenkernel41 provides Xen 4.1.  This is no longer maintained by Xen,
                     96: but as of 2014-12 receives backported security patches.  It is a
                     97: reasonable although trailing-edge choice.
                     98: 
                     99: xenkernel42 provides Xen 4.2.  This is maintained by Xen, but old as
                    100: of 2014-12.
                    101: 
                    102: Ideally newer versions of Xen will be added to pkgsrc.
                    103: 
1.26      gdt       104: Note that NetBSD support is called XEN3.  It works with 3.1 through
                    105: 4.2 because the hypercall interface has been stable.
1.20      gdt       106: 
1.19      gdt       107: Xen command program
                    108: -------------------
                    109: 
                    110: Early Xen used a program called "xm" to manipulate the system from the
                    111: dom0.  Starting in 4.1, a replacement program with similar behavior
1.27      jnemeth   112: called "xl" is provided.  In 4.2 and later, "xl" is preferred.  4.4 is
                    113: the last version that has "xm".
1.19      gdt       114: 
1.15      gdt       115: NetBSD
                    116: ------
                    117: 
                    118: The netbsd-5, netbsd-6, netbsd-7, and -current branches are all
                    119: reasonable choices, with more or less the same considerations for
                    120: non-Xen use.  Therefore, netbsd-6 is recommended as the stable version
1.29      gdt       121: of the most recent release for production use.  For those wanting to
                    122: learn Xen or without production stability concerns, netbsd-7 is likely
                    123: most appropriate.
1.15      gdt       124: 
                    125: As of NetBSD 6, a NetBSD domU will support multiple vcpus.  There is
                    126: no SMP support for NetBSD as dom0.  (The dom0 itself doesn't really
                    127: need SMP; the lack of support is really a problem when using a dom0 as
                    128: a normal computer.)
                    129: 
1.18      gdt       130: Architecture
                    131: ------------
                    132: 
1.29      gdt       133: Xen itself can run on i386 or amd64 machines.  (Practically, almost
                    134: any computer where one would want to run Xen supports amd64.)  If
                    135: using an i386 NetBSD kernel for the dom0, PAE is required (PAE
                    136: versions are built by default).  While i386 dom0 works fine, amd64 is
                    137: recommended as more normal.
                    138: 
                    139: Xen 4.2 is the last version to support i386 as a host.  TODO: Clarify
                    140: if this is about the CPU having to be amd64, or about the dom0 kernel
                    141: having to be amd64.
                    142: 
                    143: One can then run i386 domUs and amd64 domUs, in any combination.  If
                    144: running an i386 NetBSD kernel as a domU, the PAE version is required.
                    145: (Note that emacs (at least) fails if run on i386 with PAE when built
                    146: without, and vice versa, presumably due to bugs in the undump code.)
1.18      gdt       147: 
1.15      gdt       148: Recommendation
                    149: --------------
                    150: 
1.18      gdt       151: Therefore, this HOWTO recommends running xenkernel42 (and xentools42),
1.30      gdt       152: xl, the NetBSD 6 stable branch, and to use an amd64 kernel as the
                    153: dom0.  Either the i386 or amd64 of NetBSD may be used as domUs.
1.15      gdt       154: 
1.36      gdt       155: Build problems
                    156: --------------
                    157: 
                    158: Ideally, all versions of Xen in pkgsrc would build on all versions of
                    159: NetBSD on both i386 and amd64.  However, that isn't the case.  Besides
                    160: aging code and aging compilers, qemu (included in xentools for HVM
                    161: support) is difficult to build.  The following are known to fail:
                    162: 
                    163:         xenkernel3 netbsd-6 i386
                    164:         xentools42 netbsd-6 i386 
                    165: 
                    166: The following are known to work:
                    167: 
                    168:         xenkernel41 netbsd-5 amd64
                    169:         xentools41 netbsd-5 amd64
                    170:         xenkernel41 netbsd-6 i386
                    171:         xentools41 netbsd-6 i386
                    172: 
1.15      gdt       173: NetBSD as a dom0
                    174: ================
                    175: 
                    176: NetBSD can be used as a dom0 and works very well.  The following
                    177: sections address installation, updating NetBSD, and updating Xen.
1.19      gdt       178: Note that it doesn't make sense to talk about installing a dom0 OS
                    179: without also installing Xen itself.  We first address installing
                    180: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
                    181: NetBSD install to a dom0 install by just changing the kernel and boot
                    182: configuration.
1.15      gdt       183: 
1.45      gdt       184: For experimenting with Xen, a machine with as little as 1G of RAM and
                    185: 100G of disk can work.  For running many domUs in productions, far
                    186: more will be needed.
                    187: 
1.15      gdt       188: Styles of dom0 operation
                    189: ------------------------
                    190: 
                    191: There are two basic ways to use Xen.  The traditional method is for
                    192: the dom0 to do absolutely nothing other than providing support to some
                    193: number of domUs.  Such a system was probably installed for the sole
                    194: purpose of hosting domUs, and sits in a server room on a UPS.
                    195: 
                    196: The other way is to put Xen under a normal-usage computer, so that the
                    197: dom0 is what the computer would have been without Xen, perhaps a
                    198: desktop or laptop.  Then, one can run domUs at will.  Purists will
                    199: deride this as less secure than the previous approach, and for a
                    200: computer whose purpose is to run domUs, they are right.  But Xen and a
                    201: dom0 (without domUs) is not meaingfully less secure than the same
                    202: things running without Xen.  One can boot Xen or boot regular NetBSD
                    203: alternately with little problems, simply refraining from starting the
                    204: Xen daemons when not running Xen.
                    205: 
                    206: Note that NetBSD as dom0 does not support multiple CPUs.  This will
1.51      gdt       207: limit the performance of the Xen/dom0 workstation approach.  In theory
                    208: the only issue is that the "backend drivers" are not yet MPSAFE:
                    209:   http://mail-index.netbsd.org/netbsd-users/2014/08/29/msg015195.html
1.15      gdt       210: 
1.19      gdt       211: Installation of NetBSD
                    212: ----------------------
1.13      gdt       213: 
1.19      gdt       214: First,
1.27      jnemeth   215: [install NetBSD/amd64](/guide/inst/)
1.19      gdt       216: just as you would if you were not using Xen.
                    217: However, the partitioning approach is very important.
                    218: 
                    219: If you want to use RAIDframe for the dom0, there are no special issues
                    220: for Xen.  Typically one provides RAID storage for the dom0, and the
1.22      gdt       221: domU systems are unaware of RAID.  The 2nd-stage loader bootxx_* skips
                    222: over a RAID1 header to find /boot from a filesystem within a RAID
                    223: partition; this is no different when booting Xen.
1.19      gdt       224: 
                    225: There are 4 styles of providing backing storage for the virtual disks
                    226: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN,
                    227: 
                    228: With raw partitions, one has a disklabel (or gpt) partition sized for
                    229: each virtual disk to be used by the domU.  (If you are able to predict
                    230: how domU usage will evolve, please add an explanation to the HOWTO.
                    231: Seriously, needs tend to change over time.)
                    232: 
1.27      jnemeth   233: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
                    234: for domU disks.  This is almost as efficient as raw disk partitions
                    235: and more flexible.  Hence raw disk partitions should typically not
                    236: be used.
1.19      gdt       237: 
                    238: One can use files in the dom0 filesystem, typically created by dd'ing
                    239: /dev/zero to create a specific size.  This is somewhat less efficient,
                    240: but very convenient, as one can cp the files for backup, or move them
                    241: between dom0 hosts.
                    242: 
                    243: Finally, in theory one can place the files backing the domU disks in a
                    244: SAN.  (This is an invitation for someone who has done this to add a
                    245: HOWTO page.)
1.1       mspo      246: 
1.19      gdt       247: Installation of Xen
                    248: -------------------
1.1       mspo      249: 
1.20      gdt       250: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
                    251: pkgsrc (or another matching pair).
                    252: See [the pkgsrc
                    253: documentation](http://www.NetBSD.org/docs/pkgsrc/) for help with pkgsrc.
                    254: 
                    255: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm.  More
                    256: recent versions have HVM support integrated in the main xentools
                    257: package.  It is entirely reasonable to run only PV guests.
                    258: 
                    259: Next you need to install the selected Xen kernel itself, which is
                    260: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz".  Copy it to /.
                    261: For debugging, one may copy xen-debug.gz; this is conceptually similar
                    262: to DIAGNOSTIC and DEBUG in NetBSD.  xen-debug.gz is basically only
                    263: useful with a serial console.  Then, place a NetBSD XEN3_DOM0 kernel
                    264: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
                    265: of a NetBSD build.  Both xen and NetBSD may be left compressed.  (If
                    266: using i386, use releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.)
                    267: 
1.49      gdt       268: With Xen as the kernel, you must provide a dom0 NetBSD kernel to be
                    269: used as a module; place this in /.  Suitable kernels are provided in
                    270: releasedir/binary/kernel:
                    271: 
                    272:         i386 XEN3_DOM0
                    273:         i386 XEN3PAE_DOM0
                    274:        amd64 XEN3_DOM0
                    275: 
                    276: The first one is only for use with Xen 3.1 and i386-mode Xen (and you
                    277: should not do this).  Current Xen always uses PAE on i386, but you
                    278: should generally use amd64 for the dom0.  In a dom0 kernel, kernfs is
                    279: mandatory for xend to comunicate with the kernel, so ensure that /kern
                    280: is in fstab.  TODO: Say this is default, or file a PR and give a
                    281: reference.
1.20      gdt       282: 
                    283: Because you already installed NetBSD, you have a working boot setup
                    284: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
                    285: beginning of your root filesystem, /boot present, and likely
                    286: /boot.cfg.  (If not, fix before continuing!)
                    287: 
                    288: See boot.cfg(5) for an example.  The basic line is
                    289: 
1.37      gdt       290:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
1.20      gdt       291: 
                    292: which specifies that the dom0 should have 256M, leaving the rest to be
1.37      gdt       293: allocated for domUs.  In an attempt to add performance, one can also
                    294: add
                    295: 
                    296:         dom0_max_vcpus=1 dom0_vcpus_pin
                    297: 
                    298: to force only one vcpu to be provided (since NetBSD dom0 can't use
                    299: more) and to pin that vcpu to a physical cpu.  TODO: benchmark this.
1.20      gdt       300: 
                    301: As with non-Xen systems, you should have a line to boot /netbsd (a
                    302: kernel that works without Xen) and fallback versions of the non-Xen
                    303: kernel, Xen, and the dom0 kernel.
1.1       mspo      304: 
1.28      gdt       305: The [HowTo on Installing into
                    306: RAID-1](http://mail-index.NetBSD.org/port-xen/2006/03/01/0010.html)
                    307: explains how to set up booting a dom0 with Xen using grub with
                    308: NetBSD's RAIDframe.  (This is obsolete with the use of NetBSD's native
                    309: boot.)
                    310: 
1.21      gdt       311: Configuring Xen
                    312: ---------------
                    313: 
                    314: Now, you have a system that will boot Xen and the dom0 kernel, and
                    315: just run the dom0 kernel.  There will be no domUs, and none can be
1.31      gdt       316: started because you still have to configure the dom0 tools.  The
                    317: daemons which should be run vary with Xen version and with whether one
                    318: is using xm or xl.  Note that xend is for supporting "xm", and should
                    319: only be used if you plan on using "xm".  Do NOT enable xend if you
                    320: plan on using "xl" as it will cause problems.
1.21      gdt       321: 
1.43      gdt       322: The installation of NetBSD should already have created devices for xen
                    323: (xencons, xenevt), but if they are not present, create them:
                    324: 
                    325:         cd /dev && sh MAKEDEV xen
                    326: 
1.31      gdt       327: TODO: Give 3.1 advice (or remove it from pkgsrc).
                    328: 
                    329: For 3.3 (and thus xm), add to rc.conf (but note that you should have
                    330: installed 4.1 or 4.2):
                    331: 
1.32      gdt       332:         xend=YES
                    333:         xenbackendd=YES
1.31      gdt       334: 
1.33      gdt       335: For 4.1 (and thus xm; xl is believed not to work well), add to rc.conf:
1.31      gdt       336: 
                    337:         xend=YES
                    338:         xencommons=YES
                    339: 
                    340: TODO: Explain why if xm is preferred on 4.1, rc.d/xendomains has xl.
1.33      gdt       341: Or fix the package.
1.31      gdt       342: 
1.33      gdt       343: For 4.2 with xm, add to rc.conf
                    344: 
                    345:         xend=YES
                    346:         xencommons=YES
                    347: 
                    348: For 4.2 with xl (preferred), add to rc.conf:
1.31      gdt       349: 
                    350:         TODO: explain if there is a xend replacement
                    351:         xencommons=YES
                    352: 
                    353: TODO: Recommend for/against xen-watchdog.
1.27      jnemeth   354: 
1.43      gdt       355: After you have configured the daemons and either started them or
1.42      gdt       356: rebooted, run the following (or use xl) to inspect Xen's boot
                    357: messages, available resources, and running domains:
1.34      gdt       358: 
1.43      gdt       359:         # xm dmesg
                    360:        [xen's boot info]
                    361:         # xm info
                    362:        [available memory, etc.]
                    363:         # xm list
                    364:         Name              Id  Mem(MB)  CPU  State  Time(s)  Console
                    365:         Domain-0           0       64    0  r----     58.1
1.33      gdt       366: 
1.41      gdt       367: anita (for testing NetBSD)
                    368: --------------------------
                    369: 
                    370: With the setup so far, one should be able to run anita (see
                    371: pkgsrc/sysutils/py-anita) to test NetBSD releases, by doing (as root,
                    372: because anita must create a domU):
                    373: 
                    374:         anita --vmm=xm test file:///usr/obj/i386/
                    375: 
                    376: Alternatively, one can use --vmm=xl to use xl-based domU creation instead.
                    377: TODO: check this.
                    378: 
1.40      gdt       379: Xen-specific NetBSD issues
                    380: --------------------------
                    381: 
                    382: There are (at least) two additional things different about NetBSD as a
                    383: dom0 kernel compared to hardware.
                    384: 
                    385: One is that modules are not usable in DOM0 kernels, so one must
                    386: compile in what's needed.  It's not really that modules cannot work,
                    387: but that modules must be built for XEN3_DOM0 because some of the
                    388: defines change and the normal module builds don't do this.  Basically,
                    389: enabling Xen changes the kernel ABI, and the module build system
                    390: doesn't cope with this.
                    391: 
                    392: The other difference is that XEN3_DOM0 does not have exactly the same
                    393: options as GENERIC.  While it is debatable whether or not this is a
                    394: bug, users should be aware of this and can simply add missing config
                    395: items if desired.
                    396: 
1.15      gdt       397: Updating NetBSD in a dom0
                    398: -------------------------
                    399: 
                    400: This is just like updating NetBSD on bare hardware, assuming the new
                    401: version supports the version of Xen you are running.  Generally, one
                    402: replaces the kernel and reboots, and then overlays userland binaries
                    403: and adjusts /etc.
                    404: 
                    405: Note that one must update both the non-Xen kernel typically used for
                    406: rescue purposes and the DOM0 kernel used with Xen.
                    407: 
1.22      gdt       408: To convert from grub to /boot, install an mbr bootblock with fdisk,
                    409: bootxx_ with installboot, /boot and /boot.cfg.  This really should be
                    410: no different than completely reinstalling boot blocks on a non-Xen
                    411: system.
                    412: 
1.15      gdt       413: Updating Xen versions
                    414: ---------------------
                    415: 
1.21      gdt       416: Updating Xen is conceptually not difficult, but can run into all the
                    417: issues found when installing Xen.  Assuming migration from 4.1 to 4.2,
                    418: remove the xenkernel41 and xentools41 packages and install the
                    419: xenkernel42 and xentools42 packages.  Copy the 4.2 xen.gz to /.
                    420: 
                    421: Ensure that the contents of /etc/rc.d/xen* are correct.  Enable the
                    422: correct set of daemons.  Ensure that the domU config files are valid
                    423: for the new version.
1.15      gdt       424: 
1.28      gdt       425: 
                    426: Unprivileged domains (domU)
                    427: ===========================
                    428: 
                    429: This section describes general concepts about domUs.  It does not
1.33      gdt       430: address specific domU operating systems or how to install them.  The
                    431: config files for domUs are typically in /usr/pkg/etc/xen, and are
                    432: typically named so that the file anme, domU name and the domU's host
                    433: name match.
                    434: 
                    435: The domU is provided with cpu and memory by Xen, configured by the
                    436: dom0.  The domU is provided with disk and network by the dom0,
                    437: mediated by Xen, and configured in the dom0.
                    438: 
                    439: Entropy in domUs can be an issue; physical disks and network are on
                    440: the dom0.  NetBSD's /dev/random system works, but is often challenged.
                    441: 
1.48      gdt       442: Config files
                    443: ------------
                    444: 
                    445: There is no good order to present config files and the concepts
                    446: surrounding what is being configured.  We first show an example config
                    447: file, and then in the various sections give details.
                    448: 
                    449: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
                    450: for a large number of well-commented examples, mostly for running
                    451: GNU/Linux.
                    452: 
                    453: The following is an example minimal domain configuration file
                    454: "/usr/pkg/etc/xen/foo".  It is (with only a name change) an actual
                    455: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
                    456: i386 domU).  The domU serves as a network file server.
                    457: 
                    458:         # -*- mode: python; -*-
                    459: 
                    460:         kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
                    461:         memory = 1024
                    462:         vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
                    463:         disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
                    464:                  'file:/n0/xen/foo-wd1,0x1,w' ]
                    465: 
                    466: The domain will have the same name as the file.  The kernel has the
                    467: host/domU name in it, so that on the dom0 one can update the various
                    468: domUs independently.  The vif line causes an interface to be provided,
                    469: with a specific mac address (do not reuse MAC addresses!), in bridge
                    470: mode.  Two disks are provided, and they are both writable; the bits
                    471: are stored in files and Xen attaches them to a vnd(4) device in the
                    472: dom0 on domain creation.  The system treates xbd0 as the boot device
                    473: without needing explicit configuration.
                    474: 
                    475: By default xm looks for domain config files in /usr/pkg/etc/xen.  Note
                    476: that "xm create" takes the name of a config file, while other commands
                    477: take the name of a domain.  To create the domain, connect to the
                    478: console, create the domain while attaching the console, shutdown the
                    479: domain, and see if it has finished stopping, do (or xl with Xen >=
                    480: 4.2):
                    481: 
                    482:         xm create foo
                    483:         xm console foo
                    484:         xm create -c foo
                    485:         xm shutdown foo
                    486:        xm list
                    487: 
                    488: Typing ^] will exit the console session.  Shutting down a domain is
                    489: equivalent to pushing the power button; a NetBSD domU will receive a
                    490: power-press event and do a clean shutdown.  Shutting down the dom0
                    491: will trigger controlled shutdowns of all configured domUs.
                    492: 
                    493: domU kernels
                    494: ------------
                    495: 
                    496: On a physical computer, the BIOS reads sector 0, and a chain of boot
                    497: loaders finds and loads a kernel.  Normally this comes from the root
                    498: filesystem.  With Xen domUs, the process is totally different.  The
                    499: normal path is for the domU kernel to be a file in the dom0's
                    500: filesystem.  At the request of the dom0, Xen loads that kernel into a
                    501: new domU instance and starts execution.  While domU kernels can be
                    502: anyplace, reasonable places to store domU kernels on the dom0 are in /
                    503: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
                    504: config files), or in /u0/xen (where the vdisks are).
                    505: 
                    506: See the VPS section near the end for discussion of alternate ways to
                    507: obtain domU kernels.
                    508: 
1.33      gdt       509: CPU and memory
                    510: --------------
                    511: 
1.48      gdt       512: A domain is provided with some number of vcpus, less than the number
                    513: of cpus seen by the hypervisor.  (For a dom0, this is controlled by
                    514: the boot argument "dom0_max_vcpus=1".)  For a domU, it is controlled
                    515: from the config file by the "vcpus = N" directive.
                    516: 
                    517: A domain is provided with memory; this is controlled in the config
                    518: file by "memory = N" (in megabytes).  In the straightforward case, the
                    519: sum of the the memory allocated to the dom0 and all domUs must be less
1.33      gdt       520: than the available memory.
                    521: 
                    522: Xen also provides a "balloon" driver, which can be used to let domains
                    523: use more memory temporarily.  TODO: Explain better, and explain how
                    524: well it works with NetBSD.
1.28      gdt       525: 
                    526: Virtual disks
                    527: -------------
                    528: 
1.33      gdt       529: With the file/vnd style, typically one creates a directory,
                    530: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
                    531: domUs.  Then, for each domU disk, one writes zeros to a file that then
                    532: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
                    533: for the first virtual disk for the domU called foo.  Writing zeros to
                    534: the file serves two purposes.  One is that preallocating the contents
                    535: improves performance.  The other is that vnd on sparse files has
                    536: failed to work.  TODO: give working/notworking NetBSD versions for
                    537: sparse vnd.  Note that the use of file/vnd for Xen is not really
                    538: different than creating a file-backed virtual disk for some other
1.39      gdt       539: purpose, except that xentools handles the vnconfig commands.  To
                    540: create an empty 4G virtual disk, simply do
                    541: 
                    542:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
1.33      gdt       543: 
                    544: With the lvm style, one creates logical devices.  They are then used
1.48      gdt       545: similarly to vnds.  TODO: Add an example with lvm.
                    546: 
                    547: In domU config files, the disks are defined as a sequence of 3-tuples.
                    548: The first element is "method:/path/to/disk".  Common methods are
                    549: "file:" for file-backed vnd. and "phy:" for something that is already
                    550: a (TODO: character or block) device.
                    551: 
                    552: The second element is an artifact of how virtual disks are passed to
                    553: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
                    554: are given a device name to associate with the disk, and values like
                    555: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
                    556: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
                    557: second argument.  The name given is converted to a major/minor by
1.49      gdt       558: calling stat(2) on the name in /dev and this is passed to the domU.
                    559: In the general case, the dom0 and domU can be different operating
1.48      gdt       560: systems, and it is an unwarranted assumption that they have consistent
                    561: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
                    562: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
1.49      gdt       563: for the second works fine and avoids this issue.  For a GNU/Linux
                    564: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
                    565: /dev/hda1.
1.48      gdt       566: 
                    567: The third element is "w" for writable disks, and "r" for read-only
                    568: disks.
1.28      gdt       569: 
                    570: Virtual Networking
                    571: ------------------
                    572: 
1.46      gdt       573: Xen provides virtual ethernets, each of which connects the dom0 and a
                    574: domU.  For each virtual network, there is an interface "xvifN.M" in
                    575: the dom0, and in domU index N, a matching interface xennetM (NetBSD
                    576: name).  The interfaces behave as if there is an Ethernet with two
                    577: adaptors connected.  From this primitive, one can construct various
                    578: configurations.  We focus on two common and useful cases for which
                    579: there are existing scripts: bridging and NAT.
1.28      gdt       580: 
1.48      gdt       581: With bridging (in the example above), the domU perceives itself to be
                    582: on the same network as the dom0.  For server virtualization, this is
                    583: usually best.  Bridging is accomplished by creating a bridge(4) device
                    584: and adding the dom0's physical interface and the various xvifN.0
                    585: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
                    586: config file.  The bridge must be set up already in the dom0; an
                    587: example /etc/ifconfig.bridge0 is:
1.46      gdt       588: 
                    589:         create
                    590:         up
                    591:         !brconfig bridge0 add wm0
1.28      gdt       592: 
                    593: With NAT, the domU perceives itself to be behind a NAT running on the
                    594: dom0.  This is often appropriate when running Xen on a workstation.
1.48      gdt       595: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28      gdt       596: 
1.49      gdt       597: The MAC address specified is the one used for the interface in the new
                    598: domain.  The interface in domain0 will use this address XOR'd with
                    599: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
                    600: 
1.33      gdt       601: Sizing domains
                    602: --------------
                    603: 
                    604: Modern x86 hardware has vast amounts of resources.  However, many
                    605: virtual servers can function just fine on far less.  A system with
                    606: 256M of RAM and a 4G disk can be a reasonable choice.  Note that it is
                    607: far easier to adjust virtual resources than physical ones.  For
                    608: memory, it's just a config file edit and a reboot.  For disk, one can
                    609: create a new file and vnconfig it (or lvm), and then dump/restore,
                    610: just like updating physical disks, but without having to be there and
                    611: without those pesky connectors.
                    612: 
1.48      gdt       613: Starting domains automatically
                    614: ------------------------------
1.28      gdt       615: 
1.48      gdt       616: To start domains foo at bar at boot and shut them down cleanly on dom0
                    617: shutdown, in rc.conf add:
1.28      gdt       618: 
1.48      gdt       619:         xendomains="foo bar"
1.28      gdt       620: 
                    621: TODO: Explain why 4.1 rc.d/xendomains has xl, when one should use xm
1.48      gdt       622: on 4.1.  Or fix the xentools41 package to have xm
1.28      gdt       623: 
                    624: Creating specific unprivileged domains (domU)
                    625: =============================================
1.14      gdt       626: 
                    627: Creating domUs is almost entirely independent of operating system.  We
1.49      gdt       628: have already presented the basics of config files.  Note that you must
                    629: have already completed the dom0 setup so that "xl list" (or "xm list")
                    630: works.
1.14      gdt       631: 
                    632: Creating an unprivileged NetBSD domain (domU)
                    633: ---------------------------------------------
1.1       mspo      634: 
1.49      gdt       635: See the earlier config file, and adjust memory.  Decide on how much
                    636: storage you will provide, and prepare it (file or lvm).
                    637: 
                    638: While the kernel will be obtained from the dom0 filesystem, the same
                    639: file should be present in the domU as /netbsd so that tools like
                    640: savecore(8) can work.   (This is helpful but not necessary.)
                    641: 
                    642: The kernel must be specifically for Xen and for use as a domU.  The
                    643: i386 and amd64 provide the following kernels:
                    644: 
                    645:         i386 XEN3_DOMU
                    646:         i386 XEN3PAE_DOMU
                    647:        amd64 XEN3_DOMU
1.5       mspo      648: 
1.49      gdt       649: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
                    650: use the PAE version of the i386 kernel.
                    651: 
                    652: This will boot NetBSD, but this is not that useful if the disk is
                    653: empty.  One approach is to unpack sets onto the disk outside of xen
                    654: (by mounting it, just as you would prepare a physical disk for a
                    655: system you can't run the installer on).
                    656: 
                    657: A second approach is to run an INSTALL kernel, which has a miniroot
                    658: and can load sets from the network.  To do this, copy the INSTALL
                    659: kernel to / and change the kernel line in the config file to:
1.5       mspo      660: 
1.49      gdt       661:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5       mspo      662: 
1.49      gdt       663: Then, start the domain as "xl create -c configname".
1.1       mspo      664: 
1.49      gdt       665: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
                    666: line should be used in the config file.
1.1       mspo      667: 
1.3       mspo      668:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1       mspo      669: 
                    670: After booting the domain, the option to install via CDROM may be
1.49      gdt       671: selected.  The CDROM device should be changed to `xbd1d`.
1.1       mspo      672: 
1.49      gdt       673: Once done installing, "halt -p" the new domain (don't reboot or halt,
                    674: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
                    675: config file), switch the config file back to the XEN3_DOMU kernel,
                    676: and start the new domain again. Now it should be able to use "root on
                    677: xbd0a" and you should have a, functional NetBSD domU.
1.1       mspo      678: 
1.49      gdt       679: TODO: check if this is still accurate.
1.1       mspo      680: When the new domain is booting you'll see some warnings about *wscons*
                    681: and the pseudo-terminals. These can be fixed by editing the files
1.5       mspo      682: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
                    683: `/etc/ttys`, except *console*, like this:
1.1       mspo      684: 
1.3       mspo      685:     console "/usr/libexec/getty Pc"         vt100   on secure
                    686:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
                    687:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
                    688:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
                    689:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
1.1       mspo      690: 
1.5       mspo      691: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1       mspo      692: 
                    693: It is also desirable to add
                    694: 
1.49      gdt       695:         powerd=YES
1.1       mspo      696: 
1.5       mspo      697: in rc.conf. This way, the domain will be properly shut down if
                    698: `xm shutdown -R` or `xm shutdown -H` is used on the domain0.
1.1       mspo      699: 
                    700: Your domain should be now ready to work, enjoy.
                    701: 
1.14      gdt       702: Creating an unprivileged Linux domain (domU)
1.5       mspo      703: --------------------------------------------
1.1       mspo      704: 
                    705: Creating unprivileged Linux domains isn't much different from
                    706: unprivileged NetBSD domains, but there are some details to know.
                    707: 
                    708: First, the second parameter passed to the disk declaration (the '0x1' in
                    709: the example below)
                    710: 
1.3       mspo      711:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1       mspo      712: 
                    713: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49      gdt       714: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
                    715: So, hda1 which has major 3 and minor 1 on a Linux system will have
                    716: device number 0x301.  Alternatively, devices names can be used (hda,
                    717: hdb, ...)  as xentools has a table to map these names to devices
                    718: numbers.  To export a partition to a Linux guest we can use:
1.1       mspo      719: 
1.49      gdt       720:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
                    721:         root = "/dev/hda1 ro"
1.1       mspo      722: 
                    723: and it will appear as /dev/hda on the Linux system, and be used as root
                    724: partition.
                    725: 
1.49      gdt       726: To install the Linux system on the partition to be exported to the
                    727: guest domain, the following method can be used: install
                    728: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
                    729: that will be the root partition of your Linux domain, and mount it.
                    730: Then copy the files from a working Linux system, make adjustments in
                    731: `/etc` (fstab, network config).  It should also be possible to extract
                    732: binary packages such as .rpm or .deb directly to the mounted partition
                    733: using the appropriate tool, possibly running under NetBSD's Linux
                    734: emulation.  Once the filesystem has been populated, umount it.  If
                    735: desirable, the filesystem can be converted to ext3 using tune2fs -j.
                    736: It should now be possible to boot the Linux guest domain, using one of
                    737: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1       mspo      738: 
                    739: To get the linux console right, you need to add:
                    740: 
1.3       mspo      741:     extra = "xencons=tty1"
1.1       mspo      742: 
                    743: to your configuration since not all linux distributions auto-attach a
                    744: tty to the xen console.
                    745: 
1.14      gdt       746: Creating an unprivileged Solaris domain (domU)
1.5       mspo      747: ----------------------------------------------
1.1       mspo      748: 
1.50      gdt       749: See possibly outdated
                    750: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5       mspo      751: 
1.1       mspo      752: 
1.52    ! gdt       753: PCI passthrough: Using PCI devices in guest domains
        !           754: ---------------------------------------------------
1.1       mspo      755: 
1.52    ! gdt       756: The domain0 can give other domains access to selected PCI
        !           757: devices. This can allow, for example, a non-privileged domain to have
        !           758: access to a physical network interface or disk controller.  However,
        !           759: keep in mind that giving a domain access to a PCI device most likely
        !           760: will give the domain read/write access to the whole physical memory,
        !           761: as PCs don't have an IOMMU to restrict memory access to DMA-capable
        !           762: device.  Also, it's not possible to export ISA devices to non-domain0
        !           763: domains, which means that the primary VGA adapter can't be exported.
        !           764: A guest domain trying to access the VGA registers will panic.
        !           765: 
        !           766: If the domain0 is NetBSD, it has to be running Xen 3.1, as support has
        !           767: not been ported to later versions at this time.
        !           768: 
        !           769: For a PCI device to be exported to a domU, is has to be attached to
        !           770: the "pciback" driver in dom0.  Devices passed to the dom0 via the
        !           771: pciback.hide boot parameter will attach to "pciback" instead of the
        !           772: usual driver.  The list of devices is specified as "(bus:dev.func)",
1.5       mspo      773: where bus and dev are 2-digit hexadecimal numbers, and func a
                    774: single-digit number:
1.1       mspo      775: 
1.52    ! gdt       776:         pciback.hide=(00:0a.0)(00:06.0)
1.1       mspo      777: 
1.52    ! gdt       778: pciback devices should show up in the dom0's boot messages, and the
1.5       mspo      779: devices should be listed in the `/kern/xen/pci` directory.
1.1       mspo      780: 
1.52    ! gdt       781: PCI devices to be exported to a domU are listed in the "pci" array of
        !           782: the domU's config file, with the format "0000:bus:dev.func".
1.1       mspo      783: 
1.52    ! gdt       784:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1       mspo      785: 
1.52    ! gdt       786: In the domU an "xpci" device will show up, to which one or more pci
        !           787: busses will attach.  Then the PCI drivers will attach to PCI busses as
        !           788: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
        !           789: or any PCI drivers built in by default; you have to build your own
        !           790: kernel to use PCI devices in a domU.  Here's a kernel config example;
        !           791: note that only the "xpci" lines are unusual.
        !           792: 
        !           793:         include         "arch/i386/conf/XEN3_DOMU"
        !           794: 
        !           795:         # Add support for PCI busses to the XEN3_DOMU kernel
        !           796:         xpci* at xenbus ?
        !           797:         pci* at xpci ?
        !           798: 
        !           799:         # PCI USB controllers
        !           800:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
        !           801: 
        !           802:         # USB bus support
        !           803:         usb*    at uhci?
        !           804: 
        !           805:         # USB Hubs
        !           806:         uhub*   at usb?
        !           807:         uhub*   at uhub? port ? configuration ? interface ?
        !           808: 
        !           809:         # USB Mass Storage
        !           810:         umass*  at uhub? port ? configuration ? interface ?
        !           811:         wd*     at umass?
        !           812:         # SCSI controllers
        !           813:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
        !           814: 
        !           815:         # SCSI bus support (for both ahc and umass)
        !           816:         scsibus* at scsi?
        !           817: 
        !           818:         # SCSI devices
        !           819:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
        !           820:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
1.1       mspo      821: 
                    822: 
1.28      gdt       823: NetBSD as a domU in a VPS
                    824: =========================
                    825: 
                    826: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
                    827: hardware.  This section explains how to deal with Xen in a domU as a
                    828: virtual private server where you do not control or have access to the
                    829: dom0.
                    830: 
1.52    ! gdt       831: VPS operators provide varying degrees of access and mechanisms for
        !           832: configuration.  The big issue is usually how one controls which kernel
        !           833: is booted, because the kernel is nominally in the dom0 filesystem (to
        !           834: which VPS users do not normally have acesss).
        !           835: 
        !           836: A VPS user may want to compile a kernel for security updates, to run
        !           837: npf, run IPsec, or any other reason why someone would want to change
        !           838: their kernel.
        !           839: 
        !           840: One approach is to have an adminstrative interface to upload a kernel,
        !           841: or to select from a prepopulated list.
        !           842: 
        !           843: Otehr approaches are pvgrub and py-grub, which are ways to start a
        !           844: bootloader from the dom0 instead of the actual domU kernel, and for
        !           845: that loader to then load a kernel from the domU filesystem.  This is
        !           846: closer to a regular physical computer, where someone who controls a
        !           847: machine can replace the kernel.
        !           848: 
        !           849: prmgr and pvgrub
        !           850: ----------------
        !           851: 
1.28      gdt       852: TODO: Perhaps reference panix, prmgr, amazon as interesting examples.
1.52    ! gdt       853: Explain what prmgr does.
1.44      gdt       854: 
                    855: Using npf
                    856: ---------
                    857: 
                    858: In standard kernels, npf is a module, and thus cannot be loadeed in a
                    859: DOMU kernel.
                    860: 
                    861: TODO: explain how to compile npf into a custom kernel, answering:
                    862: http://mail-index.netbsd.org/netbsd-users/2014/12/26/msg015576.html

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