File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.99: download - view: text, annotated - select for diffs
Thu Mar 12 21:27:36 2015 UTC (6 years, 2 months ago) by gdt
Branches: MAIN
CVS tags: HEAD
Explain the i386/amd64 situation more clearly and more accurately.

    1: Introduction
    2: ============
    4: [![[Xen
    5: screenshot]](](
    7: Xen is a hypervisor (or virtual machine monitor) for x86 hardware
    8: (i686-class or higher), which supports running multiple guest
    9: operating systems on a single physical machine.  Xen is a Type 1 or
   10: bare-metal hypervisor; one uses the Xen kernel to control the CPU,
   11: memory and console, a dom0 operating system which mediates access to
   12: other hardware (e.g., disks, network, USB), and one or more domU
   13: operating systems which operate in an unprivileged virtualized
   14: environment.  IO requests from the domU systems are forwarded by the
   15: hypervisor (Xen) to the dom0 to be fulfilled.
   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.
   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.
   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.
   37: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   38: The dom0 can start one or more domUs.  (Booting is explained in detail
   39: in the dom0 section.)
   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
   44: and running domUs under it (NetBSD and other), and also running NetBSD
   45: as a domU in a VPS.
   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.
   52: NetBSD used to support Xen2; this has been removed.
   54: Prerequisites
   55: -------------
   57: Installing NetBSD/Xen is not extremely difficult, but it is more
   58: complex than a normal installation of NetBSD.
   59: In general, this HOWTO is occasionally overly restrictive about how
   60: things must be done, guiding the reader to stay on the established
   61: path when there are no known good reasons to stray.
   63: This HOWTO presumes a basic familiarity with the Xen system
   64: architecture.  This HOWTO presumes familiarity with installing NetBSD
   65: on i386/amd64 hardware and installing software from pkgsrc.
   66: See also the [Xen website](
   68: Versions of Xen and NetBSD
   69: ==========================
   71: Most of the installation concepts and instructions are independent
   72: of Xen version and NetBSD version.  This section gives advice on
   73: which version to choose.  Versions not in pkgsrc and older unsupported
   74: versions of NetBSD are intentionally ignored.
   76: Xen
   77: ---
   79: In NetBSD, xen is provided in pkgsrc, via matching pairs of packages
   80: xenkernel and xentools.  We will refer only to the kernel versions,
   81: but note that both packages must be installed together and must have
   82: matching versions.
   84: xenkernel3 and xenkernel33 provide Xen 3.1 and 3.3.  These no longer
   85: receive security patches and should not be used.  Xen 3.1 supports PCI
   86: passthrough.  Xen 3.1 supports non-PAE on i386.
   88: xenkernel41 provides Xen 4.1.  This is no longer maintained by Xen,
   89: but as of 2014-12 receives backported security patches.  It is a
   90: reasonable although trailing-edge choice.
   92: xenkernel42 provides Xen 4.2.  This is maintained by Xen, but old as
   93: of 2014-12.
   95: xenkernel45 provides Xen 4.5.  This is new to pkgsrc as of 2015-01 and
   96: not yet recommended for other than experimental/testing use.
   98: See also the [Xen Security Advisory page](
  100: Ideally newer versions of Xen will be added to pkgsrc.
  102: Note that NetBSD support is called XEN3.  It works with Xen 3 and Xen
  103: 4 because the hypercall interface has been stable.
  105: Xen command program
  106: -------------------
  108: Early Xen used a program called xm to manipulate the system from the
  109: dom0.  Starting in 4.1, a replacement program with similar behavior
  110: called xl is provided, but it does not work well in 4.1.  In 4.2, both
  111: xm and xl work fine.  4.4 is the last version that has xm.  You must
  112: choose one or the other, because it affects which daemons you run.
  114: NetBSD
  115: ------
  117: The netbsd-5, netbsd-6, netbsd-7, and -current branches are all
  118: reasonable choices, with more or less the same considerations for
  119: non-Xen use.  Therefore, netbsd-6 is recommended as the stable version
  120: of the most recent release for production use.  For those wanting to
  121: learn Xen or without production stability concerns, netbsd-7 is likely
  122: most appropriate.
  124: As of NetBSD 6, a NetBSD domU will support multiple vcpus.  There is
  125: no SMP support for NetBSD as dom0.  (The dom0 itself doesn't really
  126: need SMP; the lack of support is really a problem when using a dom0 as
  127: a normal computer.)
  129: Architecture
  130: ------------
  132: Xen itself can run on i386 or amd64 machines.  (Practically, almost
  133: any computer where one would want to run Xen today supports amd64.)
  135: Xen, the dom0 kernel, and each domU kernel can be either i386 or
  136: amd64.  When building a xenkernel package, one obtains i386 on an i386
  137: host, and amd64 on an amd64 host.  If the xen kernel is i386, then the
  138: dom0 kernel and all domU kernels must be i386.  With an amd64 xen
  139: kernel, an amd64 dom0 kernel is known to work, and an i386 dom0 kernel
  140: should in theory work.  An amd64 xen/dom0 is known to support both
  141: i386 and amd64 domUs.
  143: Except for an odd case with Xen 3.1, i386 dom0 and domU kernels must
  144: be PAE; these are built by default.  (Note that emacs (at least) fails
  145: if run on i386 with PAE when built without, and vice versa, presumably
  146: due to bugs in the undump code.)
  148: Because of the above, the standard approach is to use amd64 for the
  149: dom0.
  151: Xen 4.2 is the last version to support i386 as a host.  TODO: Clarify
  152: if this is about the CPU having to be amd64, or about the dom0 kernel
  153: having to be amd64.
  155: Stability
  156: ---------
  158: Mostly, NetBSD as a dom0 or domU is quite stable.
  159: However, there are some open PRs indicating problems.
  161:  - [PR 48125](
  162:  - [PR 47720](
  164: Note also that there are issues with sparse vnd(4) instances, but
  165: these are not about Xen.
  167: Recommendation
  168: --------------
  170: Therefore, this HOWTO recommends running xenkernel42 (and xentools42),
  171: xl, the NetBSD 6 stable branch, and to use an amd64 kernel as the
  172: dom0.  Either the i386 or amd64 of NetBSD may be used as domUs.
  174: Build problems
  175: --------------
  177: Ideally, all versions of Xen in pkgsrc would build on all versions of
  178: NetBSD on both i386 and amd64.  However, that isn't the case.  Besides
  179: aging code and aging compilers, qemu (included in xentools for HVM
  180: support) is difficult to build.  The following are known to work or FAIL:
  182:         xenkernel3 netbsd-5 amd64
  183:         xentools3 netbsd-5 amd64
  184:         xentools3=hvm netbsd-5 amd64 ????
  185:         xenkernel33 netbsd-5 amd64
  186:         xentools33 netbsd-5 amd64
  187:         xenkernel41 netbsd-5 amd64
  188:         xentools41 netbsd-5 amd64
  189:         xenkernel42 netbsd-5 amd64
  190:         xentools42 netbsd-5 amd64
  192:         xenkernel3 netbsd-6 i386 FAIL
  193:         xentools3 netbsd-6 i386
  194:         xentools3-hvm netbsd-6 i386 FAIL (dependencies fail)
  195:         xenkernel33 netbsd-6 i386
  196:         xentools33 netbsd-6 i386
  197:         xenkernel41 netbsd-6 i386
  198:         xentools41 netbsd-6 i386
  199:         xenkernel42 netbsd-6 i386
  200:         xentools42 netbsd-6 i386 *MIXED
  202: 	(all 3 and 33 seem to FAIL)
  203:         xenkernel41 netbsd-7 i386
  204:         xentools41 netbsd-7 i386
  205:         xenkernel42 netbsd-7 i386
  206:         xentools42 netbsd-7 i386 ??FAIL
  208: (*On netbsd-6 i386, there is a xentools42 in the 2014Q3 official builds,
  209: but it does not build for gdt.)
  211: NetBSD as a dom0
  212: ================
  214: NetBSD can be used as a dom0 and works very well.  The following
  215: sections address installation, updating NetBSD, and updating Xen.
  216: Note that it doesn't make sense to talk about installing a dom0 OS
  217: without also installing Xen itself.  We first address installing
  218: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
  219: NetBSD install to a dom0 install by just changing the kernel and boot
  220: configuration.
  222: For experimenting with Xen, a machine with as little as 1G of RAM and
  223: 100G of disk can work.  For running many domUs in productions, far
  224: more will be needed.
  226: Styles of dom0 operation
  227: ------------------------
  229: There are two basic ways to use Xen.  The traditional method is for
  230: the dom0 to do absolutely nothing other than providing support to some
  231: number of domUs.  Such a system was probably installed for the sole
  232: purpose of hosting domUs, and sits in a server room on a UPS.
  234: The other way is to put Xen under a normal-usage computer, so that the
  235: dom0 is what the computer would have been without Xen, perhaps a
  236: desktop or laptop.  Then, one can run domUs at will.  Purists will
  237: deride this as less secure than the previous approach, and for a
  238: computer whose purpose is to run domUs, they are right.  But Xen and a
  239: dom0 (without domUs) is not meaningfully less secure than the same
  240: things running without Xen.  One can boot Xen or boot regular NetBSD
  241: alternately with little problems, simply refraining from starting the
  242: Xen daemons when not running Xen.
  244: Note that NetBSD as dom0 does not support multiple CPUs.  This will
  245: limit the performance of the Xen/dom0 workstation approach.  In theory
  246: the only issue is that the "backend drivers" are not yet MPSAFE:
  249: Installation of NetBSD
  250: ----------------------
  252: First,
  253: [install NetBSD/amd64](/guide/inst/)
  254: just as you would if you were not using Xen.
  255: However, the partitioning approach is very important.
  257: If you want to use RAIDframe for the dom0, there are no special issues
  258: for Xen.  Typically one provides RAID storage for the dom0, and the
  259: domU systems are unaware of RAID.  The 2nd-stage loader bootxx_* skips
  260: over a RAID1 header to find /boot from a filesystem within a RAID
  261: partition; this is no different when booting Xen.
  263: There are 4 styles of providing backing storage for the virtual disks
  264: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN.
  266: With raw partitions, one has a disklabel (or gpt) partition sized for
  267: each virtual disk to be used by the domU.  (If you are able to predict
  268: how domU usage will evolve, please add an explanation to the HOWTO.
  269: Seriously, needs tend to change over time.)
  271: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
  272: for domU disks.  This is almost as efficient as raw disk partitions
  273: and more flexible.  Hence raw disk partitions should typically not
  274: be used.
  276: One can use files in the dom0 filesystem, typically created by dd'ing
  277: /dev/zero to create a specific size.  This is somewhat less efficient,
  278: but very convenient, as one can cp the files for backup, or move them
  279: between dom0 hosts.
  281: Finally, in theory one can place the files backing the domU disks in a
  282: SAN.  (This is an invitation for someone who has done this to add a
  283: HOWTO page.)
  285: Installation of Xen
  286: -------------------
  288: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
  289: pkgsrc (or another matching pair).
  290: See [the pkgsrc
  291: documentation]( for help with pkgsrc.
  293: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm.  More
  294: recent versions have HVM support integrated in the main xentools
  295: package.  It is entirely reasonable to run only PV guests.
  297: Next you need to install the selected Xen kernel itself, which is
  298: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz".  Copy it to /.
  299: For debugging, one may copy xen-debug.gz; this is conceptually similar
  300: to DIAGNOSTIC and DEBUG in NetBSD.  xen-debug.gz is basically only
  301: useful with a serial console.  Then, place a NetBSD XEN3_DOM0 kernel
  302: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  303: of a NetBSD build.  If using i386, use
  304: releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.  (If using Xen
  305: 3.1 and i386, you may use XEN3_DOM0 with the non-PAE Xen.  But you
  306: should not use Xen 3.1.)  Both xen and the NetBSD kernel may be (and
  307: typically are) left compressed.
  309: In a dom0 kernel, kernfs is mandatory for xend to comunicate with the
  310: kernel, so ensure that /kern is in fstab.  TODO: Say this is default,
  311: or file a PR and give a reference.
  313: Because you already installed NetBSD, you have a working boot setup
  314: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
  315: beginning of your root filesystem, /boot present, and likely
  316: /boot.cfg.  (If not, fix before continuing!)
  318: Add a line to to /boot.cfg to boot Xen.  See boot.cfg(5) for an
  319: example.  The basic line is
  321:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
  323: which specifies that the dom0 should have 256M, leaving the rest to be
  324: allocated for domUs.  To use a serial console, use
  326:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=256M console=com1 com1=9600,8n1
  328: which will use the first serial port for Xen (which counts starting
  329: from 1), forcing speed/parity, and also for NetBSD (which counts
  330: starting at 0).  In an attempt to add performance, one can also add
  332:         dom0_max_vcpus=1 dom0_vcpus_pin
  334: to force only one vcpu to be provided (since NetBSD dom0 can't use
  335: more) and to pin that vcpu to a physical cpu.  TODO: benchmark this.
  337: Xen has [many boot
  338: options](,
  339: and other tham dom0 memory and max_vcpus, they are generally not
  340: necessary.
  342: As with non-Xen systems, you should have a line to boot /netbsd (a
  343: kernel that works without Xen) and fallback versions of the non-Xen
  344: kernel, Xen, and the dom0 kernel.
  346: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  347: than GENERIC without Xen.
  349: Using grub (historic)
  350: ---------------------
  352: Before NetBSD's native bootloader could support Xen, the use of
  353: grub was recommended.  If necessary, see the
  354: [old grub information](/ports/xen/howto-grub/).
  356: The [HowTo on Installing into
  357: RAID-1](
  358: explains how to set up booting a dom0 with Xen using grub with
  359: NetBSD's RAIDframe.  (This is obsolete with the use of NetBSD's native
  360: boot.)
  362: Configuring Xen
  363: ---------------
  365: Xen logs will be in /var/log/xen.
  367: Now, you have a system that will boot Xen and the dom0 kernel, but not
  368: do anything else special.  Make sure that you have rebooted into Xen.
  369: There will be no domUs, and none can be started because you still have
  370: to configure the dom0 tools.  The daemons which should be run vary
  371: with Xen version and with whether one is using xm or xl.  Note that
  372: xend is for supporting "xm", and should only be used if you plan on
  373: using "xm".  Do NOT enable xend if you plan on using "xl" as it will
  374: cause problems.  Running xl without xencommons=YES (and starting it)
  375: will result in a hang (so don't do that; follow the HOWTO!).
  377: The installation of NetBSD should already have created devices for xen
  378: (xencons, xenevt), but if they are not present, create them:
  380:         cd /dev && sh MAKEDEV xen
  382: TODO: Give 3.1 advice (or remove it from pkgsrc).
  384: For 3.3 (and thus xm), add to rc.conf (but note that you should have
  385: installed 4.1 or 4.2):
  387:         xend=YES
  388:         xenbackendd=YES
  390: For 4.1 (and thus xm; xl is believed not to work well), add to rc.conf:
  392:         xencommons=YES
  393:         xend=YES
  395: (If you are using xentools41 from before 2014-12-26, change
  396: rc.d/xendomains to use xm rather than xl.)
  398: For 4.2 with xm, add to rc.conf
  400:         xencommons=YES
  401:         xend=YES
  403: For 4.2 with xl, add to rc.conf:
  405:         xencommons=YES
  406:         TODO: explain if there is a xend replacement
  408: For 4.5 (and thus with xl), add to rc.conf:
  410:         xencommons=YES
  411:         TODO: explain if there is a xend replacement
  413: TODO: Recommend for/against xen-watchdog.
  415: After you have configured the daemons and either started them (in the
  416: order given) or rebooted, use xm or xl to inspect Xen's boot messages,
  417: available resources, and running domains.  An example with xm follows:
  419:         # xm dmesg
  420: 	[xen's boot info]
  421:         # xm info
  422: 	[available memory, etc.]
  423:         # xm list
  424:         Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  425:         Domain-0           0       64    0  r----     58.1
  427: With xl, the commands are the same, and the output may be slightly
  428: different.  TODO: add example output for xl before the xm example,
  429: after confirming on 4.2 and resolving the TODO about rc.conf.
  431: ### Issues with xencommons
  433: xencommons starts xenstored, which stores data on behalf of dom0 and
  434: domUs.  It does not currently work to stop and start xenstored.
  435: Certainly all domUs should be shutdown first, following the sort order
  436: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  437: and is not notified when xenstored exits, leading to not recreating
  438: the state when the new xenstored starts.  Until there's a mechanism to
  439: make this work, one should not expect to be able to restart xenstored
  440: (and thus xencommons).  There is currently no reason to expect that
  441: this will get fixed any time soon.
  443: anita (for testing NetBSD)
  444: --------------------------
  446: With the setup so far (assuming 4.2/xl), one should be able to run
  447: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  448: root, because anita must create a domU):
  450:         anita --vmm=xl test file:///usr/obj/i386/
  452: Alternatively, one can use --vmm=xm to use xm-based domU creation
  453: instead (and must, on Xen <= 4.1).   TODO: confirm that anita xl really works.
  455: Xen-specific NetBSD issues
  456: --------------------------
  458: There are (at least) two additional things different about NetBSD as a
  459: dom0 kernel compared to hardware.
  461: One is that modules are not usable in DOM0 kernels, so one must
  462: compile in what's needed.  It's not really that modules cannot work,
  463: but that modules must be built for XEN3_DOM0 because some of the
  464: defines change and the normal module builds don't do this.  Basically,
  465: enabling Xen changes the kernel ABI, and the module build system
  466: doesn't cope with this.
  468: The other difference is that XEN3_DOM0 does not have exactly the same
  469: options as GENERIC.  While it is debatable whether or not this is a
  470: bug, users should be aware of this and can simply add missing config
  471: items if desired.
  473: Updating NetBSD in a dom0
  474: -------------------------
  476: This is just like updating NetBSD on bare hardware, assuming the new
  477: version supports the version of Xen you are running.  Generally, one
  478: replaces the kernel and reboots, and then overlays userland binaries
  479: and adjusts /etc.
  481: Note that one must update both the non-Xen kernel typically used for
  482: rescue purposes and the DOM0 kernel used with Xen.
  484: Converting from grub to /boot
  485: -----------------------------
  487: These instructions were [TODO: will be] used to convert a system from
  488: grub to /boot.  The system was originally installed in February of
  489: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  490: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  491: 4.1 and grub, much like the message linked earlier in the grub
  492: section.
  494:         # Install mbr bootblocks on both disks. 
  495:         fdisk -i /dev/rwd0d
  496:         fdisk -i /dev/rwd1d
  497:         # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  498:         installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  499:         installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  500:         # Install secondary boot loader
  501:         cp -p /usr/mdec/boot /
  502:         # Create boog.cfg following earlier guidance:
  503:         menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
  504:         menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=256M
  505:         menu=GENERIC:boot
  506:         menu=GENERIC single-user:boot -s
  507:         menu=GENERIC.ok:boot netbsd.ok
  508:         menu=GENERIC.ok single-user:boot netbsd.ok -s
  509:         menu=Drop to boot prompt:prompt
  510:         default=1
  511:         timeout=30
  513: TODO: actually do this and fix it if necessary.
  515: Updating Xen versions
  516: ---------------------
  518: Updating Xen is conceptually not difficult, but can run into all the
  519: issues found when installing Xen.  Assuming migration from 4.1 to 4.2,
  520: remove the xenkernel41 and xentools41 packages and install the
  521: xenkernel42 and xentools42 packages.  Copy the 4.2 xen.gz to /.
  523: Ensure that the contents of /etc/rc.d/xen* are correct.  Enable the
  524: correct set of daemons.  Ensure that the domU config files are valid
  525: for the new version.
  527: Hardware known to work
  528: ----------------------
  530: Arguably, this section is misplaced, and there should be a page of
  531: hardware that runs NetBSD/amd64 well, with the mostly-well-founded
  532: assumption that NetBSD/xen runs fine on any modern hardware that
  533: NetBSD/amd64 runs well on.  Until then, we give motherboard/CPU/RAM
  534: triples to aid those choosing a motherboard.  Note that Xen systems
  535: usually do not run X, so a listing here does not imply that X works at
  536: all.
  538:         Supermicro X9SRL-F, Xeon E5-1650 v2, 96 GiB ECC
  539:         Supermicro ??, Atom C2758 (8 core), 32 GiB ECC
  540:         ASUS M5A78L-M/USB3 AM3+ microATX, AMD Piledriver X8 4000MHz, 16 GiB ECC
  542: Older hardware:
  544:         Intel D915GEV, Pentium4 CPU 3.40GHz, 4GB 533MHz Synchronous DDR2
  546: Running Xen under qemu
  547: ----------------------
  549: The astute reader will note that this section is somewhat twisted.
  550: However, it can be useful to run Xen under qemu either because the
  551: version of NetBSD as a dom0 does not run on the hardware in use, or to
  552: generate automated test cases involving Xen.
  554: In 2015-01, the following combination was reported to mostly work:
  556:         host OS: NetBSD/amd64 6.1.4
  557:         qemu: 2.2.0 from pkgsrc
  558:         Xen kernel: xenkernel42-4.2.5nb1 from pkgsrc
  559:         dom0 kernel: NetBSD/amd64 6.1.5
  560:         Xen tools: xentools42-4.2.5 from pkgsrc
  562: See [PR 47720]( for a problem with dom0
  563: shutdown.
  565: Unprivileged domains (domU)
  566: ===========================
  568: This section describes general concepts about domUs.  It does not
  569: address specific domU operating systems or how to install them.  The
  570: config files for domUs are typically in /usr/pkg/etc/xen, and are
  571: typically named so that the file name, domU name and the domU's host
  572: name match.
  574: The domU is provided with cpu and memory by Xen, configured by the
  575: dom0.  The domU is provided with disk and network by the dom0,
  576: mediated by Xen, and configured in the dom0.
  578: Entropy in domUs can be an issue; physical disks and network are on
  579: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  581: Config files
  582: ------------
  584: There is no good order to present config files and the concepts
  585: surrounding what is being configured.  We first show an example config
  586: file, and then in the various sections give details.
  588: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
  589: for a large number of well-commented examples, mostly for running
  590: GNU/Linux.
  592: The following is an example minimal domain configuration file
  593: "/usr/pkg/etc/xen/foo".  It is (with only a name change) an actual
  594: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
  595: i386 domU).  The domU serves as a network file server.
  597:         # -*- mode: python; -*-
  599:         kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  600:         memory = 1024
  601:         vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  602:         disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  603:                  'file:/n0/xen/foo-wd1,0x1,w' ]
  605: The domain will have the same name as the file.  The kernel has the
  606: host/domU name in it, so that on the dom0 one can update the various
  607: domUs independently.  The vif line causes an interface to be provided,
  608: with a specific mac address (do not reuse MAC addresses!), in bridge
  609: mode.  Two disks are provided, and they are both writable; the bits
  610: are stored in files and Xen attaches them to a vnd(4) device in the
  611: dom0 on domain creation.  The system treates xbd0 as the boot device
  612: without needing explicit configuration.
  614: By default xm looks for domain config files in /usr/pkg/etc/xen.  Note
  615: that "xm create" takes the name of a config file, while other commands
  616: take the name of a domain.  To create the domain, connect to the
  617: console, create the domain while attaching the console, shutdown the
  618: domain, and see if it has finished stopping, do (or xl with Xen >=
  619: 4.2):
  621:         xm create foo
  622:         xm console foo
  623:         xm create -c foo
  624:         xm shutdown foo
  625:         xm list
  627: Typing ^] will exit the console session.  Shutting down a domain is
  628: equivalent to pushing the power button; a NetBSD domU will receive a
  629: power-press event and do a clean shutdown.  Shutting down the dom0
  630: will trigger controlled shutdowns of all configured domUs.
  632: domU kernels
  633: ------------
  635: On a physical computer, the BIOS reads sector 0, and a chain of boot
  636: loaders finds and loads a kernel.  Normally this comes from the root
  637: filesystem.  With Xen domUs, the process is totally different.  The
  638: normal path is for the domU kernel to be a file in the dom0's
  639: filesystem.  At the request of the dom0, Xen loads that kernel into a
  640: new domU instance and starts execution.  While domU kernels can be
  641: anyplace, reasonable places to store domU kernels on the dom0 are in /
  642: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
  643: config files), or in /u0/xen (where the vdisks are).
  645: Note that loading the domU kernel from the dom0 implies that boot
  646: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
  647: See the VPS section near the end for discussion of alternate ways to
  648: obtain domU kernels.
  650: CPU and memory
  651: --------------
  653: A domain is provided with some number of vcpus, less than the number
  654: of cpus seen by the hypervisor.  (For a dom0, this is controlled by
  655: the boot argument "dom0_max_vcpus=1".)  For a domU, it is controlled
  656: from the config file by the "vcpus = N" directive.
  658: A domain is provided with memory; this is controlled in the config
  659: file by "memory = N" (in megabytes).  In the straightforward case, the
  660: sum of the the memory allocated to the dom0 and all domUs must be less
  661: than the available memory.
  663: Xen also provides a "balloon" driver, which can be used to let domains
  664: use more memory temporarily.  TODO: Explain better, and explain how
  665: well it works with NetBSD.
  667: Virtual disks
  668: -------------
  670: With the file/vnd style, typically one creates a directory,
  671: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
  672: domUs.  Then, for each domU disk, one writes zeros to a file that then
  673: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
  674: for the first virtual disk for the domU called foo.  Writing zeros to
  675: the file serves two purposes.  One is that preallocating the contents
  676: improves performance.  The other is that vnd on sparse files has
  677: failed to work.  TODO: give working/notworking NetBSD versions for
  678: sparse vnd.  Note that the use of file/vnd for Xen is not really
  679: different than creating a file-backed virtual disk for some other
  680: purpose, except that xentools handles the vnconfig commands.  To
  681: create an empty 4G virtual disk, simply do
  683:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
  685: Do not use qemu-img-xen, because this will create sparse file.  There
  686: have been recent (2015) reports of sparse vnd(4) devices causing
  687: lockups, but there is apparently no PR.
  689: With the lvm style, one creates logical devices.  They are then used
  690: similarly to vnds.  TODO: Add an example with lvm.
  692: In domU config files, the disks are defined as a sequence of 3-tuples.
  693: The first element is "method:/path/to/disk".  Common methods are
  694: "file:" for file-backed vnd. and "phy:" for something that is already
  695: a (TODO: character or block) device.
  697: The second element is an artifact of how virtual disks are passed to
  698: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  699: are given a device name to associate with the disk, and values like
  700: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  701: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
  702: second argument.  The name given is converted to a major/minor by
  703: calling stat(2) on the name in /dev and this is passed to the domU.
  704: In the general case, the dom0 and domU can be different operating
  705: systems, and it is an unwarranted assumption that they have consistent
  706: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  707: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  708: for the second works fine and avoids this issue.  For a GNU/Linux
  709: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  710: /dev/hda1.
  712: The third element is "w" for writable disks, and "r" for read-only
  713: disks.
  715: Virtual Networking
  716: ------------------
  718: Xen provides virtual ethernets, each of which connects the dom0 and a
  719: domU.  For each virtual network, there is an interface "xvifN.M" in
  720: the dom0, and in domU index N, a matching interface xennetM (NetBSD
  721: name).  The interfaces behave as if there is an Ethernet with two
  722: adaptors connected.  From this primitive, one can construct various
  723: configurations.  We focus on two common and useful cases for which
  724: there are existing scripts: bridging and NAT.
  726: With bridging (in the example above), the domU perceives itself to be
  727: on the same network as the dom0.  For server virtualization, this is
  728: usually best.  Bridging is accomplished by creating a bridge(4) device
  729: and adding the dom0's physical interface and the various xvifN.0
  730: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  731: config file.  The bridge must be set up already in the dom0; an
  732: example /etc/ifconfig.bridge0 is:
  734:         create
  735:         up
  736:         !brconfig bridge0 add wm0
  738: With NAT, the domU perceives itself to be behind a NAT running on the
  739: dom0.  This is often appropriate when running Xen on a workstation.
  740: TODO: NAT appears to be configured by "vif = [ '' ]".
  742: The MAC address specified is the one used for the interface in the new
  743: domain.  The interface in dom0 will use this address XOR'd with
  744: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  746: Sizing domains
  747: --------------
  749: Modern x86 hardware has vast amounts of resources.  However, many
  750: virtual servers can function just fine on far less.  A system with
  751: 256M of RAM and a 4G disk can be a reasonable choice.  Note that it is
  752: far easier to adjust virtual resources than physical ones.  For
  753: memory, it's just a config file edit and a reboot.  For disk, one can
  754: create a new file and vnconfig it (or lvm), and then dump/restore,
  755: just like updating physical disks, but without having to be there and
  756: without those pesky connectors.
  758: Starting domains automatically
  759: ------------------------------
  761: To start domains foo at bar at boot and shut them down cleanly on dom0
  762: shutdown, in rc.conf add:
  764:         xendomains="foo bar"
  766: Note that earlier versions of the xentools41 xendomains rc.d scripth
  767: usd xl, when one should use xm with 4.1.
  769: Creating specific unprivileged domains (domU)
  770: =============================================
  772: Creating domUs is almost entirely independent of operating system.  We
  773: have already presented the basics of config files.  Note that you must
  774: have already completed the dom0 setup so that "xl list" (or "xm list")
  775: works.
  777: Creating an unprivileged NetBSD domain (domU)
  778: ---------------------------------------------
  780: See the earlier config file, and adjust memory.  Decide on how much
  781: storage you will provide, and prepare it (file or lvm).
  783: While the kernel will be obtained from the dom0 filesystem, the same
  784: file should be present in the domU as /netbsd so that tools like
  785: savecore(8) can work.   (This is helpful but not necessary.)
  787: The kernel must be specifically for Xen and for use as a domU.  The
  788: i386 and amd64 provide the following kernels:
  790:         i386 XEN3_DOMU
  791:         i386 XEN3PAE_DOMU
  792:         amd64 XEN3_DOMU
  794: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
  795: use the PAE version of the i386 kernel.
  797: This will boot NetBSD, but this is not that useful if the disk is
  798: empty.  One approach is to unpack sets onto the disk outside of xen
  799: (by mounting it, just as you would prepare a physical disk for a
  800: system you can't run the installer on).
  802: A second approach is to run an INSTALL kernel, which has a miniroot
  803: and can load sets from the network.  To do this, copy the INSTALL
  804: kernel to / and change the kernel line in the config file to:
  806:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  808: Then, start the domain as "xl create -c configname".
  810: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  811: line should be used in the config file.
  813:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  815: After booting the domain, the option to install via CDROM may be
  816: selected.  The CDROM device should be changed to `xbd1d`.
  818: Once done installing, "halt -p" the new domain (don't reboot or halt,
  819: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  820: config file), switch the config file back to the XEN3_DOMU kernel,
  821: and start the new domain again. Now it should be able to use "root on
  822: xbd0a" and you should have a, functional NetBSD domU.
  824: TODO: check if this is still accurate.
  825: When the new domain is booting you'll see some warnings about *wscons*
  826: and the pseudo-terminals. These can be fixed by editing the files
  827: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  828: `/etc/ttys`, except *console*, like this:
  830:     console "/usr/libexec/getty Pc"         vt100   on secure
  831:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  832:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  833:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  834:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  836: Finally, all screens must be commented out from `/etc/wscons.conf`.
  838: It is also desirable to add
  840:         powerd=YES
  842: in rc.conf. This way, the domain will be properly shut down if
  843: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  845: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  846: filesystem.  However, various programs (e.g. netstat) will use that
  847: kernel to look up symbols to read from kernel virtual memory.  If
  848: /netbsd is not the running kernel, those lookups will fail.  (This is
  849: not really a Xen-specific issue, but because the domU kernel is
  850: obtained from the dom0, it is far more likely to be out of sync or
  851: missing with Xen.)
  853: Creating an unprivileged Linux domain (domU)
  854: --------------------------------------------
  856: Creating unprivileged Linux domains isn't much different from
  857: unprivileged NetBSD domains, but there are some details to know.
  859: First, the second parameter passed to the disk declaration (the '0x1' in
  860: the example below)
  862:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  864: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  865: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  866: So, hda1 which has major 3 and minor 1 on a Linux system will have
  867: device number 0x301.  Alternatively, devices names can be used (hda,
  868: hdb, ...)  as xentools has a table to map these names to devices
  869: numbers.  To export a partition to a Linux guest we can use:
  871:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  872:         root = "/dev/hda1 ro"
  874: and it will appear as /dev/hda on the Linux system, and be used as root
  875: partition.
  877: To install the Linux system on the partition to be exported to the
  878: guest domain, the following method can be used: install
  879: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  880: that will be the root partition of your Linux domain, and mount it.
  881: Then copy the files from a working Linux system, make adjustments in
  882: `/etc` (fstab, network config).  It should also be possible to extract
  883: binary packages such as .rpm or .deb directly to the mounted partition
  884: using the appropriate tool, possibly running under NetBSD's Linux
  885: emulation.  Once the filesystem has been populated, umount it.  If
  886: desirable, the filesystem can be converted to ext3 using tune2fs -j.
  887: It should now be possible to boot the Linux guest domain, using one of
  888: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  890: To get the linux console right, you need to add:
  892:     extra = "xencons=tty1"
  894: to your configuration since not all linux distributions auto-attach a
  895: tty to the xen console.
  897: Creating an unprivileged Solaris domain (domU)
  898: ----------------------------------------------
  900: See possibly outdated
  901: [Solaris domU instructions](/ports/xen/howto-solaris/).
  904: PCI passthrough: Using PCI devices in guest domains
  905: ---------------------------------------------------
  907: The dom0 can give other domains access to selected PCI
  908: devices. This can allow, for example, a non-privileged domain to have
  909: access to a physical network interface or disk controller.  However,
  910: keep in mind that giving a domain access to a PCI device most likely
  911: will give the domain read/write access to the whole physical memory,
  912: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  913: device.  Also, it's not possible to export ISA devices to non-dom0
  914: domains, which means that the primary VGA adapter can't be exported.
  915: A guest domain trying to access the VGA registers will panic.
  917: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  918: not been ported to later versions at this time.
  920: For a PCI device to be exported to a domU, is has to be attached to
  921: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  922: pciback.hide boot parameter will attach to "pciback" instead of the
  923: usual driver.  The list of devices is specified as "(bus:dev.func)",
  924: where bus and dev are 2-digit hexadecimal numbers, and func a
  925: single-digit number:
  927:         pciback.hide=(00:0a.0)(00:06.0)
  929: pciback devices should show up in the dom0's boot messages, and the
  930: devices should be listed in the `/kern/xen/pci` directory.
  932: PCI devices to be exported to a domU are listed in the "pci" array of
  933: the domU's config file, with the format "0000:bus:dev.func".
  935:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  937: In the domU an "xpci" device will show up, to which one or more pci
  938: busses will attach.  Then the PCI drivers will attach to PCI busses as
  939: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  940: or any PCI drivers built in by default; you have to build your own
  941: kernel to use PCI devices in a domU.  Here's a kernel config example;
  942: note that only the "xpci" lines are unusual.
  944:         include         "arch/i386/conf/XEN3_DOMU"
  946:         # Add support for PCI busses to the XEN3_DOMU kernel
  947:         xpci* at xenbus ?
  948:         pci* at xpci ?
  950:         # PCI USB controllers
  951:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  953:         # USB bus support
  954:         usb*    at uhci?
  956:         # USB Hubs
  957:         uhub*   at usb?
  958:         uhub*   at uhub? port ? configuration ? interface ?
  960:         # USB Mass Storage
  961:         umass*  at uhub? port ? configuration ? interface ?
  962:         wd*     at umass?
  963:         # SCSI controllers
  964:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  966:         # SCSI bus support (for both ahc and umass)
  967:         scsibus* at scsi?
  969:         # SCSI devices
  970:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  971:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  974: NetBSD as a domU in a VPS
  975: =========================
  977: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  978: hardware.  This section explains how to deal with Xen in a domU as a
  979: virtual private server where you do not control or have access to the
  980: dom0.  This is not intended to be an exhaustive list of VPS providers;
  981: only a few are mentioned that specifically support NetBSD.
  983: VPS operators provide varying degrees of access and mechanisms for
  984: configuration.  The big issue is usually how one controls which kernel
  985: is booted, because the kernel is nominally in the dom0 filesystem (to
  986: which VPS users do not normally have acesss).  A second issue is how
  987: to install NetBSD.
  988: A VPS user may want to compile a kernel for security updates, to run
  989: npf, run IPsec, or any other reason why someone would want to change
  990: their kernel.
  992: One approach is to have an adminstrative interface to upload a kernel,
  993: or to select from a prepopulated list.  Other approaches are pygrub
  994: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  995: kernel from the domU filesystem.  This is closer to a regular physical
  996: computer, where someone who controls a machine can replace the kernel.
  998: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  999: multiple vcpus, and it is typical for VPS providers to enable multiple
 1000: CPUs for NetBSD domUs.
 1002: pygrub
 1003: -------
 1005: pygrub runs in the dom0 and looks into the domU filesystem.  This
 1006: implies that the domU must have a kernel in a filesystem in a format
 1007: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
 1008: interest.
 1010: pvgrub
 1011: ------
 1013: pvgrub is a version of grub that uses PV operations instead of BIOS
 1014: calls.  It is booted from the dom0 as the domU kernel, and then reads
 1015: /grub/menu.lst and loads a kernel from the domU filesystem.
 1017: [Panix]( lets users use pvgrub.  Panix reports
 1018: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
 1019: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
 1020: page](, which describes only
 1021: Linux but should be updated to cover NetBSD :-).
 1023: []( also lets users with pvgrub to boot
 1024: their own kernel.  See then [ NetBSD
 1025: HOWTO](
 1026: (which is in need of updating).
 1028: It appears that [grub's FFS
 1029: code](
 1030: does not support all aspects of modern FFS, but there are also reports
 1031: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
 1032: partition for the kernel with the intent that grub can understand it,
 1033: which leads to /netbsd not being the actual kernel.  One must remember
 1034: to update the special boot partiion.
 1036: Amazon
 1037: ------
 1039: See the [Amazon EC2 page](../amazon_ec2/).
 1041: Using npf
 1042: ---------
 1044: In standard kernels, npf is a module, and thus cannot be loaded in a
 1045: DOMU kernel.
 1047: TODO: Explain how to compile npf into a custom kernel, answering (but
 1048: note that the problem was caused by not booting the right kernel)
 1049: [this email to
 1050: netbsd-users](
 1052: TODO items for improving NetBSD/xen
 1053: ===================================
 1055: * Make the NetBSD dom0 kernel work with SMP.
 1056: * Test the Xen 4.5 packages adequately to be able to recommend them as
 1057:   the standard approach.
 1058: * Get PCI passthrough working on Xen 4.5
 1059: * Get pvgrub into pkgsrc, either via xentools or separately.
 1060: * grub
 1061:   * Check/add support to pkgsrc grub2 for UFS2 and arbitrary
 1062:     fragsize/blocksize (UFS2 support may be present; the point is to
 1063:     make it so that with any UFS1/UFS2 filesystem setup that works
 1064:     with NetBSD grub will also work).
 1065:     See [pkg/40258](
 1066:   * Push patches upstream.
 1067:   * Get UFS2 patches into pvgrub.
 1068: * Add support for PV ops to a version of /boot, and make it usable as
 1069:   a kernel in Xen, similar to pvgrub.
 1070: * Solve somehow the issue with modules for GENERIC not being loadable
 1071:   in a Xen dom0 or domU kernel.
 1073: Random pointers
 1074: ===============
 1076: TODO: This section contains links from elsewhere not yet integrated
 1077: into the HOWTO.
 1079: *
 1080: *

CVSweb for NetBSD wikisrc <> software: FreeBSD-CVSweb