File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.87: download - view: text, annotated - select for diffs
Wed Jan 28 19:27:11 2015 UTC (4 years, 10 months ago) by gdt
Branches: MAIN
CVS tags: HEAD
Admonish to start xencommons before xl.

    1: Introduction
    2: ============
    3: 
    4: [![[Xen
    5: screenshot]](http://www.netbsd.org/gallery/in-Action/hubertf-xens.png)](http://www.netbsd.org/gallery/in-Action/hubertf-xen.png)
    6: 
    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.
   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: 
   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.
   36: 
   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.)
   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
   44: and running domUs under it (NetBSD and other), and also running NetBSD
   45: as a domU in a VPS.
   46: 
   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: 
   52: NetBSD used to support Xen2; this has been removed.
   53: 
   54: Prerequisites
   55: -------------
   56: 
   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.
   62: 
   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](http://www.xenproject.org/).
   67: 
   68: Versions of Xen and NetBSD
   69: ==========================
   70: 
   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.
   75: 
   76: Xen
   77: ---
   78: 
   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.
   83: 
   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.
   87: 
   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.
   91: 
   92: xenkernel42 provides Xen 4.2.  This is maintained by Xen, but old as
   93: of 2014-12.
   94: 
   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.
   97: 
   98: Ideally newer versions of Xen will be added to pkgsrc.
   99: 
  100: Note that NetBSD support is called XEN3.  It works with Xen 3 and Xen
  101: 4 because the hypercall interface has been stable.
  102: 
  103: Xen command program
  104: -------------------
  105: 
  106: Early Xen used a program called xm to manipulate the system from the
  107: dom0.  Starting in 4.1, a replacement program with similar behavior
  108: called xl is provided, but it does not work well in 4.1.  In 4.2, both
  109: xm and xl work fine.  4.4 is the last version that has xm.  You must
  110: choose one or the other, because it affects which daemons you run.
  111: 
  112: NetBSD
  113: ------
  114: 
  115: The netbsd-5, netbsd-6, netbsd-7, and -current branches are all
  116: reasonable choices, with more or less the same considerations for
  117: non-Xen use.  Therefore, netbsd-6 is recommended as the stable version
  118: of the most recent release for production use.  For those wanting to
  119: learn Xen or without production stability concerns, netbsd-7 is likely
  120: most appropriate.
  121: 
  122: As of NetBSD 6, a NetBSD domU will support multiple vcpus.  There is
  123: no SMP support for NetBSD as dom0.  (The dom0 itself doesn't really
  124: need SMP; the lack of support is really a problem when using a dom0 as
  125: a normal computer.)
  126: 
  127: Architecture
  128: ------------
  129: 
  130: Xen itself can run on i386 or amd64 machines.  (Practically, almost
  131: any computer where one would want to run Xen supports amd64.)  If
  132: using an i386 NetBSD kernel for the dom0, PAE is required (PAE
  133: versions are built by default).  While i386 dom0 works fine, amd64 is
  134: recommended as more normal.
  135: 
  136: Xen 4.2 is the last version to support i386 as a host.  TODO: Clarify
  137: if this is about the CPU having to be amd64, or about the dom0 kernel
  138: having to be amd64.
  139: 
  140: One can then run i386 domUs and amd64 domUs, in any combination.  If
  141: running an i386 NetBSD kernel as a domU, the PAE version is required.
  142: (Note that emacs (at least) fails if run on i386 with PAE when built
  143: without, and vice versa, presumably due to bugs in the undump code.)
  144: 
  145: Recommendation
  146: --------------
  147: 
  148: Therefore, this HOWTO recommends running xenkernel42 (and xentools42),
  149: xl, the NetBSD 6 stable branch, and to use an amd64 kernel as the
  150: dom0.  Either the i386 or amd64 of NetBSD may be used as domUs.
  151: 
  152: Build problems
  153: --------------
  154: 
  155: Ideally, all versions of Xen in pkgsrc would build on all versions of
  156: NetBSD on both i386 and amd64.  However, that isn't the case.  Besides
  157: aging code and aging compilers, qemu (included in xentools for HVM
  158: support) is difficult to build.  The following are known to work or FAIL:
  159: 
  160:         xenkernel3 netbsd-5 amd64
  161:         xentools3 netbsd-5 amd64
  162:         xentools3=hvm netbsd-5 amd64 ????
  163:         xenkernel33 netbsd-5 amd64
  164:         xentools33 netbsd-5 amd64
  165:         xenkernel41 netbsd-5 amd64
  166:         xentools41 netbsd-5 amd64
  167:         xenkernel42 netbsd-5 amd64
  168:         xentools42 netbsd-5 amd64
  169: 
  170:         xenkernel3 netbsd-6 i386 FAIL
  171:         xentools3 netbsd-6 i386
  172:         xentools3-hvm netbsd-6 i386 FAIL (dependencies fail)
  173:         xenkernel33 netbsd-6 i386
  174:         xentools33 netbsd-6 i386
  175:         xenkernel41 netbsd-6 i386
  176:         xentools41 netbsd-6 i386
  177:         xenkernel42 netbsd-6 i386
  178:         xentools42 netbsd-6 i386 *MIXED
  179: 
  180: 	(all 3 and 33 seem to FAIL)
  181:         xenkernel41 netbsd-7 i386
  182:         xentools41 netbsd-7 i386
  183:         xenkernel42 netbsd-7 i386
  184:         xentools42 netbsd-7 i386 ??FAIL
  185: 
  186: (*On netbsd-6 i386, there is a xentools42 in the 2014Q3 official builds,
  187: but it does not build for gdt.)
  188: 
  189: NetBSD as a dom0
  190: ================
  191: 
  192: NetBSD can be used as a dom0 and works very well.  The following
  193: sections address installation, updating NetBSD, and updating Xen.
  194: Note that it doesn't make sense to talk about installing a dom0 OS
  195: without also installing Xen itself.  We first address installing
  196: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
  197: NetBSD install to a dom0 install by just changing the kernel and boot
  198: configuration.
  199: 
  200: For experimenting with Xen, a machine with as little as 1G of RAM and
  201: 100G of disk can work.  For running many domUs in productions, far
  202: more will be needed.
  203: 
  204: Styles of dom0 operation
  205: ------------------------
  206: 
  207: There are two basic ways to use Xen.  The traditional method is for
  208: the dom0 to do absolutely nothing other than providing support to some
  209: number of domUs.  Such a system was probably installed for the sole
  210: purpose of hosting domUs, and sits in a server room on a UPS.
  211: 
  212: The other way is to put Xen under a normal-usage computer, so that the
  213: dom0 is what the computer would have been without Xen, perhaps a
  214: desktop or laptop.  Then, one can run domUs at will.  Purists will
  215: deride this as less secure than the previous approach, and for a
  216: computer whose purpose is to run domUs, they are right.  But Xen and a
  217: dom0 (without domUs) is not meaingfully less secure than the same
  218: things running without Xen.  One can boot Xen or boot regular NetBSD
  219: alternately with little problems, simply refraining from starting the
  220: Xen daemons when not running Xen.
  221: 
  222: Note that NetBSD as dom0 does not support multiple CPUs.  This will
  223: limit the performance of the Xen/dom0 workstation approach.  In theory
  224: the only issue is that the "backend drivers" are not yet MPSAFE:
  225:   http://mail-index.netbsd.org/netbsd-users/2014/08/29/msg015195.html
  226: 
  227: Installation of NetBSD
  228: ----------------------
  229: 
  230: First,
  231: [install NetBSD/amd64](/guide/inst/)
  232: just as you would if you were not using Xen.
  233: However, the partitioning approach is very important.
  234: 
  235: If you want to use RAIDframe for the dom0, there are no special issues
  236: for Xen.  Typically one provides RAID storage for the dom0, and the
  237: domU systems are unaware of RAID.  The 2nd-stage loader bootxx_* skips
  238: over a RAID1 header to find /boot from a filesystem within a RAID
  239: partition; this is no different when booting Xen.
  240: 
  241: There are 4 styles of providing backing storage for the virtual disks
  242: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN,
  243: 
  244: With raw partitions, one has a disklabel (or gpt) partition sized for
  245: each virtual disk to be used by the domU.  (If you are able to predict
  246: how domU usage will evolve, please add an explanation to the HOWTO.
  247: Seriously, needs tend to change over time.)
  248: 
  249: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
  250: for domU disks.  This is almost as efficient as raw disk partitions
  251: and more flexible.  Hence raw disk partitions should typically not
  252: be used.
  253: 
  254: One can use files in the dom0 filesystem, typically created by dd'ing
  255: /dev/zero to create a specific size.  This is somewhat less efficient,
  256: but very convenient, as one can cp the files for backup, or move them
  257: between dom0 hosts.
  258: 
  259: Finally, in theory one can place the files backing the domU disks in a
  260: SAN.  (This is an invitation for someone who has done this to add a
  261: HOWTO page.)
  262: 
  263: Installation of Xen
  264: -------------------
  265: 
  266: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
  267: pkgsrc (or another matching pair).
  268: See [the pkgsrc
  269: documentation](http://www.NetBSD.org/docs/pkgsrc/) for help with pkgsrc.
  270: 
  271: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm.  More
  272: recent versions have HVM support integrated in the main xentools
  273: package.  It is entirely reasonable to run only PV guests.
  274: 
  275: Next you need to install the selected Xen kernel itself, which is
  276: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz".  Copy it to /.
  277: For debugging, one may copy xen-debug.gz; this is conceptually similar
  278: to DIAGNOSTIC and DEBUG in NetBSD.  xen-debug.gz is basically only
  279: useful with a serial console.  Then, place a NetBSD XEN3_DOM0 kernel
  280: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  281: of a NetBSD build.  If using i386, use
  282: releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.  (If using Xen
  283: 3.1 and i386, you may use XEN3_DOM0 with the non-PAE Xen.  But you
  284: should not use Xen 3.1.)  Both xen and the NetBSD kernel may be (and
  285: typically are) left compressed.
  286: 
  287: In a dom0 kernel, kernfs is mandatory for xend to comunicate with the
  288: kernel, so ensure that /kern is in fstab.  TODO: Say this is default,
  289: or file a PR and give a reference.
  290: 
  291: Because you already installed NetBSD, you have a working boot setup
  292: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
  293: beginning of your root filesystem, /boot present, and likely
  294: /boot.cfg.  (If not, fix before continuing!)
  295: 
  296: Add a line to to /boot.cfg to boot Xen.  See boot.cfg(5) for an
  297: example.  The basic line is
  298: 
  299:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
  300: 
  301: which specifies that the dom0 should have 256M, leaving the rest to be
  302: allocated for domUs.  To use a serial console, use
  303: 
  304:         menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=256M console=com1 com1=9600,8n1
  305: 
  306: which will use the first serial port for Xen (which counts starting
  307: from 1), forcing speed/parity, and also for NetBSD (which counts
  308: starting at 0).  In an attempt to add performance, one can also add
  309: 
  310:         dom0_max_vcpus=1 dom0_vcpus_pin
  311: 
  312: to force only one vcpu to be provided (since NetBSD dom0 can't use
  313: more) and to pin that vcpu to a physical cpu.  TODO: benchmark this.
  314: 
  315: As with non-Xen systems, you should have a line to boot /netbsd (a
  316: kernel that works without Xen) and fallback versions of the non-Xen
  317: kernel, Xen, and the dom0 kernel.
  318: 
  319: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  320: than GENERIC without Xen.
  321: 
  322: Using grub (historic)
  323: ---------------------
  324: 
  325: Before NetBSD's native bootloader could support Xen, the use of
  326: grub was recommended.  If necessary, see the
  327: [old grub information](/ports/xen/howto-grub/).
  328: 
  329: The [HowTo on Installing into
  330: RAID-1](http://mail-index.NetBSD.org/port-xen/2006/03/01/0010.html)
  331: explains how to set up booting a dom0 with Xen using grub with
  332: NetBSD's RAIDframe.  (This is obsolete with the use of NetBSD's native
  333: boot.)
  334: 
  335: Configuring Xen
  336: ---------------
  337: 
  338: Xen logs will be in /var/log/xen.
  339: 
  340: Now, you have a system that will boot Xen and the dom0 kernel, but not
  341: do anything else special.  Make sure that you have rebooted into Xen.
  342: There will be no domUs, and none can be started because you still have
  343: to configure the dom0 tools.  The daemons which should be run vary
  344: with Xen version and with whether one is using xm or xl.  Note that
  345: xend is for supporting "xm", and should only be used if you plan on
  346: using "xm".  Do NOT enable xend if you plan on using "xl" as it will
  347: cause problems.  Running xl without xencommons=YES (and starting it)
  348: will result in a hang (so don't do that; follow the HOWTO!).
  349: 
  350: The installation of NetBSD should already have created devices for xen
  351: (xencons, xenevt), but if they are not present, create them:
  352: 
  353:         cd /dev && sh MAKEDEV xen
  354: 
  355: TODO: Give 3.1 advice (or remove it from pkgsrc).
  356: 
  357: For 3.3 (and thus xm), add to rc.conf (but note that you should have
  358: installed 4.1 or 4.2):
  359: 
  360:         xend=YES
  361:         xenbackendd=YES
  362: 
  363: For 4.1 (and thus xm; xl is believed not to work well), add to rc.conf:
  364: 
  365:         xencommons=YES
  366:         xend=YES
  367: 
  368: (If you are using xentools41 from before 2014-12-26, change
  369: rc.d/xendomains to use xm rather than xl.)
  370: 
  371: For 4.2 with xm, add to rc.conf
  372: 
  373:         xencommons=YES
  374:         xend=YES
  375: 
  376: For 4.2 with xl, add to rc.conf:
  377: 
  378:         xencommons=YES
  379:         TODO: explain if there is a xend replacement
  380: 
  381: For 4.5 (and thus with xl), add to rc.conf:
  382: 
  383:         xencommons=YES
  384:         TODO: explain if there is a xend replacement
  385: 
  386: TODO: Recommend for/against xen-watchdog.
  387: 
  388: After you have configured the daemons and either started them (in the
  389: order given) or rebooted, use xm or xl to inspect Xen's boot messages,
  390: available resources, and running domains.  An example with xm follows:
  391: 
  392:         # xm dmesg
  393: 	[xen's boot info]
  394:         # xm info
  395: 	[available memory, etc.]
  396:         # xm list
  397:         Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  398:         Domain-0           0       64    0  r----     58.1
  399: 
  400: With xl, the commands are the same, and the output may be slightly
  401: different.  TODO: add example output for xl before the xm example,
  402: after confirming on 4.2 and resolving the TODO about rc.conf.
  403: 
  404: anita (for testing NetBSD)
  405: --------------------------
  406: 
  407: With the setup so far (assuming 4.2/xl), one should be able to run
  408: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  409: root, because anita must create a domU):
  410: 
  411:         anita --vmm=xl test file:///usr/obj/i386/
  412: 
  413: Alternatively, one can use --vmm=xm to use xm-based domU creation
  414: instead (and must, on Xen <= 4.1).   TODO: confirm that anita xl really works.
  415:     
  416: Xen-specific NetBSD issues
  417: --------------------------
  418: 
  419: There are (at least) two additional things different about NetBSD as a
  420: dom0 kernel compared to hardware.
  421: 
  422: One is that modules are not usable in DOM0 kernels, so one must
  423: compile in what's needed.  It's not really that modules cannot work,
  424: but that modules must be built for XEN3_DOM0 because some of the
  425: defines change and the normal module builds don't do this.  Basically,
  426: enabling Xen changes the kernel ABI, and the module build system
  427: doesn't cope with this.
  428: 
  429: The other difference is that XEN3_DOM0 does not have exactly the same
  430: options as GENERIC.  While it is debatable whether or not this is a
  431: bug, users should be aware of this and can simply add missing config
  432: items if desired.
  433: 
  434: Updating NetBSD in a dom0
  435: -------------------------
  436: 
  437: This is just like updating NetBSD on bare hardware, assuming the new
  438: version supports the version of Xen you are running.  Generally, one
  439: replaces the kernel and reboots, and then overlays userland binaries
  440: and adjusts /etc.
  441: 
  442: Note that one must update both the non-Xen kernel typically used for
  443: rescue purposes and the DOM0 kernel used with Xen.
  444: 
  445: Converting from grub to /boot
  446: -----------------------------
  447: 
  448: These instructions were [TODO: will be] used to convert a system from
  449: grub to /boot.  The system was originally installed in February of
  450: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  451: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  452: 4.1 and grub, much like the message linked earlier in the grub
  453: section.
  454: 
  455:         # Install mbr bootblocks on both disks. 
  456:         fdisk -i /dev/rwd0d
  457:         fdisk -i /dev/rwd1d
  458:         # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  459:         installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  460:         installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  461:         # Install secondary boot loader
  462:         cp -p /usr/mdec/boot /
  463:         # Create boog.cfg following earlier guidance:
  464:         menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
  465:         menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=256M
  466:         menu=GENERIC:boot
  467:         menu=GENERIC single-user:boot -s
  468:         menu=GENERIC.ok:boot netbsd.ok
  469:         menu=GENERIC.ok single-user:boot netbsd.ok -s
  470:         menu=Drop to boot prompt:prompt
  471:         default=1
  472:         timeout=30
  473: 
  474: TODO: actually do this and fix it if necessary.
  475: 
  476: Updating Xen versions
  477: ---------------------
  478: 
  479: Updating Xen is conceptually not difficult, but can run into all the
  480: issues found when installing Xen.  Assuming migration from 4.1 to 4.2,
  481: remove the xenkernel41 and xentools41 packages and install the
  482: xenkernel42 and xentools42 packages.  Copy the 4.2 xen.gz to /.
  483: 
  484: Ensure that the contents of /etc/rc.d/xen* are correct.  Enable the
  485: correct set of daemons.  Ensure that the domU config files are valid
  486: for the new version.
  487: 
  488: 
  489: Running Xen under qemu
  490: ----------------------
  491: 
  492: The astute reader will note that this section is somewhat twisted.
  493: However, it can be useful to run Xen under qemu either because the
  494: version of NetBSD as a dom0 does not run on the hardware in use, or to
  495: generate automated test cases involving Xen.
  496: 
  497: In 2015-01, the following combination was reported to mostly work:
  498: 
  499:         host OS: NetBSD/amd64 6.1.4
  500:         qemu: 2.2.0 from pkgsrc
  501:         Xen kernel: xenkernel42-4.2.5nb1 from pkgsrc
  502:         dom0 kernel: NetBSD/amd64 6.1.5
  503:         Xen tools: xentools42-4.2.5 from pkgsrc
  504: 
  505: See http://gnats.netbsd.org/47720 for a problem with dom0 shutdown.
  506: 
  507: Unprivileged domains (domU)
  508: ===========================
  509: 
  510: This section describes general concepts about domUs.  It does not
  511: address specific domU operating systems or how to install them.  The
  512: config files for domUs are typically in /usr/pkg/etc/xen, and are
  513: typically named so that the file name, domU name and the domU's host
  514: name match.
  515: 
  516: The domU is provided with cpu and memory by Xen, configured by the
  517: dom0.  The domU is provided with disk and network by the dom0,
  518: mediated by Xen, and configured in the dom0.
  519: 
  520: Entropy in domUs can be an issue; physical disks and network are on
  521: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  522: 
  523: Config files
  524: ------------
  525: 
  526: There is no good order to present config files and the concepts
  527: surrounding what is being configured.  We first show an example config
  528: file, and then in the various sections give details.
  529: 
  530: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
  531: for a large number of well-commented examples, mostly for running
  532: GNU/Linux.
  533: 
  534: The following is an example minimal domain configuration file
  535: "/usr/pkg/etc/xen/foo".  It is (with only a name change) an actual
  536: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
  537: i386 domU).  The domU serves as a network file server.
  538: 
  539:         # -*- mode: python; -*-
  540: 
  541:         kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  542:         memory = 1024
  543:         vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  544:         disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  545:                  'file:/n0/xen/foo-wd1,0x1,w' ]
  546: 
  547: The domain will have the same name as the file.  The kernel has the
  548: host/domU name in it, so that on the dom0 one can update the various
  549: domUs independently.  The vif line causes an interface to be provided,
  550: with a specific mac address (do not reuse MAC addresses!), in bridge
  551: mode.  Two disks are provided, and they are both writable; the bits
  552: are stored in files and Xen attaches them to a vnd(4) device in the
  553: dom0 on domain creation.  The system treates xbd0 as the boot device
  554: without needing explicit configuration.
  555: 
  556: By default xm looks for domain config files in /usr/pkg/etc/xen.  Note
  557: that "xm create" takes the name of a config file, while other commands
  558: take the name of a domain.  To create the domain, connect to the
  559: console, create the domain while attaching the console, shutdown the
  560: domain, and see if it has finished stopping, do (or xl with Xen >=
  561: 4.2):
  562: 
  563:         xm create foo
  564:         xm console foo
  565:         xm create -c foo
  566:         xm shutdown foo
  567: 	xm list
  568: 
  569: Typing ^] will exit the console session.  Shutting down a domain is
  570: equivalent to pushing the power button; a NetBSD domU will receive a
  571: power-press event and do a clean shutdown.  Shutting down the dom0
  572: will trigger controlled shutdowns of all configured domUs.
  573: 
  574: domU kernels
  575: ------------
  576: 
  577: On a physical computer, the BIOS reads sector 0, and a chain of boot
  578: loaders finds and loads a kernel.  Normally this comes from the root
  579: filesystem.  With Xen domUs, the process is totally different.  The
  580: normal path is for the domU kernel to be a file in the dom0's
  581: filesystem.  At the request of the dom0, Xen loads that kernel into a
  582: new domU instance and starts execution.  While domU kernels can be
  583: anyplace, reasonable places to store domU kernels on the dom0 are in /
  584: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
  585: config files), or in /u0/xen (where the vdisks are).
  586: 
  587: Note that loading the domU kernel from the dom0 implies that boot
  588: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
  589: See the VPS section near the end for discussion of alternate ways to
  590: obtain domU kernels.
  591: 
  592: CPU and memory
  593: --------------
  594: 
  595: A domain is provided with some number of vcpus, less than the number
  596: of cpus seen by the hypervisor.  (For a dom0, this is controlled by
  597: the boot argument "dom0_max_vcpus=1".)  For a domU, it is controlled
  598: from the config file by the "vcpus = N" directive.
  599: 
  600: A domain is provided with memory; this is controlled in the config
  601: file by "memory = N" (in megabytes).  In the straightforward case, the
  602: sum of the the memory allocated to the dom0 and all domUs must be less
  603: than the available memory.
  604: 
  605: Xen also provides a "balloon" driver, which can be used to let domains
  606: use more memory temporarily.  TODO: Explain better, and explain how
  607: well it works with NetBSD.
  608: 
  609: Virtual disks
  610: -------------
  611: 
  612: With the file/vnd style, typically one creates a directory,
  613: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
  614: domUs.  Then, for each domU disk, one writes zeros to a file that then
  615: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
  616: for the first virtual disk for the domU called foo.  Writing zeros to
  617: the file serves two purposes.  One is that preallocating the contents
  618: improves performance.  The other is that vnd on sparse files has
  619: failed to work.  TODO: give working/notworking NetBSD versions for
  620: sparse vnd.  Note that the use of file/vnd for Xen is not really
  621: different than creating a file-backed virtual disk for some other
  622: purpose, except that xentools handles the vnconfig commands.  To
  623: create an empty 4G virtual disk, simply do
  624: 
  625:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
  626: 
  627: With the lvm style, one creates logical devices.  They are then used
  628: similarly to vnds.  TODO: Add an example with lvm.
  629: 
  630: In domU config files, the disks are defined as a sequence of 3-tuples.
  631: The first element is "method:/path/to/disk".  Common methods are
  632: "file:" for file-backed vnd. and "phy:" for something that is already
  633: a (TODO: character or block) device.
  634: 
  635: The second element is an artifact of how virtual disks are passed to
  636: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  637: are given a device name to associate with the disk, and values like
  638: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  639: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
  640: second argument.  The name given is converted to a major/minor by
  641: calling stat(2) on the name in /dev and this is passed to the domU.
  642: In the general case, the dom0 and domU can be different operating
  643: systems, and it is an unwarranted assumption that they have consistent
  644: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  645: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  646: for the second works fine and avoids this issue.  For a GNU/Linux
  647: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  648: /dev/hda1.
  649: 
  650: The third element is "w" for writable disks, and "r" for read-only
  651: disks.
  652: 
  653: Virtual Networking
  654: ------------------
  655: 
  656: Xen provides virtual ethernets, each of which connects the dom0 and a
  657: domU.  For each virtual network, there is an interface "xvifN.M" in
  658: the dom0, and in domU index N, a matching interface xennetM (NetBSD
  659: name).  The interfaces behave as if there is an Ethernet with two
  660: adaptors connected.  From this primitive, one can construct various
  661: configurations.  We focus on two common and useful cases for which
  662: there are existing scripts: bridging and NAT.
  663: 
  664: With bridging (in the example above), the domU perceives itself to be
  665: on the same network as the dom0.  For server virtualization, this is
  666: usually best.  Bridging is accomplished by creating a bridge(4) device
  667: and adding the dom0's physical interface and the various xvifN.0
  668: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  669: config file.  The bridge must be set up already in the dom0; an
  670: example /etc/ifconfig.bridge0 is:
  671: 
  672:         create
  673:         up
  674:         !brconfig bridge0 add wm0
  675: 
  676: With NAT, the domU perceives itself to be behind a NAT running on the
  677: dom0.  This is often appropriate when running Xen on a workstation.
  678: TODO: NAT appears to be configured by "vif = [ '' ]".
  679: 
  680: The MAC address specified is the one used for the interface in the new
  681: domain.  The interface in dom0 will use this address XOR'd with
  682: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  683: 
  684: Sizing domains
  685: --------------
  686: 
  687: Modern x86 hardware has vast amounts of resources.  However, many
  688: virtual servers can function just fine on far less.  A system with
  689: 256M of RAM and a 4G disk can be a reasonable choice.  Note that it is
  690: far easier to adjust virtual resources than physical ones.  For
  691: memory, it's just a config file edit and a reboot.  For disk, one can
  692: create a new file and vnconfig it (or lvm), and then dump/restore,
  693: just like updating physical disks, but without having to be there and
  694: without those pesky connectors.
  695: 
  696: Starting domains automatically
  697: ------------------------------
  698: 
  699: To start domains foo at bar at boot and shut them down cleanly on dom0
  700: shutdown, in rc.conf add:
  701: 
  702:         xendomains="foo bar"
  703: 
  704: Note that earlier versions of the xentools41 xendomains rc.d scripth
  705: usd xl, when one should use xm with 4.1.
  706: 
  707: Creating specific unprivileged domains (domU)
  708: =============================================
  709: 
  710: Creating domUs is almost entirely independent of operating system.  We
  711: have already presented the basics of config files.  Note that you must
  712: have already completed the dom0 setup so that "xl list" (or "xm list")
  713: works.
  714: 
  715: Creating an unprivileged NetBSD domain (domU)
  716: ---------------------------------------------
  717: 
  718: See the earlier config file, and adjust memory.  Decide on how much
  719: storage you will provide, and prepare it (file or lvm).
  720: 
  721: While the kernel will be obtained from the dom0 filesystem, the same
  722: file should be present in the domU as /netbsd so that tools like
  723: savecore(8) can work.   (This is helpful but not necessary.)
  724: 
  725: The kernel must be specifically for Xen and for use as a domU.  The
  726: i386 and amd64 provide the following kernels:
  727: 
  728:         i386 XEN3_DOMU
  729:         i386 XEN3PAE_DOMU
  730: 	amd64 XEN3_DOMU
  731: 
  732: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
  733: use the PAE version of the i386 kernel.
  734: 
  735: This will boot NetBSD, but this is not that useful if the disk is
  736: empty.  One approach is to unpack sets onto the disk outside of xen
  737: (by mounting it, just as you would prepare a physical disk for a
  738: system you can't run the installer on).
  739: 
  740: A second approach is to run an INSTALL kernel, which has a miniroot
  741: and can load sets from the network.  To do this, copy the INSTALL
  742: kernel to / and change the kernel line in the config file to:
  743: 
  744:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  745: 
  746: Then, start the domain as "xl create -c configname".
  747: 
  748: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  749: line should be used in the config file.
  750: 
  751:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  752: 
  753: After booting the domain, the option to install via CDROM may be
  754: selected.  The CDROM device should be changed to `xbd1d`.
  755: 
  756: Once done installing, "halt -p" the new domain (don't reboot or halt,
  757: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  758: config file), switch the config file back to the XEN3_DOMU kernel,
  759: and start the new domain again. Now it should be able to use "root on
  760: xbd0a" and you should have a, functional NetBSD domU.
  761: 
  762: TODO: check if this is still accurate.
  763: When the new domain is booting you'll see some warnings about *wscons*
  764: and the pseudo-terminals. These can be fixed by editing the files
  765: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  766: `/etc/ttys`, except *console*, like this:
  767: 
  768:     console "/usr/libexec/getty Pc"         vt100   on secure
  769:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  770:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  771:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  772:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  773: 
  774: Finally, all screens must be commented out from `/etc/wscons.conf`.
  775: 
  776: It is also desirable to add
  777: 
  778:         powerd=YES
  779: 
  780: in rc.conf. This way, the domain will be properly shut down if
  781: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  782: 
  783: Your domain should be now ready to work, enjoy.
  784: 
  785: Creating an unprivileged Linux domain (domU)
  786: --------------------------------------------
  787: 
  788: Creating unprivileged Linux domains isn't much different from
  789: unprivileged NetBSD domains, but there are some details to know.
  790: 
  791: First, the second parameter passed to the disk declaration (the '0x1' in
  792: the example below)
  793: 
  794:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  795: 
  796: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  797: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  798: So, hda1 which has major 3 and minor 1 on a Linux system will have
  799: device number 0x301.  Alternatively, devices names can be used (hda,
  800: hdb, ...)  as xentools has a table to map these names to devices
  801: numbers.  To export a partition to a Linux guest we can use:
  802: 
  803:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  804:         root = "/dev/hda1 ro"
  805: 
  806: and it will appear as /dev/hda on the Linux system, and be used as root
  807: partition.
  808: 
  809: To install the Linux system on the partition to be exported to the
  810: guest domain, the following method can be used: install
  811: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  812: that will be the root partition of your Linux domain, and mount it.
  813: Then copy the files from a working Linux system, make adjustments in
  814: `/etc` (fstab, network config).  It should also be possible to extract
  815: binary packages such as .rpm or .deb directly to the mounted partition
  816: using the appropriate tool, possibly running under NetBSD's Linux
  817: emulation.  Once the filesystem has been populated, umount it.  If
  818: desirable, the filesystem can be converted to ext3 using tune2fs -j.
  819: It should now be possible to boot the Linux guest domain, using one of
  820: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  821: 
  822: To get the linux console right, you need to add:
  823: 
  824:     extra = "xencons=tty1"
  825: 
  826: to your configuration since not all linux distributions auto-attach a
  827: tty to the xen console.
  828: 
  829: Creating an unprivileged Solaris domain (domU)
  830: ----------------------------------------------
  831: 
  832: See possibly outdated
  833: [Solaris domU instructions](/ports/xen/howto-solaris/).
  834: 
  835: 
  836: PCI passthrough: Using PCI devices in guest domains
  837: ---------------------------------------------------
  838: 
  839: The dom0 can give other domains access to selected PCI
  840: devices. This can allow, for example, a non-privileged domain to have
  841: access to a physical network interface or disk controller.  However,
  842: keep in mind that giving a domain access to a PCI device most likely
  843: will give the domain read/write access to the whole physical memory,
  844: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  845: device.  Also, it's not possible to export ISA devices to non-dom0
  846: domains, which means that the primary VGA adapter can't be exported.
  847: A guest domain trying to access the VGA registers will panic.
  848: 
  849: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  850: not been ported to later versions at this time.
  851: 
  852: For a PCI device to be exported to a domU, is has to be attached to
  853: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  854: pciback.hide boot parameter will attach to "pciback" instead of the
  855: usual driver.  The list of devices is specified as "(bus:dev.func)",
  856: where bus and dev are 2-digit hexadecimal numbers, and func a
  857: single-digit number:
  858: 
  859:         pciback.hide=(00:0a.0)(00:06.0)
  860: 
  861: pciback devices should show up in the dom0's boot messages, and the
  862: devices should be listed in the `/kern/xen/pci` directory.
  863: 
  864: PCI devices to be exported to a domU are listed in the "pci" array of
  865: the domU's config file, with the format "0000:bus:dev.func".
  866: 
  867:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  868: 
  869: In the domU an "xpci" device will show up, to which one or more pci
  870: busses will attach.  Then the PCI drivers will attach to PCI busses as
  871: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  872: or any PCI drivers built in by default; you have to build your own
  873: kernel to use PCI devices in a domU.  Here's a kernel config example;
  874: note that only the "xpci" lines are unusual.
  875: 
  876:         include         "arch/i386/conf/XEN3_DOMU"
  877: 
  878:         # Add support for PCI busses to the XEN3_DOMU kernel
  879:         xpci* at xenbus ?
  880:         pci* at xpci ?
  881: 
  882:         # PCI USB controllers
  883:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  884: 
  885:         # USB bus support
  886:         usb*    at uhci?
  887: 
  888:         # USB Hubs
  889:         uhub*   at usb?
  890:         uhub*   at uhub? port ? configuration ? interface ?
  891: 
  892:         # USB Mass Storage
  893:         umass*  at uhub? port ? configuration ? interface ?
  894:         wd*     at umass?
  895:         # SCSI controllers
  896:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  897: 
  898:         # SCSI bus support (for both ahc and umass)
  899:         scsibus* at scsi?
  900: 
  901:         # SCSI devices
  902:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  903:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  904: 
  905: 
  906: NetBSD as a domU in a VPS
  907: =========================
  908: 
  909: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  910: hardware.  This section explains how to deal with Xen in a domU as a
  911: virtual private server where you do not control or have access to the
  912: dom0.  This is not intended to be an exhaustive list of VPS providers;
  913: only a few are mentioned that specifically support NetBSD.
  914: 
  915: VPS operators provide varying degrees of access and mechanisms for
  916: configuration.  The big issue is usually how one controls which kernel
  917: is booted, because the kernel is nominally in the dom0 filesystem (to
  918: which VPS users do not normally have acesss).  A second issue is how
  919: to install NetBSD.
  920: A VPS user may want to compile a kernel for security updates, to run
  921: npf, run IPsec, or any other reason why someone would want to change
  922: their kernel.
  923: 
  924: One approach is to have an adminstrative interface to upload a kernel,
  925: or to select from a prepopulated list.  Other approaches are pygrub
  926: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  927: kernel from the domU filesystem.  This is closer to a regular physical
  928: computer, where someone who controls a machine can replace the kernel.
  929: 
  930: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  931: multiple vcpus, and it is typical for VPS providers to enable multiple
  932: CPUs for NetBSD domUs.
  933: 
  934: pygrub
  935: -------
  936: 
  937: pygrub runs in the dom0 and looks into the domU filesystem.  This
  938: implies that the domU must have a kernel in a filesystem in a format
  939: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
  940: interest.
  941: 
  942: pvgrub
  943: ------
  944: 
  945: pvgrub is a version of grub that uses PV operations instead of BIOS
  946: calls.  It is booted from the dom0 as the domU kernel, and then reads
  947: /grub/menu.lst and loads a kernel from the domU filesystem.
  948: 
  949: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  950: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  951: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  952: page](http://www.panix.com/v-colo/grub.html), which describes only
  953: Linux but should be updated to cover NetBSD :-).
  954: 
  955: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  956: their own kernel.  See then [prgmr.com NetBSD
  957: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  958: (which is in need of updating).
  959: 
  960: It appears that [grub's FFS
  961: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  962: does not support all aspects of modern FFS, but there are also reports
  963: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  964: partition for the kernel with the intent that grub can understand it,
  965: which leads to /netbsd not being the actual kernel.  One must remember
  966: to update the special boot partiion.
  967: 
  968: Amazon
  969: ------
  970: 
  971: TODO: add link to NetBSD amazon howto.
  972: 
  973: Using npf
  974: ---------
  975: 
  976: In standard kernels, npf is a module, and thus cannot be loaded in a
  977: DOMU kernel.
  978: 
  979: TODO: explain how to compile npf into a custom kernel, answering (but
  980: note that the problem was caused by not booting the right kernel):
  981: http://mail-index.netbsd.org/netbsd-users/2014/12/26/msg015576.html
  982: 
  983: TODO items for improving NetBSD/xen
  984: ===================================
  985: 
  986: * Package Xen 4.4.
  987: * Get PCI passthrough working on Xen 4.2 (or 4.4).
  988: * Get pvgrub into pkgsrc, either via xentools or separately.
  989: * grub
  990:   * Check/add support to pkgsrc grub2 for UFS2 and arbitrary
  991:     fragsize/blocksize (UFS2 support may be present; the point is to
  992:     make it so that with any UFS1/UFS2 filesystem setup that works
  993:     with NetBSD grub will also work).
  994:     See [pkg/40258](http://gnats.netbsd.org/40258).
  995:   * Push patches upstream.
  996:   * Get UFS2 patches into pvgrub.
  997: * Add support for PV ops to a version of /boot, and make it usable as
  998:   a kernel in Xen, similar to pvgrub.

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