File:  [NetBSD Developer Wiki] / wikisrc / ports / xen / howto.mdwn
Revision 1.153: download - view: text, annotated - select for diffs
Thu Jul 26 12:21:24 2018 UTC (16 months, 2 weeks ago) by maxv
Branches: MAIN
CVS tags: HEAD
less noise, more swag

    1: [[!meta title="Xen HowTo"]]
    2: 
    3: Introduction
    4: ============
    5: 
    6: [![[Xen
    7: screenshot]](https://www.netbsd.org/gallery/in-Action/hubertf-xens.png)](https://www.netbsd.org/gallery/in-Action/hubertf-xen.png)
    8: 
    9: Xen is a hypervisor for x86 hardware, which supports running multiple guest
   10: operating systems on a single physical machine.  Xen is a Type 1 or
   11: bare-metal hypervisor; one uses the Xen kernel to control the CPU,
   12: memory and console, a dom0 operating system which mediates access to
   13: other hardware (e.g., disks, network, USB), and one or more domU
   14: operating systems which operate in an unprivileged virtualized
   15: environment.  IO requests from the domU systems are forwarded by the
   16: Xen hypervisor to the dom0 to be fulfilled.
   17: 
   18: Xen supports different styles of guest:
   19: 
   20: [[!table data="""
   21: Style of guest	|Supported by NetBSD
   22: PV		|Yes (dom0, domU)
   23: HVM		|Yes (domU)
   24: PVHVM		|No
   25: PVH		|No
   26: """]]
   27: 
   28: In Para-Virtualized (PV) mode, the guest OS does not attempt to access
   29: hardware directly, but instead makes hypercalls to the hypervisor; PV
   30: guests must be specifically coded for Xen. In HVM mode, no guest
   31: modification is required; however, hardware support is required, such
   32: as VT-x on Intel CPUs and SVM on AMD CPUs.
   33: 
   34: There are further features for IOMMU virtualization, Intel's VT-d and
   35: AMD's AMD-Vi.  TODO: Explain whether Xen on NetBSD makes use of these
   36: features.  TODO: Review by someone who really understands this.
   37: 
   38: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
   39: The dom0 can start one or more domUs.  (Booting is explained in detail
   40: in the dom0 section.)
   41: 
   42: This HOWTO presumes a basic familiarity with the Xen system
   43: architecture, with installing NetBSD on i386/amd64 hardware, and with
   44: installing software from pkgsrc.  See also the [Xen
   45: website](http://www.xenproject.org/).
   46: 
   47: This HOWTO attempts to address both the case of running a NetBSD dom0
   48: on hardware and running domUs under it (NetBSD and other), and also
   49: running NetBSD as a domU in a VPS.
   50: 
   51: Versions of Xen and NetBSD
   52: ==========================
   53: 
   54: Most of the installation concepts and instructions are independent
   55: of Xen version and NetBSD version.  This section gives advice on
   56: which version to choose.  Versions not in pkgsrc and older unsupported
   57: versions of NetBSD are intentionally ignored.
   58: 
   59: The term "amd64" is used to refer to both the NetBSD port and to the
   60: hardware architecture on which it runs.  Such hardware is generally
   61: made by both Intel and AMD, and common on PC computers.
   62: 
   63: Xen versions
   64: ------------
   65: 
   66: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
   67: xenkernel and xentools.  We will refer only to the kernel versions,
   68: but note that both packages must be installed together and must have
   69: matching versions.
   70: 
   71: Versions available in pkgsrc:
   72: 
   73: [[!table data="""
   74: Xen Version	|Package Name	|Xen CPU Support	|EOL'ed By Upstream
   75: 4.2		|xenkernel42	|32bit, 64bit		|Yes
   76: 4.5		|xenkernel45	|64bit			|Yes
   77: 4.6		|xenkernel46	|64bit			|Partially
   78: 4.8		|xenkernel48	|64bit			|No
   79: 4.11		|xenkernel411	|64bit			|No
   80: """]]
   81: 
   82: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
   83: 
   84: Note: Xen 4.2 was the last version to support 32bit CPUs.
   85: 
   86: Xen command program
   87: -------------------
   88: 
   89: Early Xen used a program called xm to manipulate the system from the
   90: dom0.  Starting in 4.1, a replacement program with similar behavior
   91: called xl is provided, but it does not work well in 4.1.  In 4.2, both
   92: xm and xl work fine.  4.4 is the last version that has xm.
   93: 
   94: You must make a global choice to use xm or xl, because it affects not
   95: only which command you use, but the command used by rc.d scripts
   96: (specifically xendomains) and which daemons should be run.  The
   97: xentools packages provide xl for 4.2 and up.
   98: 
   99: In 4.2, you can choose to use xm by simply changing the ctl_command
  100: variable and setting xend=YES in rc.conf.
  101: 
  102: With xl, virtual devices are configured in parallel, which can cause
  103: problems if they are written assuming serial operation (e.g., updating
  104: firewall rules without explicit locking).  There is now locking for
  105: the provided scripts, which works for normal casses (e.g, file-backed
  106: xbd, where a vnd must be allocated).  But, as of 201612, it has not
  107: been adequately tested for a complex custom setup with a large number
  108: of interfaces.
  109: 
  110: NetBSD versions
  111: ---------------
  112: 
  113: The netbsd-7, netbsd-8, and -current branches are all reasonable
  114: choices, with more or less the same considerations for non-Xen use.
  115: NetBSD 8 is recommended as the stable version of the most recent
  116: release for production use.
  117: 
  118: For developing Xen, netbsd-current may be appropriate.
  119: 
  120: As of NetBSD 6, a NetBSD domU will support multiple vcpus.  There is
  121: no SMP support for NetBSD as dom0.  (The dom0 itself doesn't really
  122: need SMP for dom0 functions; the lack of support is really a problem
  123: when using a dom0 as a normal computer.)
  124: 
  125: Note: NetBSD support is called XEN3. However, it does support Xen 4,
  126: because the hypercall interface has remained identical.
  127: 
  128: Architecture
  129: ------------
  130: 
  131: Xen itself can run on i386 (Xen < 4.2) or amd64 hardware (all Xen
  132: versions).  Practically, almost any computer where one would want to
  133: run Xen today supports amd64.
  134: 
  135: Xen, the dom0 system, and each domU system can be either i386 or
  136: amd64.  When building a xenkernel package, one obtains an i386 Xen
  137: kernel on an i386 host, and an amd64 Xen kernel on an amd64 host.  If
  138: the Xen kernel is i386, then the dom0 kernel and all domU kernels must
  139: be i386.  With an amd64 Xen kernel, an amd64 dom0 kernel is known to
  140: work, and an i386 dom0 kernel should in theory work.  An amd64
  141: Xen/dom0 is known to support both i386 and amd64 domUs.
  142: 
  143: i386 dom0 and domU kernels must be PAE. PAE kernels are included in
  144: the NetBSD default build.
  145: 
  146: Because of the above, the standard approach is to use an amd64 Xen
  147: kernel and NetBSD/amd64 for the dom0.  For domUs, NetBSD/i386 (with
  148: the PAE kernel) and NetBSD/amd64 are in widespread use, and there is
  149: little to no Xen-specific reason to prefer one over the other.
  150: 
  151: Note that to use an i386 dom0 with Xen 4.5 or higher, one must build
  152: (or obtain from pre-built packages) an amd64 Xen kernel and install
  153: that on the system.  (One must also use a PAE i386 kernel, but this is
  154: also required with an i386 Xen kernel.).  Almost no one in the
  155: NetBSD/Xen community does this, and the standard, well-tested,
  156: approach is to use an amd64 dom0.
  157: 
  158: A [posting on
  159: xen-devel](https://lists.xen.org/archives/html/xen-devel/2012-07/msg00085.html)
  160: explained that PV system call overhead was higher on amd64, and thus
  161: there is some notion that i386 guests are faster.  It goes on to
  162: caution that the total situation is complex and not entirely
  163: understood. On top of that caution, the post is about Linux, not
  164: NetBSD.  TODO: Include link to benchmarks, if someone posts them.
  165: 
  166: NetBSD as a dom0
  167: ================
  168: 
  169: NetBSD can be used as a dom0 and works very well.  The following
  170: sections address installation, updating NetBSD, and updating Xen.
  171: Note that it doesn't make sense to talk about installing a dom0 OS
  172: without also installing Xen itself.  We first address installing
  173: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
  174: NetBSD install to a dom0 install by just changing the kernel and boot
  175: configuration.
  176: 
  177: For experimenting with Xen, a machine with as little as 1G of RAM and
  178: 100G of disk can work.  For running many domUs in productions, far
  179: more will be needed; e.g. 4-8G and 1T of disk is reasonable for a
  180: half-dozen domUs of 512M and 32G each.  Basically, the RAM and disk
  181: have to be bigger than the sum of the RAM/disk needs of the dom0 and
  182: all the domUs.
  183: 
  184: In 2018-05, trouble booting a dom0 was reported with 256M of RAM: with
  185: 512M it worked reliably.  This does not make sense, but if you see
  186: "not ELF" after Xen boots, try increasing dom0 RAM.
  187: 
  188: Styles of dom0 operation
  189: ------------------------
  190: 
  191: There are two basic ways to use Xen.  The traditional method is for
  192: the dom0 to do absolutely nothing other than providing support to some
  193: number of domUs.  Such a system was probably installed for the sole
  194: purpose of hosting domUs, and sits in a server room on a UPS.
  195: 
  196: The other way is to put Xen under a normal-usage computer, so that the
  197: dom0 is what the computer would have been without Xen, perhaps a
  198: desktop or laptop.  Then, one can run domUs at will.  Purists will
  199: deride this as less secure than the previous approach, and for a
  200: computer whose purpose is to run domUs, they are right.  But Xen and a
  201: dom0 (without domUs) is not meaningfully less secure than the same
  202: things running without Xen.  One can boot Xen or boot regular NetBSD
  203: alternately with little problems, simply refraining from starting the
  204: Xen daemons when not running Xen.
  205: 
  206: Note that NetBSD as dom0 does not support multiple CPUs.  This will
  207: limit the performance of the Xen/dom0 workstation approach.  In theory
  208: the only issue is that the "backend drivers" are not yet MPSAFE:
  209:   https://mail-index.netbsd.org/netbsd-users/2014/08/29/msg015195.html
  210: 
  211: Installation of NetBSD
  212: ----------------------
  213: 
  214: First,
  215: [install NetBSD/amd64](/guide/inst/)
  216: just as you would if you were not using Xen.
  217: However, the partitioning approach is very important.
  218: 
  219: If you want to use RAIDframe for the dom0, there are no special issues
  220: for Xen.  Typically one provides RAID storage for the dom0, and the
  221: domU systems are unaware of RAID.  The 2nd-stage loader bootxx_* skips
  222: over a RAID1 header to find /boot from a file system within a RAID
  223: partition; this is no different when booting Xen.
  224: 
  225: There are 4 styles of providing backing storage for the virtual disks
  226: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN.
  227: 
  228: With raw partitions, one has a disklabel (or gpt) partition sized for
  229: each virtual disk to be used by the domU.  (If you are able to predict
  230: how domU usage will evolve, please add an explanation to the HOWTO.
  231: Seriously, needs tend to change over time.)
  232: 
  233: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
  234: for domU disks.  This is almost as efficient as raw disk partitions
  235: and more flexible.  Hence raw disk partitions should typically not
  236: be used.
  237: 
  238: One can use files in the dom0 file system, typically created by dd'ing
  239: /dev/zero to create a specific size.  This is somewhat less efficient,
  240: but very convenient, as one can cp the files for backup, or move them
  241: between dom0 hosts.
  242: 
  243: Finally, in theory one can place the files backing the domU disks in a
  244: SAN.  (This is an invitation for someone who has done this to add a
  245: HOWTO page.)
  246: 
  247: Installation of Xen
  248: -------------------
  249: 
  250: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
  251: pkgsrc (or another matching pair).  See [the pkgsrc
  252: documentation](https://www.NetBSD.org/docs/pkgsrc/) for help with
  253: pkgsrc.  Ensure that your packages are recent; the HOWTO does not
  254: contemplate old builds.
  255: 
  256: Next you need to install the selected Xen kernel itself, which is
  257: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz".  Copy it to /.
  258: For debugging, one may copy xen-debug.gz; this is conceptually similar
  259: to DIAGNOSTIC and DEBUG in NetBSD.  xen-debug.gz is basically only
  260: useful with a serial console.  Then, place a NetBSD XEN3_DOM0 kernel
  261: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
  262: of a NetBSD build.  If using i386, use
  263: releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.  Both xen and
  264: the NetBSD kernel may be (and typically are) left compressed.
  265: 
  266: In a dom0, kernfs is mandatory for xend to communicate with the
  267: kernel, so ensure that /kern is in fstab.  (A standard NetBSD install
  268: should already mount /kern.)
  269: 
  270: Because you already installed NetBSD, you have a working boot setup
  271: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
  272: beginning of your root file system, have /boot, and likely also
  273: /boot.cfg.  (If not, fix before continuing!)
  274: 
  275: Add a line to /boot.cfg to boot Xen.  See boot.cfg(5) for an
  276: example.  The basic line is:
  277: 
  278: [[!template id=programlisting text="""
  279: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  280: """]]
  281: 
  282: which specifies that the dom0 should have 512M, leaving the rest to be
  283: allocated for domUs.  To use a serial console, use
  284: 
  285: [[!template id=programlisting text="""
  286: menu=Xen:load /netbsd-XEN3_DOM0.gz;multiboot /xen.gz dom0_mem=512M console=com1 com1=9600,8n1
  287: """]]
  288: 
  289: which will use the first serial port for Xen (which counts starting
  290: from 1, unlike NetBSD which counts starting from 0), forcing
  291: speed/parity.  Because the NetBSD command line lacks a
  292: "console=pc" argument, it will use the default "xencons" console device,
  293: which directs the console I/O through Xen to the same console device Xen
  294: itself uses (in this case, the serial port).
  295: 
  296: In an attempt to add performance, one can also add:
  297: 
  298: [[!template id=programlisting text="""
  299: dom0_max_vcpus=1 dom0_vcpus_pin
  300: """]]
  301: 
  302: to force only one vcpu to be provided (since NetBSD dom0 can't use
  303: more) and to pin that vcpu to a physical CPU.  TODO: benchmark this.
  304: 
  305: Xen has [many boot
  306: options](http://xenbits.xenproject.org/docs/4.5-testing/misc/xen-command-line.html),
  307: and other than dom0 memory and max_vcpus, they are generally not
  308: necessary.
  309: 
  310: As with non-Xen systems, you should have a line to boot /netbsd (a
  311: kernel that works without Xen).  Consider a line to boot /netbsd.ok (a
  312: fallback version of the non-Xen kernel, updated manually when you are
  313: sure /netbsd is ok).  Consider also a line to boot fallback versions
  314: of Xen and the dom0 kernel, but note that non-Xen NetBSD can be used
  315: to resolve Xen booting issues.
  316: 
  317: Now, reboot so that you are running a DOM0 kernel under Xen, rather
  318: than GENERIC without Xen.
  319: 
  320: Using grub (historic)
  321: ---------------------
  322: 
  323: Before NetBSD's native bootloader could support Xen, the use of
  324: grub was recommended.  If necessary, see the
  325: [old grub information](/ports/xen/howto-grub).
  326: 
  327: The [HowTo on Installing into
  328: RAID-1](https://mail-index.NetBSD.org/port-xen/2006/03/01/0010.html)
  329: explains how to set up booting a dom0 with Xen using grub with
  330: NetBSD's RAIDframe.  (This is obsolete with the use of NetBSD's native
  331: boot.  Now, just create a system with RAID-1, and alter /boot.cfg as
  332: described above.)
  333: 
  334: Configuring Xen
  335: ---------------
  336: 
  337: Now, you have a system that will boot Xen and the dom0 kernel, but not
  338: do anything else special.  Make sure that you have rebooted into Xen.
  339: There will be no domUs, and none can be started because you still have
  340: to configure the dom0 daemons.
  341: 
  342: The daemons which should be run vary with Xen version and with whether
  343: one is using xm or xl. Xen 4.2 and up packages use xl. To use xm with 4.2,
  344: edit xendomains to use xm instead.
  345: 
  346: For 4.1 and up, you should enable xencommons.  Not enabling xencommons
  347: will result in a hang; it is necessary to hit ^C on the console to let
  348: the machine finish booting.  If you are using xm (default in 4.1, or
  349: if you changed xendomains in 4.2), you should also enable xend:
  350: 
  351: [[!template id=programlisting text="""
  352: xend=YES # only if using xm, and only installed <= 4.2
  353: xencommons=YES
  354: """]]
  355: 
  356: TODO: Recommend for/against xen-watchdog.
  357: 
  358: After you have configured the daemons and either started them (in the
  359: order given) or rebooted, use xm or xl to inspect Xen's boot messages,
  360: available resources, and running domains.  An example with xl follows:
  361: 
  362: [[!template id=programlisting text="""
  363: # xl dmesg
  364: ... xen's boot info ...
  365: # xl info
  366: ... available memory, etc ...
  367: # xl list
  368: Name              Id  Mem(MB)  CPU  State  Time(s)  Console
  369: Domain-0           0       64    0  r----     58.1
  370: """]]
  371: 
  372: Xen logs will be in /var/log/xen.
  373: 
  374: ### Issues with xencommons
  375: 
  376: xencommons starts xenstored, which stores data on behalf of dom0 and
  377: domUs.  It does not currently work to stop and start xenstored.
  378: Certainly all domUs should be shutdown first, following the sort order
  379: of the rc.d scripts.  However, the dom0 sets up state with xenstored,
  380: and is not notified when xenstored exits, leading to not recreating
  381: the state when the new xenstored starts.  Until there's a mechanism to
  382: make this work, one should not expect to be able to restart xenstored
  383: (and thus xencommons).  There is currently no reason to expect that
  384: this will get fixed any time soon.
  385: 
  386: ### No-longer needed advice about devices
  387: 
  388: The installation of NetBSD should already have created devices for xen
  389: (xencons, xenevt, xsd_kva), but if they are not present, create them:
  390: 
  391: [[!template id=programlisting text="""
  392: cd /dev && sh MAKEDEV xen
  393: """]]
  394: 
  395: anita (for testing NetBSD)
  396: --------------------------
  397: 
  398: With the setup so far (assuming 4.2/xl), one should be able to run
  399: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
  400: root, because anita must create a domU):
  401: 
  402: [[!template id=programlisting text="""
  403: anita --vmm=xl test file:///usr/obj/i386/
  404: """]]
  405: 
  406: Alternatively, one can use --vmm=xm to use xm-based domU creation
  407: instead (and must, on Xen <= 4.1).   TODO: confirm that anita xl really works.
  408:     
  409: Xen-specific NetBSD issues
  410: --------------------------
  411: 
  412: There are (at least) two additional things different about NetBSD as a
  413: dom0 kernel compared to hardware.
  414: 
  415: One is that the module ABI is different because some of the #defines
  416: change, so one must build modules for Xen.  As of netbsd-7, the build
  417: system does this automatically.  TODO: check this.  (Before building
  418: Xen modules was added, it was awkward to use modules to the point
  419: where it was considered that it did not work.)
  420: 
  421: The other difference is that XEN3_DOM0 does not have exactly the same
  422: options as GENERIC.  While it is debatable whether or not this is a
  423: bug, users should be aware of this and can simply add missing config
  424: items if desired.
  425: 
  426: Updating NetBSD in a dom0
  427: -------------------------
  428: 
  429: This is just like updating NetBSD on bare hardware, assuming the new
  430: version supports the version of Xen you are running.  Generally, one
  431: replaces the kernel and reboots, and then overlays userland binaries
  432: and adjusts /etc.
  433: 
  434: Note that one must update both the non-Xen kernel typically used for
  435: rescue purposes and the DOM0 kernel used with Xen.
  436: 
  437: Converting from grub to /boot
  438: -----------------------------
  439: 
  440: These instructions were [TODO: will be] used to convert a system from
  441: grub to /boot.  The system was originally installed in February of
  442: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
  443: over time.  Before these commands, it was running NetBSD 6 i386, Xen
  444: 4.1 and grub, much like the message linked earlier in the grub
  445: section.
  446: 
  447:         # Install MBR bootblocks on both disks. 
  448:         fdisk -i /dev/rwd0d
  449:         fdisk -i /dev/rwd1d
  450:         # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
  451:         installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
  452:         installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
  453:         # Install secondary boot loader
  454:         cp -p /usr/mdec/boot /
  455:         # Create boot.cfg following earlier guidance:
  456:         menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=512M
  457:         menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=512M
  458:         menu=GENERIC:boot
  459:         menu=GENERIC single-user:boot -s
  460:         menu=GENERIC.ok:boot netbsd.ok
  461:         menu=GENERIC.ok single-user:boot netbsd.ok -s
  462:         menu=Drop to boot prompt:prompt
  463:         default=1
  464:         timeout=30
  465: 
  466: TODO: actually do this and fix it if necessary.
  467: 
  468: Upgrading Xen versions
  469: ---------------------
  470: 
  471: Minor version upgrades are trivial.  Just rebuild/replace the
  472: xenkernel version and copy the new xen.gz to / (where /boot.cfg
  473: references it), and reboot.
  474: 
  475: Major version upgrades are conceptually not difficult, but can run
  476: into all the issues found when installing Xen.  Assuming migration
  477: from 4.1 to 4.2, remove the xenkernel41 and xentools41 packages and
  478: install the xenkernel42 and xentools42 packages.  Copy the 4.2 xen.gz
  479: to /.
  480: 
  481: Ensure that the contents of /etc/rc.d/xen* are correct.  Specifically,
  482: they must match the package you just installed and not be left over
  483: from some previous installation.
  484: 
  485: Enable the correct set of daemons; see the configuring section above.
  486: (Upgrading from 3.x to 4.x without doing this will result in a hang.)
  487: 
  488: Ensure that the domU config files are valid for the new version.
  489: Specifically, for 4.x remove autorestart=True, and ensure that disks
  490: are specified with numbers as the second argument, as the examples
  491: above show, and not NetBSD device names.
  492: 
  493: Unprivileged domains (domU)
  494: ===========================
  495: 
  496: This section describes general concepts about domUs.  It does not
  497: address specific domU operating systems or how to install them.  The
  498: config files for domUs are typically in /usr/pkg/etc/xen, and are
  499: typically named so that the file name, domU name and the domU's host
  500: name match.
  501: 
  502: The domU is provided with CPU and memory by Xen, configured by the
  503: dom0.  The domU is provided with disk and network by the dom0,
  504: mediated by Xen, and configured in the dom0.
  505: 
  506: Entropy in domUs can be an issue; physical disks and network are on
  507: the dom0.  NetBSD's /dev/random system works, but is often challenged.
  508: 
  509: Config files
  510: ------------
  511: 
  512: There is no good order to present config files and the concepts
  513: surrounding what is being configured.  We first show an example config
  514: file, and then in the various sections give details.
  515: 
  516: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
  517: for a large number of well-commented examples, mostly for running
  518: GNU/Linux.
  519: 
  520: The following is an example minimal domain configuration file
  521: "/usr/pkg/etc/xen/foo".  It is (with only a name change) an actual
  522: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
  523: i386 domU).  The domU serves as a network file server.
  524: 
  525:         # -*- mode: python; -*-
  526: 
  527:         kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
  528:         memory = 1024
  529:         vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
  530:         disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
  531:                  'file:/n0/xen/foo-wd1,0x1,w' ]
  532: 
  533: The domain will have the same name as the file.  The kernel has the
  534: host/domU name in it, so that on the dom0 one can update the various
  535: domUs independently.  The vif line causes an interface to be provided,
  536: with a specific mac address (do not reuse MAC addresses!), in bridge
  537: mode.  Two disks are provided, and they are both writable; the bits
  538: are stored in files and Xen attaches them to a vnd(4) device in the
  539: dom0 on domain creation.  The system treats xbd0 as the boot device
  540: without needing explicit configuration.
  541: 
  542: By default xm looks for domain config files in /usr/pkg/etc/xen.  Note
  543: that "xm create" takes the name of a config file, while other commands
  544: take the name of a domain.  To create the domain, connect to the
  545: console, create the domain while attaching the console, shutdown the
  546: domain, and see if it has finished stopping, do (or xl with Xen >=
  547: 4.2):
  548: 
  549:         xm create foo
  550:         xm console foo
  551:         xm create -c foo
  552:         xm shutdown foo
  553:         xm list
  554: 
  555: Typing ^] will exit the console session.  Shutting down a domain is
  556: equivalent to pushing the power button; a NetBSD domU will receive a
  557: power-press event and do a clean shutdown.  Shutting down the dom0
  558: will trigger controlled shutdowns of all configured domUs.
  559: 
  560: domU kernels
  561: ------------
  562: 
  563: On a physical computer, the BIOS reads sector 0, and a chain of boot
  564: loaders finds and loads a kernel.  Normally this comes from the root
  565: file system.  With Xen domUs, the process is totally different.  The
  566: normal path is for the domU kernel to be a file in the dom0's
  567: file system.  At the request of the dom0, Xen loads that kernel into a
  568: new domU instance and starts execution.  While domU kernels can be
  569: anyplace, reasonable places to store domU kernels on the dom0 are in /
  570: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
  571: config files), or in /u0/xen (where the vdisks are).
  572: 
  573: Note that loading the domU kernel from the dom0 implies that boot
  574: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
  575: See the VPS section near the end for discussion of alternate ways to
  576: obtain domU kernels.
  577: 
  578: CPU and memory
  579: --------------
  580: 
  581: A domain is provided with some number of vcpus, less than the number
  582: of CPUs seen by the hypervisor.  (For a dom0, this is controlled by
  583: the boot argument "dom0_max_vcpus=1".)  For a domU, it is controlled
  584: from the config file by the "vcpus = N" directive.
  585: 
  586: A domain is provided with memory; this is controlled in the config
  587: file by "memory = N" (in megabytes).  In the straightforward case, the
  588: sum of the the memory allocated to the dom0 and all domUs must be less
  589: than the available memory.
  590: 
  591: Xen also provides a "balloon" driver, which can be used to let domains
  592: use more memory temporarily.  TODO: Explain better, and explain how
  593: well it works with NetBSD.
  594: 
  595: Virtual disks
  596: -------------
  597: 
  598: With the file/vnd style, typically one creates a directory,
  599: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
  600: domUs.  Then, for each domU disk, one writes zeros to a file that then
  601: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
  602: for the first virtual disk for the domU called foo.  Writing zeros to
  603: the file serves two purposes.  One is that preallocating the contents
  604: improves performance.  The other is that vnd on sparse files has
  605: failed to work.  TODO: give working/notworking NetBSD versions for
  606: sparse vnd and gnats reference.  Note that the use of file/vnd for Xen
  607: is not really different than creating a file-backed virtual disk for
  608: some other purpose, except that xentools handles the vnconfig
  609: commands.  To create an empty 4G virtual disk, simply do
  610: 
  611:         dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
  612: 
  613: Do not use qemu-img-xen, because this will create sparse file.  There
  614: have been recent (2015) reports of sparse vnd(4) devices causing
  615: lockups, but there is apparently no PR.
  616: 
  617: With the lvm style, one creates logical devices.  They are then used
  618: similarly to vnds.  TODO: Add an example with lvm.
  619: 
  620: In domU config files, the disks are defined as a sequence of 3-tuples.
  621: The first element is "method:/path/to/disk".  Common methods are
  622: "file:" for file-backed vnd. and "phy:" for something that is already
  623: a (TODO: character or block) device.
  624: 
  625: The second element is an artifact of how virtual disks are passed to
  626: Linux, and a source of confusion with NetBSD Xen usage.  Linux domUs
  627: are given a device name to associate with the disk, and values like
  628: "hda1" or "sda1" are common.  In a NetBSD domU, the first disk appears
  629: as xbd0, the second as xbd1, and so on.  However, xm/xl demand a
  630: second argument.  The name given is converted to a major/minor by
  631: calling stat(2) on the name in /dev and this is passed to the domU.
  632: In the general case, the dom0 and domU can be different operating
  633: systems, and it is an unwarranted assumption that they have consistent
  634: numbering in /dev, or even that the dom0 OS has a /dev.  With NetBSD
  635: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
  636: for the second works fine and avoids this issue.  For a GNU/Linux
  637: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
  638: /dev/hda1.
  639: 
  640: The third element is "w" for writable disks, and "r" for read-only
  641: disks.
  642: 
  643: Note that NetBSD by default creates only vnd[0123].  If you need more
  644: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
  645: dom0.
  646: 
  647: Note that NetBSD by default creates only xbd[0123].  If you need more
  648: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
  649: 
  650: Virtual Networking
  651: ------------------
  652: 
  653: Xen provides virtual Ethernets, each of which connects the dom0 and a
  654: domU.  For each virtual network, there is an interface "xvifN.M" in
  655: the dom0, and in domU index N, a matching interface xennetM (NetBSD
  656: name).  The interfaces behave as if there is an Ethernet with two
  657: adapters connected.  From this primitive, one can construct various
  658: configurations.  We focus on two common and useful cases for which
  659: there are existing scripts: bridging and NAT.
  660: 
  661: With bridging (in the example above), the domU perceives itself to be
  662: on the same network as the dom0.  For server virtualization, this is
  663: usually best.  Bridging is accomplished by creating a bridge(4) device
  664: and adding the dom0's physical interface and the various xvifN.0
  665: interfaces to the bridge.  One specifies "bridge=bridge0" in the domU
  666: config file.  The bridge must be set up already in the dom0; an
  667: example /etc/ifconfig.bridge0 is:
  668: 
  669:         create
  670:         up
  671:         !brconfig bridge0 add wm0
  672: 
  673: With NAT, the domU perceives itself to be behind a NAT running on the
  674: dom0.  This is often appropriate when running Xen on a workstation.
  675: TODO: NAT appears to be configured by "vif = [ '' ]".
  676: 
  677: The MAC address specified is the one used for the interface in the new
  678: domain.  The interface in dom0 will use this address XOR'd with
  679: 00:00:00:01:00:00.  Random MAC addresses are assigned if not given.
  680: 
  681: Sizing domains
  682: --------------
  683: 
  684: Modern x86 hardware has vast amounts of resources.  However, many
  685: virtual servers can function just fine on far less.  A system with
  686: 512M of RAM and a 4G disk can be a reasonable choice.  Note that it is
  687: far easier to adjust virtual resources than physical ones.  For
  688: memory, it's just a config file edit and a reboot.  For disk, one can
  689: create a new file and vnconfig it (or lvm), and then dump/restore,
  690: just like updating physical disks, but without having to be there and
  691: without those pesky connectors.
  692: 
  693: Starting domains automatically
  694: ------------------------------
  695: 
  696: To start domains foo at bar at boot and shut them down cleanly on dom0
  697: shutdown, in rc.conf add:
  698: 
  699:         xendomains="foo bar"
  700: 
  701: Note that earlier versions of the xentools41 xendomains rc.d script
  702: used xl, when one should use xm with 4.1.
  703: 
  704: Creating specific unprivileged domains (domU)
  705: =============================================
  706: 
  707: Creating domUs is almost entirely independent of operating system.  We
  708: have already presented the basics of config files.  Note that you must
  709: have already completed the dom0 setup so that "xl list" (or "xm list")
  710: works.
  711: 
  712: Creating an unprivileged NetBSD domain (domU)
  713: ---------------------------------------------
  714: 
  715: See the earlier config file, and adjust memory.  Decide on how much
  716: storage you will provide, and prepare it (file or lvm).
  717: 
  718: While the kernel will be obtained from the dom0 file system, the same
  719: file should be present in the domU as /netbsd so that tools like
  720: savecore(8) can work.   (This is helpful but not necessary.)
  721: 
  722: The kernel must be specifically for Xen and for use as a domU.  The
  723: i386 and amd64 provide the following kernels:
  724: 
  725:         i386 XEN3_DOMU
  726:         i386 XEN3PAE_DOMU
  727:         amd64 XEN3_DOMU
  728: 
  729: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
  730: use the PAE version of the i386 kernel.
  731: 
  732: This will boot NetBSD, but this is not that useful if the disk is
  733: empty.  One approach is to unpack sets onto the disk outside of xen
  734: (by mounting it, just as you would prepare a physical disk for a
  735: system you can't run the installer on).
  736: 
  737: A second approach is to run an INSTALL kernel, which has a miniroot
  738: and can load sets from the network.  To do this, copy the INSTALL
  739: kernel to / and change the kernel line in the config file to:
  740: 
  741:         kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
  742: 
  743: Then, start the domain as "xl create -c configname".
  744: 
  745: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
  746: line should be used in the config file.
  747: 
  748:     disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
  749: 
  750: After booting the domain, the option to install via CDROM may be
  751: selected.  The CDROM device should be changed to `xbd1d`.
  752: 
  753: Once done installing, "halt -p" the new domain (don't reboot or halt,
  754: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
  755: config file), switch the config file back to the XEN3_DOMU kernel,
  756: and start the new domain again. Now it should be able to use "root on
  757: xbd0a" and you should have a, functional NetBSD domU.
  758: 
  759: TODO: check if this is still accurate.
  760: When the new domain is booting you'll see some warnings about *wscons*
  761: and the pseudo-terminals. These can be fixed by editing the files
  762: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
  763: `/etc/ttys`, except *console*, like this:
  764: 
  765:     console "/usr/libexec/getty Pc"         vt100   on secure
  766:     ttyE0   "/usr/libexec/getty Pc"         vt220   off secure
  767:     ttyE1   "/usr/libexec/getty Pc"         vt220   off secure
  768:     ttyE2   "/usr/libexec/getty Pc"         vt220   off secure
  769:     ttyE3   "/usr/libexec/getty Pc"         vt220   off secure
  770: 
  771: Finally, all screens must be commented out from `/etc/wscons.conf`.
  772: 
  773: It is also desirable to add
  774: 
  775:         powerd=YES
  776: 
  777: in rc.conf. This way, the domain will be properly shut down if
  778: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
  779: 
  780: It is not strictly necessary to have a kernel (as /netbsd) in the domU
  781: file system.  However, various programs (e.g. netstat) will use that
  782: kernel to look up symbols to read from kernel virtual memory.  If
  783: /netbsd is not the running kernel, those lookups will fail.  (This is
  784: not really a Xen-specific issue, but because the domU kernel is
  785: obtained from the dom0, it is far more likely to be out of sync or
  786: missing with Xen.)
  787: 
  788: Creating an unprivileged Linux domain (domU)
  789: --------------------------------------------
  790: 
  791: Creating unprivileged Linux domains isn't much different from
  792: unprivileged NetBSD domains, but there are some details to know.
  793: 
  794: First, the second parameter passed to the disk declaration (the '0x1' in
  795: the example below)
  796: 
  797:     disk = [ 'phy:/dev/wd0e,0x1,w' ]
  798: 
  799: does matter to Linux. It wants a Linux device number here (e.g. 0x300
  800: for hda).  Linux builds device numbers as: (major \<\< 8 + minor).
  801: So, hda1 which has major 3 and minor 1 on a Linux system will have
  802: device number 0x301.  Alternatively, devices names can be used (hda,
  803: hdb, ...)  as xentools has a table to map these names to devices
  804: numbers.  To export a partition to a Linux guest we can use:
  805: 
  806:         disk = [ 'phy:/dev/wd0e,0x300,w' ]
  807:         root = "/dev/hda1 ro"
  808: 
  809: and it will appear as /dev/hda on the Linux system, and be used as root
  810: partition.
  811: 
  812: To install the Linux system on the partition to be exported to the
  813: guest domain, the following method can be used: install
  814: sysutils/e2fsprogs from pkgsrc.  Use mke2fs to format the partition
  815: that will be the root partition of your Linux domain, and mount it.
  816: Then copy the files from a working Linux system, make adjustments in
  817: `/etc` (fstab, network config).  It should also be possible to extract
  818: binary packages such as .rpm or .deb directly to the mounted partition
  819: using the appropriate tool, possibly running under NetBSD's Linux
  820: emulation.  Once the file system has been populated, umount it.  If
  821: desirable, the file system can be converted to ext3 using tune2fs -j.
  822: It should now be possible to boot the Linux guest domain, using one of
  823: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
  824: 
  825: To get the Linux console right, you need to add:
  826: 
  827:     extra = "xencons=tty1"
  828: 
  829: to your configuration since not all Linux distributions auto-attach a
  830: tty to the xen console.
  831: 
  832: Creating an unprivileged Solaris domain (domU)
  833: ----------------------------------------------
  834: 
  835: See possibly outdated
  836: [Solaris domU instructions](/ports/xen/howto-solaris/).
  837: 
  838: 
  839: PCI passthrough: Using PCI devices in guest domains
  840: ---------------------------------------------------
  841: 
  842: The dom0 can give other domains access to selected PCI
  843: devices. This can allow, for example, a non-privileged domain to have
  844: access to a physical network interface or disk controller.  However,
  845: keep in mind that giving a domain access to a PCI device most likely
  846: will give the domain read/write access to the whole physical memory,
  847: as PCs don't have an IOMMU to restrict memory access to DMA-capable
  848: device.  Also, it's not possible to export ISA devices to non-dom0
  849: domains, which means that the primary VGA adapter can't be exported.
  850: A guest domain trying to access the VGA registers will panic.
  851: 
  852: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
  853: not been ported to later versions at this time.
  854: 
  855: For a PCI device to be exported to a domU, is has to be attached to
  856: the "pciback" driver in dom0.  Devices passed to the dom0 via the
  857: pciback.hide boot parameter will attach to "pciback" instead of the
  858: usual driver.  The list of devices is specified as "(bus:dev.func)",
  859: where bus and dev are 2-digit hexadecimal numbers, and func a
  860: single-digit number:
  861: 
  862:         pciback.hide=(00:0a.0)(00:06.0)
  863: 
  864: pciback devices should show up in the dom0's boot messages, and the
  865: devices should be listed in the `/kern/xen/pci` directory.
  866: 
  867: PCI devices to be exported to a domU are listed in the "pci" array of
  868: the domU's config file, with the format "0000:bus:dev.func".
  869: 
  870:         pci = [ '0000:00:06.0', '0000:00:0a.0' ]
  871: 
  872: In the domU an "xpci" device will show up, to which one or more pci
  873: buses will attach.  Then the PCI drivers will attach to PCI buses as
  874: usual.  Note that the default NetBSD DOMU kernels do not have "xpci"
  875: or any PCI drivers built in by default; you have to build your own
  876: kernel to use PCI devices in a domU.  Here's a kernel config example;
  877: note that only the "xpci" lines are unusual.
  878: 
  879:         include         "arch/i386/conf/XEN3_DOMU"
  880: 
  881:         # Add support for PCI buses to the XEN3_DOMU kernel
  882:         xpci* at xenbus ?
  883:         pci* at xpci ?
  884: 
  885:         # PCI USB controllers
  886:         uhci*   at pci? dev ? function ?        # Universal Host Controller (Intel)
  887: 
  888:         # USB bus support
  889:         usb*    at uhci?
  890: 
  891:         # USB Hubs
  892:         uhub*   at usb?
  893:         uhub*   at uhub? port ? configuration ? interface ?
  894: 
  895:         # USB Mass Storage
  896:         umass*  at uhub? port ? configuration ? interface ?
  897:         wd*     at umass?
  898:         # SCSI controllers
  899:         ahc*    at pci? dev ? function ?        # Adaptec [23]94x, aic78x0 SCSI
  900: 
  901:         # SCSI bus support (for both ahc and umass)
  902:         scsibus* at scsi?
  903: 
  904:         # SCSI devices
  905:         sd*     at scsibus? target ? lun ?      # SCSI disk drives
  906:         cd*     at scsibus? target ? lun ?      # SCSI CD-ROM drives
  907: 
  908: 
  909: NetBSD as a domU in a VPS
  910: =========================
  911: 
  912: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
  913: hardware.  This section explains how to deal with Xen in a domU as a
  914: virtual private server where you do not control or have access to the
  915: dom0.  This is not intended to be an exhaustive list of VPS providers;
  916: only a few are mentioned that specifically support NetBSD.
  917: 
  918: VPS operators provide varying degrees of access and mechanisms for
  919: configuration.  The big issue is usually how one controls which kernel
  920: is booted, because the kernel is nominally in the dom0 file system (to
  921: which VPS users do not normally have access).  A second issue is how
  922: to install NetBSD.
  923: A VPS user may want to compile a kernel for security updates, to run
  924: npf, run IPsec, or any other reason why someone would want to change
  925: their kernel.
  926: 
  927: One approach is to have an administrative interface to upload a kernel,
  928: or to select from a prepopulated list.  Other approaches are pygrub
  929: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
  930: kernel from the domU file system.  This is closer to a regular physical
  931: computer, where someone who controls a machine can replace the kernel.
  932: 
  933: A second issue is multiple CPUs.  With NetBSD 6, domUs support
  934: multiple vcpus, and it is typical for VPS providers to enable multiple
  935: CPUs for NetBSD domUs.
  936: 
  937: pygrub
  938: -------
  939: 
  940: pygrub runs in the dom0 and looks into the domU file system.  This
  941: implies that the domU must have a kernel in a file system in a format
  942: known to pygrub.  As of 2014, pygrub seems to be of mostly historical
  943: interest.
  944: 
  945: pvgrub
  946: ------
  947: 
  948: pvgrub is a version of grub that uses PV operations instead of BIOS
  949: calls.  It is booted from the dom0 as the domU kernel, and then reads
  950: /grub/menu.lst and loads a kernel from the domU file system.
  951: 
  952: [Panix](http://www.panix.com/) lets users use pvgrub.  Panix reports
  953: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
  954: (and hence with defaults from "newfs -O 2").  See [Panix's pvgrub
  955: page](http://www.panix.com/v-colo/grub.html), which describes only
  956: Linux but should be updated to cover NetBSD :-).
  957: 
  958: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
  959: their own kernel.  See then [prgmr.com NetBSD
  960: HOWTO](http://wiki.prgmr.com/mediawiki/index.php/NetBSD_as_a_DomU)
  961: (which is in need of updating).
  962: 
  963: It appears that [grub's FFS
  964: code](http://xenbits.xensource.com/hg/xen-unstable.hg/file/bca284f67702/tools/libfsimage/ufs/fsys_ufs.c)
  965: does not support all aspects of modern FFS, but there are also reports
  966: that FFSv2 works fine.  At prgmr, typically one has an ext2 or FAT
  967: partition for the kernel with the intent that grub can understand it,
  968: which leads to /netbsd not being the actual kernel.  One must remember
  969: to update the special boot partition.
  970: 
  971: Amazon
  972: ------
  973: 
  974: See the [Amazon EC2 page](/amazon_ec2/).
  975: 
  976: TODO items for improving NetBSD/xen
  977: ===================================
  978: 
  979: * Make the NetBSD dom0 kernel work with SMP.
  980: * Test the Xen 4.5 packages adequately to be able to recommend them as
  981:   the standard approach.
  982: * Get PCI passthrough working on Xen 4.5
  983: * Get pvgrub into pkgsrc, either via xentools or separately.
  984: * grub
  985:   * Check/add support to pkgsrc grub2 for UFS2 and arbitrary
  986:     fragsize/blocksize (UFS2 support may be present; the point is to
  987:     make it so that with any UFS1/UFS2 file system setup that works
  988:     with NetBSD grub will also work).
  989:     See [pkg/40258](https://gnats.netbsd.org/40258).
  990:   * Push patches upstream.
  991:   * Get UFS2 patches into pvgrub.
  992: * Add support for PV ops to a version of /boot, and make it usable as
  993:   a kernel in Xen, similar to pvgrub.
  994: 
  995: Random pointers
  996: ===============
  997: 
  998: This section contains links from elsewhere not yet integrated into the
  999: HOWTO, and other guides.
 1000: 
 1001: * http://www.lumbercartel.ca/library/xen/
 1002: * http://pbraun.nethence.com/doc/sysutils/xen_netbsd_dom0.html
 1003: * https://gmplib.org/~tege/xen.html

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