7: Xen is a hypervisor (or virtual machine monitor) for x86 hardware
8: (i686-class or higher), which supports running multiple guest
9: operating systems on a single physical machine. Xen is a Type 1 or
10: bare-metal hypervisor; one uses the Xen kernel to control the CPU,
11: memory and console, a dom0 operating system which mediates access to
12: other hardware (e.g., disks, network, USB), and one or more domU
13: operating systems which operate in an unprivileged virtualized
14: environment. IO requests from the domU systems are forwarded by the
15: hypervisor (Xen) to the dom0 to be fulfilled.
17: Xen supports two styles of guests. The original is Para-Virtualized
18: (PV) which means that the guest OS does not attempt to access hardware
19: directly, but instead makes hypercalls to the hypervisor. This is
20: analogous to a user-space program making system calls. (The dom0
21: operating system uses PV calls for some functions, such as updating
22: memory mapping page tables, but has direct hardware access for disk
23: and network.) PV guests must be specifically coded for Xen.
25: The more recent style is HVM, which means that the guest does not have
26: code for Xen and need not be aware that it is running under Xen.
27: Attempts to access hardware registers are trapped and emulated. This
28: style is less efficient but can run unmodified guests.
30: Generally any machine that runs NetBSD/amd64 will work with Xen and PV
31: guests. In theory i386 computers (without x86_64/amd64 support) can
32: be used for Xen <= 4.2, but we have no recent reports of this working
33: (this is a hint). For HVM guests, hardware support is needed, but it
34: is common on recent machines. For Intel CPUs, one needs the VT-x
35: extension, shown in "cpuctl identify 0" as VMX. For AMD CPUs, one
36: needs the AMD-V extensions, shown in "cpuctl identify 0" as SVM.
37: There are further features for IOMMU virtualization, Intel's VT-d and
38: AMD's AMD-Vi. TODO: Explain whether Xen on NetBSD makes use of these
39: features. TODO: Review by someone who really understands this.
41: Note that a FreeBSD dom0 requires VT-x and VT-d (or equivalent); this
42: is because the FreeBSD dom0 does not run in PV mode.
44: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
45: The dom0 can start one or more domUs. (Booting is explained in detail
46: in the dom0 section.)
48: NetBSD supports Xen in that it can serve as dom0, be used as a domU,
49: and that Xen kernels and tools are available in pkgsrc. This HOWTO
50: attempts to address both the case of running a NetBSD dom0 on hardware
51: and running domUs under it (NetBSD and other), and also running NetBSD
52: as a domU in a VPS.
54: Xen 3.1 in pkgsrc supports "PCI passthrough", which means that
55: specific PCI devices can be made available to a specific domU instead
56: of the dom0. This can be useful to let a domU run X11, or access some
57: network interface or other peripheral.
59: NetBSD 6 and earlier supported Xen 2; support was removed from NetBSD
60: 7. Xen 2 has been removed from pkgsrc.
65: Installing NetBSD/Xen is not extremely difficult, but it is more
66: complex than a normal installation of NetBSD.
67: In general, this HOWTO is occasionally overly restrictive about how
68: things must be done, guiding the reader to stay on the established
69: path when there are no known good reasons to stray.
71: This HOWTO presumes a basic familiarity with the Xen system
72: architecture, with installing NetBSD on i386/amd64 hardware, and with
73: installing software from pkgsrc. See also the [Xen
76: Versions of Xen and NetBSD
79: Most of the installation concepts and instructions are independent
80: of Xen version and NetBSD version. This section gives advice on
81: which version to choose. Versions not in pkgsrc and older unsupported
82: versions of NetBSD are intentionally ignored.
87: In NetBSD, Xen is provided in pkgsrc, via matching pairs of packages
88: xenkernel and xentools. We will refer only to the kernel versions,
89: but note that both packages must be installed together and must have
90: matching versions.
92: xenkernel3 provides Xen 3.1. It is no longer maintained by Xen, and the last applied security patch was in
93: 2011. Thus, it should not be used. It supports PCI passthrough,
94: which is why people use it anyway. Xen 3.1 supports i386, both PAE and
97: xenkernel33 provides Xen 3.3. It is no longer maintained by Xen, and
98: the last applied security patch was in 2012. Thus, it should not be
99: used. Xen 3.3 supports i386, but only in PAE mode. There are no good
100: reasons to run this version.
102: xenkernel41 provides Xen 4.1. It is no longer maintained by Xen, but
103: as of 2016-12 received backported security patches. Xen 4.1 supports
104: i386, but only in PAE mode. There are no good reasons to run this
107: xenkernel42 provides Xen 4.2. It is no longer maintained by Xen, but
108: as of 2016-12 received backported security patches. Xen 4.2 supports
109: i386, but only in PAE mode. The only reason to run this is if you
110: need to use xm instead of xl, or if you need to run an i386 dom0
111: (because your hardware is i386 only).
113: xenkernel45 provides Xen 4.5. As of 2016-12, security patches were
114: released by Xen and applied to pkgsrc. Xen 4.5 requires using a dom0
115: running NetBSD's amd64 port (Intel or AMD hardware is fine), but domUs
116: can be amd64 or i386 PAE. TODO: It is either a conservative choice or
117: somewhat old.
119: xenkernel46 provides Xen 4.6. It is new to pkgsrc as of 2016-05. As
120: of 2016-12, security patches were released by Xen and applied to
121: pkgsrc. Xen 4.6 similarly requires a NetBSD/amd64 dom0, but domUs can
122: be amd64 or i386 PAE. TODO: It is either a somewhat aggressive choice
123: or the standard choice.
125: Xen 4.7 (released 2016-06) and 4.8 (released 2016-12) are not yet in
128: See also the [Xen Security Advisory page](http://xenbits.xen.org/xsa/).
130: Note that NetBSD support is called XEN3. It works with Xen 3 and Xen
131: 4 because the hypercall interface has been stable.
133: Xen command program
136: Early Xen used a program called xm to manipulate the system from the
137: dom0. Starting in 4.1, a replacement program with similar behavior
138: called xl is provided, but it does not work well in 4.1. In 4.2, both
139: xm and xl work fine. 4.4 is the last version that has xm.
141: You must make a global choice to use xm or xl, because it affects not
142: only which command you use, but the command used by rc.d scripts
143: (specifically xendomains) and which daemons should be run. The
144: xentools packages provide xm for 3.1, 3.3 and 4.1 and xl for 4.2 and up.
146: In 4.2, you can choose to use xm by simply changing the ctl_command
149: With xl, virtual devices are configured in parallel, which can cause
150: problems if they are written assuming serial operation (e.g., updating
151: firewall rules without explicit locking).
153: There is now locking for xl startup, but as of 201612 it has not been
154: adequately tested.
159: The netbsd-6, netbsd-7, and -current branches are all reasonable
160: choices, with more or less the same considerations for non-Xen use.
161: Therefore, netbsd-7 is recommended as the stable version of the most
162: recent release for production use. In addition, netbsd-7 and -current
163: have a important scheduler fix (in November of 2015) affecting
164: contention between dom0 and domUs; see
165: https://releng.netbsd.org/cgi-bin/req-7.cgi?show=1040 for a
166: description. For those wanting to learn Xen or without production
167: stability concerns, netbsd-7 is still likely most appropriate, but
168: -current is also a reasonable choice. (Xen runs ok on netbsd-5, but
169: the xentools packages are likely difficult to build, and netbsd-5 is
170: not supported.)
172: As of NetBSD 6, a NetBSD domU will support multiple vcpus. There is
173: no SMP support for NetBSD as dom0. (The dom0 itself doesn't really
174: need SMP for dom0 functions; the lack of support is really a problem
175: when using a dom0 as a normal computer.)
180: Xen itself can run on i386 (Xen < 3.1) or amd64 machines (all Xen
181: versions). (Practically, almost any computer where one would want to
182: run Xen today supports amd64.)
184: Xen, the dom0 kernel, and each domU kernel can be either i386 or
185: amd64. When building a xenkernel package, one obtains i386 on an i386
186: host, and amd64 on an amd64 host. If the Xen kernel is i386, then the
187: dom0 kernel and all domU kernels must be i386. With an amd64 Xen
188: kernel, an amd64 dom0 kernel is known to work, and an i386PAE dom0
189: kernel should in theory work. An amd64 Xen/dom0 is known to support
190: both i386PAE and amd64 domUs.
192: i386 dom0 and domU kernels must be PAE (except for Xen 3.1); these are
193: built by default. (Note that emacs (at least) fails if run on i386
194: with PAE when built without, and vice versa, presumably due to bugs in
195: the undump code.)
197: Because of the above, the standard approach is to use amd64 for the
200: Xen 4.2 is the last version to support i386 as a host. TODO: Clarify
201: if this is about the CPU, the Xen kernel, or the dom0 kernel having to
202: be amd64.
208: Mostly, NetBSD as a dom0 or domU is quite stable.
209: However, there are some open PRs indicating problems.
211: - [PR 48125](http://gnats.netbsd.org/48125)
212: - [PR 47720](http://gnats.netbsd.org/47720)
214: Note also that there are issues with sparse vnd(4) instances, but
215: these are not about Xen -- they just are noticed with sparse vnd(4)
216: instances in support of virtual disks in a dom0.
221: Therefore, this HOWTO recommends running xenkernel45 or xenkernel46,
222: xl, the NetBSD 7 stable branch, and to use an amd64 kernel as the
223: dom0. Either the i386PAE or amd64 version of NetBSD may be used as
229: Ideally, all versions of Xen in pkgsrc would build on all supported
230: versions of NetBSD/amd64, to the point where this section would be
231: silly. However, that has not always been the case. Besides aging
232: code and aging compilers, qemu (included in xentools for HVM support)
233: is difficult to build. Note that there is intentionally no data for
234: 4.5+ up for i386, and often omits xentools info if the corresponding
235: kernel fails.
237: The following table gives status, with the date last checked
238: (generally on the most recent quarterly branch). The first code is
239: "builds" if it builds ok, and "FAIL" for a failure to build. The
240: second code/date only appears for xenkernel* and is "works" if it runs
241: ok as a dom0, and "FAIL" if it won't boot or run a domU.
243: xenkernel3 netbsd-6 i386 FAIL 201612
244: xenkernel33 netbsd-6 i386 FAIL 201612
245: xenkernel41 netbsd-6 i386 builds 201612
246: xenkernel42 netbsd-6 i386 builds 201612
247: xentools3 netbsd-6 i386 FAIL 201612
248: xentools3-hvm netbsd-6 i386 FAIL 201412
249: xentools33 netbsd-6 i386 builds 201412
250: xentools41 netbsd-6 i386 builds 201612
251: xentools42 netbsd-6 i386 FAIL 201612
253: xenkernel3 netbsd-7 i386 FAIL 201412
254: xenkernel33 netbsd-7 i386 FAIL 201412
255: xenkernel41 netbsd-7 i386 builds 201412
256: xenkernel42 netbsd-7 i386 builds 201412
257: xentools41 netbsd-7 i386 builds 201412
258: xentools42 netbsd-7 i386 ??FAIL 201412
260: xenkernel3 netbsd-6 amd64 FAIL 201612
261: xenkernel33 netbsd-6 amd64 FAIL 201612
262: xenkernel41 netbsd-6 amd64 builds 201612 works 201612
263: xenkernel42 netbsd-6 amd64 builds 201612 works 201612
264: xenkernel45 netbsd-6 amd64 builds 201612
265: xenkernel46 netbsd-6 amd64 builds 201612
266: xentools41 netbsd-6 amd64 builds 201612
267: xentools42 netbsd-6 amd64 builds 201612
268: xentools45 netbsd-6 amd64 builds 201612
269: xentools46 netbsd-6 amd64 FAIL 201612
271: xenkernel3 netbsd-7 amd64 builds 201612
272: xenkernel33 netbsd-7 amd64 builds 201612
273: xenkernel41 netbsd-7 amd64 builds 201612
274: xenkernel42 netbsd-7 amd64 builds 201612
275: xenkernel45 netbsd-7 amd64 builds 201612
276: xenkernel46 netbsd-7 amd64 builds 201612
277: xentools3 netbsd-7 amd64 builds 201612
278: xentools3-hvm netbsd-7 amd64 builds 201612
279: xentools33 netbsd-7 amd64 FAIL 201612
280: xentools41 netbsd-7 amd64 builds 201612
281: xentools42 netbsd-7 amd64 builds 201612
282: xentools45 netbsd-7 amd64 builds 201612
283: xentools46 netbsd-7 amd64 builds 201612
285: NetBSD as a dom0
288: NetBSD can be used as a dom0 and works very well. The following
289: sections address installation, updating NetBSD, and updating Xen.
290: Note that it doesn't make sense to talk about installing a dom0 OS
291: without also installing Xen itself. We first address installing
292: NetBSD, which is not yet a dom0, and then adding Xen, pivoting the
293: NetBSD install to a dom0 install by just changing the kernel and boot
296: For experimenting with Xen, a machine with as little as 1G of RAM and
297: 100G of disk can work. For running many domUs in productions, far
298: more will be needed.
300: Styles of dom0 operation
303: There are two basic ways to use Xen. The traditional method is for
304: the dom0 to do absolutely nothing other than providing support to some
305: number of domUs. Such a system was probably installed for the sole
306: purpose of hosting domUs, and sits in a server room on a UPS.
308: The other way is to put Xen under a normal-usage computer, so that the
309: dom0 is what the computer would have been without Xen, perhaps a
310: desktop or laptop. Then, one can run domUs at will. Purists will
311: deride this as less secure than the previous approach, and for a
312: computer whose purpose is to run domUs, they are right. But Xen and a
313: dom0 (without domUs) is not meaningfully less secure than the same
314: things running without Xen. One can boot Xen or boot regular NetBSD
315: alternately with little problems, simply refraining from starting the
316: Xen daemons when not running Xen.
318: Note that NetBSD as dom0 does not support multiple CPUs. This will
319: limit the performance of the Xen/dom0 workstation approach. In theory
320: the only issue is that the "backend drivers" are not yet MPSAFE:
323: Installation of NetBSD
327: [install NetBSD/amd64](/guide/inst/)
328: just as you would if you were not using Xen.
329: However, the partitioning approach is very important.
331: If you want to use RAIDframe for the dom0, there are no special issues
332: for Xen. Typically one provides RAID storage for the dom0, and the
333: domU systems are unaware of RAID. The 2nd-stage loader bootxx_* skips
334: over a RAID1 header to find /boot from a file system within a RAID
335: partition; this is no different when booting Xen.
337: There are 4 styles of providing backing storage for the virtual disks
338: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN.
340: With raw partitions, one has a disklabel (or gpt) partition sized for
341: each virtual disk to be used by the domU. (If you are able to predict
342: how domU usage will evolve, please add an explanation to the HOWTO.
343: Seriously, needs tend to change over time.)
345: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
346: for domU disks. This is almost as efficient as raw disk partitions
347: and more flexible. Hence raw disk partitions should typically not
348: be used.
350: One can use files in the dom0 file system, typically created by dd'ing
351: /dev/zero to create a specific size. This is somewhat less efficient,
352: but very convenient, as one can cp the files for backup, or move them
353: between dom0 hosts.
355: Finally, in theory one can place the files backing the domU disks in a
356: SAN. (This is an invitation for someone who has done this to add a
357: HOWTO page.)
359: Installation of Xen
362: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
363: pkgsrc (or another matching pair). See [the pkgsrc
364: documentation](http://www.NetBSD.org/docs/pkgsrc/) for help with
365: pkgsrc. Ensure that your packages are recent; the HOWTO does not
366: contemplate old builds.
369: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm. More
370: recent versions have HVM support integrated in the main xentools
371: package. It is entirely reasonable to run only PV guests.
373: Next you need to install the selected Xen kernel itself, which is
374: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz". Copy it to /.
375: For debugging, one may copy xen-debug.gz; this is conceptually similar
376: to DIAGNOSTIC and DEBUG in NetBSD. xen-debug.gz is basically only
377: useful with a serial console. Then, place a NetBSD XEN3_DOM0 kernel
378: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
379: of a NetBSD build. If using i386, use
380: releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz. (If using Xen
381: 3.1 and i386, you may use XEN3_DOM0 with the non-PAE Xen. But you
382: should not use Xen 3.1.) Both xen and the NetBSD kernel may be (and
383: typically are) left compressed.
385: In a dom0 kernel, kernfs is mandatory for xend to communicate with the
386: kernel, so ensure that /kern is in fstab. TODO: Say this is default,
387: or file a PR and give a reference.
389: Because you already installed NetBSD, you have a working boot setup
390: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
391: beginning of your root file system, /boot present, and likely
392: /boot.cfg. (If not, fix before continuing!)
394: Add a line to to /boot.cfg to boot Xen. See boot.cfg(5) for an
395: example. The basic line is
397: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
399: which specifies that the dom0 should have 256M, leaving the rest to be
400: allocated for domUs. To use a serial console, use
402: menu=Xen:load /netbsd-XEN3_DOM0.gz console=com0;multiboot /xen.gz dom0_mem=256M console=com1 com1=9600,8n1
404: which will use the first serial port for Xen (which counts starting
405: from 1), forcing speed/parity, and also for NetBSD (which counts
406: starting at 0). In an attempt to add performance, one can also add
408: dom0_max_vcpus=1 dom0_vcpus_pin
410: to force only one vcpu to be provided (since NetBSD dom0 can't use
411: more) and to pin that vcpu to a physical CPU. TODO: benchmark this.
413: Xen has [many boot
415: and other than dom0 memory and max_vcpus, they are generally not
418: As with non-Xen systems, you should have a line to boot /netbsd (a
419: kernel that works without Xen). Consider a line to boot /netbsd.ok (a
420: fallback version of the non-Xen kernel, updated manually when you are
421: sure /netbsd is ok). Consider also a line to boot fallback versions
422: of Xen and the dom0 kernel, but note that non-Xen NetBSD can be used
423: to resolve Xen booting issues.
425: Probably you want a default=N line to choose Xen in the absence of
428: Now, reboot so that you are running a DOM0 kernel under Xen, rather
429: than GENERIC without Xen.
431: Using grub (historic)
434: Before NetBSD's native bootloader could support Xen, the use of
435: grub was recommended. If necessary, see the
436: [old grub information](/ports/xen/howto-grub/).
438: The [HowTo on Installing into
440: explains how to set up booting a dom0 with Xen using grub with
441: NetBSD's RAIDframe. (This is obsolete with the use of NetBSD's native
444: Configuring Xen
447: Xen logs will be in /var/log/xen.
449: Now, you have a system that will boot Xen and the dom0 kernel, but not
450: do anything else special. Make sure that you have rebooted into Xen.
451: There will be no domUs, and none can be started because you still have
452: to configure the dom0 daemons.
454: The daemons which should be run vary with Xen version and with whether
455: one is using xm or xl. The Xen 3.1, 3.3 and 4.1 packages use xm. Xen
456: 4.2 and up packages use xl. To use xm with 4.2, edit xendomains to
457: use xm instead.
460: For "xm", you should enable xend and xenbackendd:
465: For "xl", you should enable xend and xencommons (xenstored). Trying
466: to use xl without xencommons=YES will result in a hang; it is
467: necessary to hit ^C on the console to let the machine finish booting.
468: TODO: Check if xend really should be used, and if not why it is
474: TODO: Recommend for/against xen-watchdog.
476: After you have configured the daemons and either started them (in the
477: order given) or rebooted, use xm or xl to inspect Xen's boot messages,
478: available resources, and running domains. An example with xl follows:
480: # xl dmesg
481: [xen's boot info]
482: # xl info
483: [available memory, etc.]
484: # xl list
485: Name Id Mem(MB) CPU State Time(s) Console
486: Domain-0 0 64 0 r---- 58.1
488: ### Issues with xencommons
490: xencommons starts xenstored, which stores data on behalf of dom0 and
491: domUs. It does not currently work to stop and start xenstored.
492: Certainly all domUs should be shutdown first, following the sort order
493: of the rc.d scripts. However, the dom0 sets up state with xenstored,
494: and is not notified when xenstored exits, leading to not recreating
495: the state when the new xenstored starts. Until there's a mechanism to
496: make this work, one should not expect to be able to restart xenstored
497: (and thus xencommons). There is currently no reason to expect that
498: this will get fixed any time soon.
500: ### No-longer needed advice about devices
502: The installation of NetBSD should already have created devices for xen
503: (xencons, xenevt, xsd_kva), but if they are not present, create them:
505: cd /dev && sh MAKEDEV xen
507: anita (for testing NetBSD)
510: With the setup so far (assuming 4.2/xl), one should be able to run
511: anita (see pkgsrc/misc/py-anita) to test NetBSD releases, by doing (as
512: root, because anita must create a domU):
514: anita --vmm=xl test file:///usr/obj/i386/
516: Alternatively, one can use --vmm=xm to use xm-based domU creation
517: instead (and must, on Xen <= 4.1). TODO: confirm that anita xl really works.
519: Xen-specific NetBSD issues
522: There are (at least) two additional things different about NetBSD as a
523: dom0 kernel compared to hardware.
525: One is that the module ABI is different because some of the #defines
526: change, so one must build modules for Xen. As of netbsd-7, the build
527: system does this automatically. TODO: check this. (Before building
528: Xen modules was added, it was awkward to use modules to the point
529: where it was considered that it did not work.)
531: The other difference is that XEN3_DOM0 does not have exactly the same
532: options as GENERIC. While it is debatable whether or not this is a
533: bug, users should be aware of this and can simply add missing config
534: items if desired.
536: Updating NetBSD in a dom0
539: This is just like updating NetBSD on bare hardware, assuming the new
540: version supports the version of Xen you are running. Generally, one
541: replaces the kernel and reboots, and then overlays userland binaries
542: and adjusts /etc.
544: Note that one must update both the non-Xen kernel typically used for
545: rescue purposes and the DOM0 kernel used with Xen.
547: Converting from grub to /boot
550: These instructions were [TODO: will be] used to convert a system from
551: grub to /boot. The system was originally installed in February of
552: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
553: over time. Before these commands, it was running NetBSD 6 i386, Xen
554: 4.1 and grub, much like the message linked earlier in the grub
557: # Install MBR bootblocks on both disks.
558: fdisk -i /dev/rwd0d
559: fdisk -i /dev/rwd1d
560: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
561: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
562: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
563: # Install secondary boot loader
564: cp -p /usr/mdec/boot /
565: # Create boot.cfg following earlier guidance:
566: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
567: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=256M
569: menu=GENERIC single-user:boot -s
570: menu=GENERIC.ok:boot netbsd.ok
571: menu=GENERIC.ok single-user:boot netbsd.ok -s
572: menu=Drop to boot prompt:prompt
576: TODO: actually do this and fix it if necessary.
578: Upgrading Xen versions
581: Minor version upgrades are trivial. Just rebuild/replace the
582: xenkernel version and copy the new xen.gz to / (where /boot.cfg
583: references it), and reboot.
585: Major version upgrades are conceptually not difficult, but can run
586: into all the issues found when installing Xen. Assuming migration
587: from 4.1 to 4.2, remove the xenkernel41 and xentools41 packages and
588: install the xenkernel42 and xentools42 packages. Copy the 4.2 xen.gz
589: to /.
591: Ensure that the contents of /etc/rc.d/xen* are correct. Specifically,
592: they must match the package you just installed and not be left over
593: from some previous installation.
595: Enable the correct set of daemons; see the configuring section above.
596: (Upgrading from 3.x to 4.x without doing this will result in a hang.)
598: Ensure that the domU config files are valid for the new version.
599: Specifically, for 4.x remove autorestart=True, and ensure that disks
600: are specified with numbers as the second argument, as the examples
601: above show, and not NetBSD device names.
603: Hardware known to work
606: Arguably, this section is misplaced, and there should be a page of
607: hardware that runs NetBSD/amd64 well, with the mostly-well-founded
608: assumption that NetBSD/xen runs fine on any modern hardware that
609: NetBSD/amd64 runs well on. Until then, we give motherboard/CPU/RAM
610: triples to aid those choosing a motherboard. Note that Xen systems
611: usually do not run X, so a listing here does not imply that X works at
614: Supermicro X9SRL-F, Xeon E5-1650 v2, 96 GiB ECC
615: Supermicro ??, Atom C2758 (8 core), 32 GiB ECC
616: ASUS M5A78L-M/USB3 AM3+ microATX, AMD Piledriver X8 4000MHz, 16 GiB ECC
618: Older hardware:
620: Intel D915GEV, Pentium4 CPU 3.40GHz, 4GB 533MHz Synchronous DDR2
622: Running Xen under qemu
625: The astute reader will note that this section is somewhat twisted.
626: However, it can be useful to run Xen under qemu either because the
627: version of NetBSD as a dom0 does not run on the hardware in use, or to
628: generate automated test cases involving Xen.
630: In 2015-01, the following combination was reported to mostly work:
632: host OS: NetBSD/amd64 6.1.4
633: qemu: 2.2.0 from pkgsrc
634: Xen kernel: xenkernel42-4.2.5nb1 from pkgsrc
635: dom0 kernel: NetBSD/amd64 6.1.5
636: Xen tools: xentools42-4.2.5 from pkgsrc
638: See [PR 47720](http://gnats.netbsd.org/47720) for a problem with dom0
641: Unprivileged domains (domU)
644: This section describes general concepts about domUs. It does not
645: address specific domU operating systems or how to install them. The
646: config files for domUs are typically in /usr/pkg/etc/xen, and are
647: typically named so that the file name, domU name and the domU's host
648: name match.
650: The domU is provided with CPU and memory by Xen, configured by the
651: dom0. The domU is provided with disk and network by the dom0,
652: mediated by Xen, and configured in the dom0.
654: Entropy in domUs can be an issue; physical disks and network are on
655: the dom0. NetBSD's /dev/random system works, but is often challenged.
657: Config files
660: There is no good order to present config files and the concepts
661: surrounding what is being configured. We first show an example config
662: file, and then in the various sections give details.
664: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
665: for a large number of well-commented examples, mostly for running
668: The following is an example minimal domain configuration file
669: "/usr/pkg/etc/xen/foo". It is (with only a name change) an actual
670: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
671: i386 domU). The domU serves as a network file server.
673: # -*- mode: python; -*-
675: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
676: memory = 1024
677: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
678: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
679: 'file:/n0/xen/foo-wd1,0x1,w' ]
681: The domain will have the same name as the file. The kernel has the
682: host/domU name in it, so that on the dom0 one can update the various
683: domUs independently. The vif line causes an interface to be provided,
684: with a specific mac address (do not reuse MAC addresses!), in bridge
685: mode. Two disks are provided, and they are both writable; the bits
686: are stored in files and Xen attaches them to a vnd(4) device in the
687: dom0 on domain creation. The system treats xbd0 as the boot device
688: without needing explicit configuration.
690: By default xm looks for domain config files in /usr/pkg/etc/xen. Note
691: that "xm create" takes the name of a config file, while other commands
692: take the name of a domain. To create the domain, connect to the
693: console, create the domain while attaching the console, shutdown the
694: domain, and see if it has finished stopping, do (or xl with Xen >=
697: xm create foo
698: xm console foo
699: xm create -c foo
700: xm shutdown foo
701: xm list
703: Typing ^] will exit the console session. Shutting down a domain is
704: equivalent to pushing the power button; a NetBSD domU will receive a
705: power-press event and do a clean shutdown. Shutting down the dom0
706: will trigger controlled shutdowns of all configured domUs.
708: domU kernels
711: On a physical computer, the BIOS reads sector 0, and a chain of boot
712: loaders finds and loads a kernel. Normally this comes from the root
713: file system. With Xen domUs, the process is totally different. The
714: normal path is for the domU kernel to be a file in the dom0's
715: file system. At the request of the dom0, Xen loads that kernel into a
716: new domU instance and starts execution. While domU kernels can be
717: anyplace, reasonable places to store domU kernels on the dom0 are in /
718: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
719: config files), or in /u0/xen (where the vdisks are).
721: Note that loading the domU kernel from the dom0 implies that boot
722: blocks, /boot, /boot.cfg, and so on are all ignored in the domU.
723: See the VPS section near the end for discussion of alternate ways to
724: obtain domU kernels.
726: CPU and memory
729: A domain is provided with some number of vcpus, less than the number
730: of CPUs seen by the hypervisor. (For a dom0, this is controlled by
731: the boot argument "dom0_max_vcpus=1".) For a domU, it is controlled
732: from the config file by the "vcpus = N" directive.
734: A domain is provided with memory; this is controlled in the config
735: file by "memory = N" (in megabytes). In the straightforward case, the
736: sum of the the memory allocated to the dom0 and all domUs must be less
737: than the available memory.
739: Xen also provides a "balloon" driver, which can be used to let domains
740: use more memory temporarily. TODO: Explain better, and explain how
741: well it works with NetBSD.
743: Virtual disks
746: With the file/vnd style, typically one creates a directory,
747: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
748: domUs. Then, for each domU disk, one writes zeros to a file that then
749: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
750: for the first virtual disk for the domU called foo. Writing zeros to
751: the file serves two purposes. One is that preallocating the contents
752: improves performance. The other is that vnd on sparse files has
753: failed to work. TODO: give working/notworking NetBSD versions for
754: sparse vnd and gnats reference. Note that the use of file/vnd for Xen
755: is not really different than creating a file-backed virtual disk for
756: some other purpose, except that xentools handles the vnconfig
757: commands. To create an empty 4G virtual disk, simply do
759: dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
761: Do not use qemu-img-xen, because this will create sparse file. There
762: have been recent (2015) reports of sparse vnd(4) devices causing
763: lockups, but there is apparently no PR.
765: With the lvm style, one creates logical devices. They are then used
766: similarly to vnds. TODO: Add an example with lvm.
768: In domU config files, the disks are defined as a sequence of 3-tuples.
769: The first element is "method:/path/to/disk". Common methods are
770: "file:" for file-backed vnd. and "phy:" for something that is already
771: a (TODO: character or block) device.
773: The second element is an artifact of how virtual disks are passed to
774: Linux, and a source of confusion with NetBSD Xen usage. Linux domUs
775: are given a device name to associate with the disk, and values like
776: "hda1" or "sda1" are common. In a NetBSD domU, the first disk appears
777: as xbd0, the second as xbd1, and so on. However, xm/xl demand a
778: second argument. The name given is converted to a major/minor by
779: calling stat(2) on the name in /dev and this is passed to the domU.
780: In the general case, the dom0 and domU can be different operating
781: systems, and it is an unwarranted assumption that they have consistent
782: numbering in /dev, or even that the dom0 OS has a /dev. With NetBSD
783: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
784: for the second works fine and avoids this issue. For a GNU/Linux
785: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
788: The third element is "w" for writable disks, and "r" for read-only
791: Note that NetBSD by default creates only vnd. If you need more
792: than 4 total virtual disks at a time, run e.g. "./MAKEDEV vnd4" in the
795: Note that NetBSD by default creates only xbd. If you need more
796: virtual disks in a domU, run e.g. "./MAKEDEV xbd4" in the domU.
798: Virtual Networking
801: Xen provides virtual Ethernets, each of which connects the dom0 and a
802: domU. For each virtual network, there is an interface "xvifN.M" in
803: the dom0, and in domU index N, a matching interface xennetM (NetBSD
804: name). The interfaces behave as if there is an Ethernet with two
805: adapters connected. From this primitive, one can construct various
806: configurations. We focus on two common and useful cases for which
807: there are existing scripts: bridging and NAT.
809: With bridging (in the example above), the domU perceives itself to be
810: on the same network as the dom0. For server virtualization, this is
811: usually best. Bridging is accomplished by creating a bridge(4) device
812: and adding the dom0's physical interface and the various xvifN.0
813: interfaces to the bridge. One specifies "bridge=bridge0" in the domU
814: config file. The bridge must be set up already in the dom0; an
815: example /etc/ifconfig.bridge0 is:
819: !brconfig bridge0 add wm0
821: With NAT, the domU perceives itself to be behind a NAT running on the
822: dom0. This is often appropriate when running Xen on a workstation.
823: TODO: NAT appears to be configured by "vif = [ '' ]".
825: The MAC address specified is the one used for the interface in the new
826: domain. The interface in dom0 will use this address XOR'd with
827: 00:00:00:01:00:00. Random MAC addresses are assigned if not given.
829: Sizing domains
832: Modern x86 hardware has vast amounts of resources. However, many
833: virtual servers can function just fine on far less. A system with
834: 256M of RAM and a 4G disk can be a reasonable choice. Note that it is
835: far easier to adjust virtual resources than physical ones. For
836: memory, it's just a config file edit and a reboot. For disk, one can
837: create a new file and vnconfig it (or lvm), and then dump/restore,
838: just like updating physical disks, but without having to be there and
839: without those pesky connectors.
841: Starting domains automatically
844: To start domains foo at bar at boot and shut them down cleanly on dom0
845: shutdown, in rc.conf add:
847: xendomains="foo bar"
849: Note that earlier versions of the xentools41 xendomains rc.d script
850: used xl, when one should use xm with 4.1.
852: Creating specific unprivileged domains (domU)
855: Creating domUs is almost entirely independent of operating system. We
856: have already presented the basics of config files. Note that you must
857: have already completed the dom0 setup so that "xl list" (or "xm list")
860: Creating an unprivileged NetBSD domain (domU)
863: See the earlier config file, and adjust memory. Decide on how much
864: storage you will provide, and prepare it (file or lvm).
866: While the kernel will be obtained from the dom0 file system, the same
867: file should be present in the domU as /netbsd so that tools like
868: savecore(8) can work. (This is helpful but not necessary.)
870: The kernel must be specifically for Xen and for use as a domU. The
871: i386 and amd64 provide the following kernels:
873: i386 XEN3_DOMU
874: i386 XEN3PAE_DOMU
875: amd64 XEN3_DOMU
877: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
878: use the PAE version of the i386 kernel.
880: This will boot NetBSD, but this is not that useful if the disk is
881: empty. One approach is to unpack sets onto the disk outside of xen
882: (by mounting it, just as you would prepare a physical disk for a
883: system you can't run the installer on).
885: A second approach is to run an INSTALL kernel, which has a miniroot
886: and can load sets from the network. To do this, copy the INSTALL
887: kernel to / and change the kernel line in the config file to:
889: kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
891: Then, start the domain as "xl create -c configname".
893: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
894: line should be used in the config file.
896: disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
898: After booting the domain, the option to install via CDROM may be
899: selected. The CDROM device should be changed to `xbd1d`.
901: Once done installing, "halt -p" the new domain (don't reboot or halt,
902: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
903: config file), switch the config file back to the XEN3_DOMU kernel,
904: and start the new domain again. Now it should be able to use "root on
905: xbd0a" and you should have a, functional NetBSD domU.
907: TODO: check if this is still accurate.
908: When the new domain is booting you'll see some warnings about *wscons*
909: and the pseudo-terminals. These can be fixed by editing the files
910: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
911: `/etc/ttys`, except *console*, like this:
913: console "/usr/libexec/getty Pc" vt100 on secure
914: ttyE0 "/usr/libexec/getty Pc" vt220 off secure
915: ttyE1 "/usr/libexec/getty Pc" vt220 off secure
916: ttyE2 "/usr/libexec/getty Pc" vt220 off secure
917: ttyE3 "/usr/libexec/getty Pc" vt220 off secure
919: Finally, all screens must be commented out from `/etc/wscons.conf`.
921: It is also desirable to add
925: in rc.conf. This way, the domain will be properly shut down if
926: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
928: It is not strictly necessary to have a kernel (as /netbsd) in the domU
929: file system. However, various programs (e.g. netstat) will use that
930: kernel to look up symbols to read from kernel virtual memory. If
931: /netbsd is not the running kernel, those lookups will fail. (This is
932: not really a Xen-specific issue, but because the domU kernel is
933: obtained from the dom0, it is far more likely to be out of sync or
934: missing with Xen.)
936: Creating an unprivileged Linux domain (domU)
939: Creating unprivileged Linux domains isn't much different from
940: unprivileged NetBSD domains, but there are some details to know.
942: First, the second parameter passed to the disk declaration (the '0x1' in
943: the example below)
945: disk = [ 'phy:/dev/wd0e,0x1,w' ]
947: does matter to Linux. It wants a Linux device number here (e.g. 0x300
948: for hda). Linux builds device numbers as: (major \<\< 8 + minor).
949: So, hda1 which has major 3 and minor 1 on a Linux system will have
950: device number 0x301. Alternatively, devices names can be used (hda,
951: hdb, ...) as xentools has a table to map these names to devices
952: numbers. To export a partition to a Linux guest we can use:
954: disk = [ 'phy:/dev/wd0e,0x300,w' ]
955: root = "/dev/hda1 ro"
957: and it will appear as /dev/hda on the Linux system, and be used as root
960: To install the Linux system on the partition to be exported to the
961: guest domain, the following method can be used: install
962: sysutils/e2fsprogs from pkgsrc. Use mke2fs to format the partition
963: that will be the root partition of your Linux domain, and mount it.
964: Then copy the files from a working Linux system, make adjustments in
965: `/etc` (fstab, network config). It should also be possible to extract
966: binary packages such as .rpm or .deb directly to the mounted partition
967: using the appropriate tool, possibly running under NetBSD's Linux
968: emulation. Once the file system has been populated, umount it. If
969: desirable, the file system can be converted to ext3 using tune2fs -j.
970: It should now be possible to boot the Linux guest domain, using one of
971: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
973: To get the Linux console right, you need to add:
975: extra = "xencons=tty1"
977: to your configuration since not all Linux distributions auto-attach a
978: tty to the xen console.
980: Creating an unprivileged Solaris domain (domU)
983: See possibly outdated
984: [Solaris domU instructions](/ports/xen/howto-solaris/).
987: PCI passthrough: Using PCI devices in guest domains
990: The dom0 can give other domains access to selected PCI
991: devices. This can allow, for example, a non-privileged domain to have
992: access to a physical network interface or disk controller. However,
993: keep in mind that giving a domain access to a PCI device most likely
994: will give the domain read/write access to the whole physical memory,
995: as PCs don't have an IOMMU to restrict memory access to DMA-capable
996: device. Also, it's not possible to export ISA devices to non-dom0
997: domains, which means that the primary VGA adapter can't be exported.
998: A guest domain trying to access the VGA registers will panic.
1000: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1001: not been ported to later versions at this time.
1003: For a PCI device to be exported to a domU, is has to be attached to
1004: the "pciback" driver in dom0. Devices passed to the dom0 via the
1005: pciback.hide boot parameter will attach to "pciback" instead of the
1006: usual driver. The list of devices is specified as "(bus:dev.func)",
1007: where bus and dev are 2-digit hexadecimal numbers, and func a
1008: single-digit number:
1012: pciback devices should show up in the dom0's boot messages, and the
1013: devices should be listed in the `/kern/xen/pci` directory.
1015: PCI devices to be exported to a domU are listed in the "pci" array of
1016: the domU's config file, with the format "0000:bus:dev.func".
1018: pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1020: In the domU an "xpci" device will show up, to which one or more pci
1021: buses will attach. Then the PCI drivers will attach to PCI buses as
1022: usual. Note that the default NetBSD DOMU kernels do not have "xpci"
1023: or any PCI drivers built in by default; you have to build your own
1024: kernel to use PCI devices in a domU. Here's a kernel config example;
1025: note that only the "xpci" lines are unusual.
1027: include "arch/i386/conf/XEN3_DOMU"
1029: # Add support for PCI buses to the XEN3_DOMU kernel
1030: xpci* at xenbus ?
1031: pci* at xpci ?
1033: # PCI USB controllers
1034: uhci* at pci? dev ? function ? # Universal Host Controller (Intel)
1036: # USB bus support
1037: usb* at uhci?
1039: # USB Hubs
1040: uhub* at usb?
1041: uhub* at uhub? port ? configuration ? interface ?
1043: # USB Mass Storage
1044: umass* at uhub? port ? configuration ? interface ?
1045: wd* at umass?
1046: # SCSI controllers
1047: ahc* at pci? dev ? function ? # Adaptec 94x, aic78x0 SCSI
1049: # SCSI bus support (for both ahc and umass)
1050: scsibus* at scsi?
1052: # SCSI devices
1053: sd* at scsibus? target ? lun ? # SCSI disk drives
1054: cd* at scsibus? target ? lun ? # SCSI CD-ROM drives
1057: NetBSD as a domU in a VPS
1060: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
1061: hardware. This section explains how to deal with Xen in a domU as a
1062: virtual private server where you do not control or have access to the
1063: dom0. This is not intended to be an exhaustive list of VPS providers;
1064: only a few are mentioned that specifically support NetBSD.
1066: VPS operators provide varying degrees of access and mechanisms for
1067: configuration. The big issue is usually how one controls which kernel
1068: is booted, because the kernel is nominally in the dom0 file system (to
1069: which VPS users do not normally have access). A second issue is how
1070: to install NetBSD.
1071: A VPS user may want to compile a kernel for security updates, to run
1072: npf, run IPsec, or any other reason why someone would want to change
1073: their kernel.
1075: One approach is to have an administrative interface to upload a kernel,
1076: or to select from a prepopulated list. Other approaches are pygrub
1077: (deprecated) and pvgrub, which are ways to have a bootloader obtain a
1078: kernel from the domU file system. This is closer to a regular physical
1079: computer, where someone who controls a machine can replace the kernel.
1081: A second issue is multiple CPUs. With NetBSD 6, domUs support
1082: multiple vcpus, and it is typical for VPS providers to enable multiple
1083: CPUs for NetBSD domUs.
1088: pygrub runs in the dom0 and looks into the domU file system. This
1089: implies that the domU must have a kernel in a file system in a format
1090: known to pygrub. As of 2014, pygrub seems to be of mostly historical
1096: pvgrub is a version of grub that uses PV operations instead of BIOS
1097: calls. It is booted from the dom0 as the domU kernel, and then reads
1098: /grub/menu.lst and loads a kernel from the domU file system.
1100: [Panix](http://www.panix.com/) lets users use pvgrub. Panix reports
1101: that pvgrub works with FFsv2 with 16K/2K and 32K/4K block/frag sizes
1102: (and hence with defaults from "newfs -O 2"). See [Panix's pvgrub
1103: page](http://www.panix.com/v-colo/grub.html), which describes only
1104: Linux but should be updated to cover NetBSD :-).
1106: [prgmr.com](http://prgmr.com/) also lets users with pvgrub to boot
1107: their own kernel. See then [prgmr.com NetBSD
1109: (which is in need of updating).
1111: It appears that [grub's FFS
1113: does not support all aspects of modern FFS, but there are also reports
1114: that FFSv2 works fine. At prgmr, typically one has an ext2 or FAT
1115: partition for the kernel with the intent that grub can understand it,
1116: which leads to /netbsd not being the actual kernel. One must remember
1117: to update the special boot partition.
1122: See the [Amazon EC2 page](../amazon_ec2/).
1124: Using npf
1127: In standard kernels, npf is a module, and thus cannot be loaded in a
1128: DOMU kernel.
1130: TODO: Explain how to compile npf into a custom kernel, answering (but
1131: note that the problem was caused by not booting the right kernel)
1132: [this email to
1135: TODO items for improving NetBSD/xen
1138: * Make the NetBSD dom0 kernel work with SMP.
1139: * Test the Xen 4.5 packages adequately to be able to recommend them as
1140: the standard approach.
1141: * Get PCI passthrough working on Xen 4.5
1142: * Get pvgrub into pkgsrc, either via xentools or separately.
1143: * grub
1144: * Check/add support to pkgsrc grub2 for UFS2 and arbitrary
1145: fragsize/blocksize (UFS2 support may be present; the point is to
1146: make it so that with any UFS1/UFS2 file system setup that works
1147: with NetBSD grub will also work).
1148: See [pkg/40258](http://gnats.netbsd.org/40258).
1149: * Push patches upstream.
1150: * Get UFS2 patches into pvgrub.
1151: * Add support for PV ops to a version of /boot, and make it usable as
1152: a kernel in Xen, similar to pvgrub.
1153: * Solve somehow the issue with modules for GENERIC not being loadable
1154: in a Xen dom0 or domU kernel.
1156: Random pointers
1159: This section contains links from elsewhere not yet integrated into the
1160: HOWTO, and other guides.
1162: * http://www.lumbercartel.ca/library/xen/
1163: * http://pbraun.nethence.com/doc/sysutils/xen_netbsd_dom0.html
1164: * https://gmplib.org/~tege/xen.html
CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb