Annotation of wikisrc/ports/xen/howto.mdwn, revision 1.58
1.5 mspo 1: Introduction
1.13 gdt 2: ============
1.1 mspo 3:
4: [![[Xen
1.57 gdt 5: screenshot]](http://www.netbsd.org/gallery/in-Action/hubertf-xens.png)](http://www.netbsd.org/gallery/in-Action/hubertf-xen.png)
1.1 mspo 6:
1.58 ! gdt 7: Xen is a hypervisor (or virtual machine monitor) for x86 hardware
1.12 gdt 8: (i686-class or higher), which supports running multiple guest
1.58 ! gdt 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.
1.12 gdt 16:
17: Xen supports two styles of guests. The original is Para-Virtualized
18: (PV) which means that the guest OS does not attempt to access hardware
19: directly, but instead makes hypercalls to the hypervisor. This is
20: analogous to a user-space program making system calls. (The dom0
21: operating system uses PV calls for some functions, such as updating
22: memory mapping page tables, but has direct hardware access for disk
23: and network.) PV guests must be specifically coded for Xen.
24:
25: The more recent style is HVM, which means that the guest does not have
26: code for Xen and need not be aware that it is running under Xen.
27: Attempts to access hardware registers are trapped and emulated. This
28: style is less efficient but can run unmodified guests.
29:
1.29 gdt 30: Generally any amd64 machine will work with Xen and PV guests. In
31: theory i386 computers without amd64 support can be used for Xen <=
32: 4.2, but we have no recent reports of this working (this is a hint).
33: For HVM guests, the VT or VMX cpu feature (Intel) or SVM/HVM/VT
34: (amd64) is needed; "cpuctl identify 0" will show this. TODO: Clean up
35: and check the above features.
1.19 gdt 36:
1.27 jnemeth 37: At boot, the dom0 kernel is loaded as a module with Xen as the kernel.
1.12 gdt 38: The dom0 can start one or more domUs. (Booting is explained in detail
39: in the dom0 section.)
40:
41: NetBSD supports Xen in that it can serve as dom0, be used as a domU,
42: and that Xen kernels and tools are available in pkgsrc. This HOWTO
43: attempts to address both the case of running a NetBSD dom0 on hardware
1.24 gdt 44: and running domUs under it (NetBSD and other), and also running NetBSD
45: as a domU in a VPS.
1.12 gdt 46:
1.20 gdt 47: Some versions of Xen support "PCI passthrough", which means that
48: specific PCI devices can be made available to a specific domU instead
49: of the dom0. This can be useful to let a domU run X11, or access some
50: network interface or other peripheral.
51:
1.54 gdt 52: NetBSD used to support Xen2; this has been removed.
53:
1.12 gdt 54: Prerequisites
1.13 gdt 55: -------------
1.12 gdt 56:
57: Installing NetBSD/Xen is not extremely difficult, but it is more
58: complex than a normal installation of NetBSD.
1.15 gdt 59: In general, this HOWTO is occasionally overly restrictive about how
60: things must be done, guiding the reader to stay on the established
61: path when there are no known good reasons to stray.
1.12 gdt 62:
63: This HOWTO presumes a basic familiarity with the Xen system
1.16 gdt 64: architecture. This HOWTO presumes familiarity with installing NetBSD
65: on i386/amd64 hardware and installing software from pkgsrc.
1.27 jnemeth 66: See also the [Xen website](http://www.xenproject.org/).
1.1 mspo 67:
1.15 gdt 68: Versions of Xen and NetBSD
69: ==========================
70:
1.27 jnemeth 71: Most of the installation concepts and instructions are independent
72: of Xen version and NetBSD version. This section gives advice on
73: which version to choose. Versions not in pkgsrc and older unsupported
74: versions of NetBSD are intentionally ignored.
1.15 gdt 75:
76: Xen
77: ---
78:
79: In NetBSD, xen is provided in pkgsrc, via matching pairs of packages
80: xenkernel and xentools. We will refer only to the kernel versions,
81: but note that both packages must be installed together and must have
82: matching versions.
83:
84: xenkernel3 and xenkernel33 provide Xen 3.1 and 3.3. These no longer
1.20 gdt 85: receive security patches and should not be used. Xen 3.1 supports PCI
1.29 gdt 86: passthrough. Xen 3.1 supports non-PAE on i386.
1.15 gdt 87:
88: xenkernel41 provides Xen 4.1. This is no longer maintained by Xen,
89: but as of 2014-12 receives backported security patches. It is a
90: reasonable although trailing-edge choice.
91:
92: xenkernel42 provides Xen 4.2. This is maintained by Xen, but old as
93: of 2014-12.
94:
95: Ideally newer versions of Xen will be added to pkgsrc.
96:
1.26 gdt 97: Note that NetBSD support is called XEN3. It works with 3.1 through
98: 4.2 because the hypercall interface has been stable.
1.20 gdt 99:
1.19 gdt 100: Xen command program
101: -------------------
102:
103: Early Xen used a program called "xm" to manipulate the system from the
104: dom0. Starting in 4.1, a replacement program with similar behavior
1.27 jnemeth 105: called "xl" is provided. In 4.2 and later, "xl" is preferred. 4.4 is
106: the last version that has "xm".
1.19 gdt 107:
1.15 gdt 108: NetBSD
109: ------
110:
111: The netbsd-5, netbsd-6, netbsd-7, and -current branches are all
112: reasonable choices, with more or less the same considerations for
113: non-Xen use. Therefore, netbsd-6 is recommended as the stable version
1.29 gdt 114: of the most recent release for production use. For those wanting to
115: learn Xen or without production stability concerns, netbsd-7 is likely
116: most appropriate.
1.15 gdt 117:
118: As of NetBSD 6, a NetBSD domU will support multiple vcpus. There is
119: no SMP support for NetBSD as dom0. (The dom0 itself doesn't really
120: need SMP; the lack of support is really a problem when using a dom0 as
121: a normal computer.)
122:
1.18 gdt 123: Architecture
124: ------------
125:
1.29 gdt 126: Xen itself can run on i386 or amd64 machines. (Practically, almost
127: any computer where one would want to run Xen supports amd64.) If
128: using an i386 NetBSD kernel for the dom0, PAE is required (PAE
129: versions are built by default). While i386 dom0 works fine, amd64 is
130: recommended as more normal.
131:
132: Xen 4.2 is the last version to support i386 as a host. TODO: Clarify
133: if this is about the CPU having to be amd64, or about the dom0 kernel
134: having to be amd64.
135:
136: One can then run i386 domUs and amd64 domUs, in any combination. If
137: running an i386 NetBSD kernel as a domU, the PAE version is required.
138: (Note that emacs (at least) fails if run on i386 with PAE when built
139: without, and vice versa, presumably due to bugs in the undump code.)
1.18 gdt 140:
1.15 gdt 141: Recommendation
142: --------------
143:
1.18 gdt 144: Therefore, this HOWTO recommends running xenkernel42 (and xentools42),
1.30 gdt 145: xl, the NetBSD 6 stable branch, and to use an amd64 kernel as the
146: dom0. Either the i386 or amd64 of NetBSD may be used as domUs.
1.15 gdt 147:
1.36 gdt 148: Build problems
149: --------------
150:
151: Ideally, all versions of Xen in pkgsrc would build on all versions of
152: NetBSD on both i386 and amd64. However, that isn't the case. Besides
153: aging code and aging compilers, qemu (included in xentools for HVM
154: support) is difficult to build. The following are known to fail:
155:
156: xenkernel3 netbsd-6 i386
157: xentools42 netbsd-6 i386
158:
159: The following are known to work:
160:
161: xenkernel41 netbsd-5 amd64
162: xentools41 netbsd-5 amd64
163: xenkernel41 netbsd-6 i386
164: xentools41 netbsd-6 i386
165:
1.15 gdt 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.
1.19 gdt 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.
1.15 gdt 176:
1.45 gdt 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.
180:
1.15 gdt 181: Styles of dom0 operation
182: ------------------------
183:
184: There are two basic ways to use Xen. The traditional method is for
185: the dom0 to do absolutely nothing other than providing support to some
186: number of domUs. Such a system was probably installed for the sole
187: purpose of hosting domUs, and sits in a server room on a UPS.
188:
189: The other way is to put Xen under a normal-usage computer, so that the
190: dom0 is what the computer would have been without Xen, perhaps a
191: desktop or laptop. Then, one can run domUs at will. Purists will
192: deride this as less secure than the previous approach, and for a
193: computer whose purpose is to run domUs, they are right. But Xen and a
194: dom0 (without domUs) is not meaingfully less secure than the same
195: things running without Xen. One can boot Xen or boot regular NetBSD
196: alternately with little problems, simply refraining from starting the
197: Xen daemons when not running Xen.
198:
199: Note that NetBSD as dom0 does not support multiple CPUs. This will
1.51 gdt 200: limit the performance of the Xen/dom0 workstation approach. In theory
201: the only issue is that the "backend drivers" are not yet MPSAFE:
202: http://mail-index.netbsd.org/netbsd-users/2014/08/29/msg015195.html
1.15 gdt 203:
1.19 gdt 204: Installation of NetBSD
205: ----------------------
1.13 gdt 206:
1.19 gdt 207: First,
1.27 jnemeth 208: [install NetBSD/amd64](/guide/inst/)
1.19 gdt 209: just as you would if you were not using Xen.
210: However, the partitioning approach is very important.
211:
212: If you want to use RAIDframe for the dom0, there are no special issues
213: for Xen. Typically one provides RAID storage for the dom0, and the
1.22 gdt 214: domU systems are unaware of RAID. The 2nd-stage loader bootxx_* skips
215: over a RAID1 header to find /boot from a filesystem within a RAID
216: partition; this is no different when booting Xen.
1.19 gdt 217:
218: There are 4 styles of providing backing storage for the virtual disks
219: used by domUs: raw partitions, LVM, file-backed vnd(4), and SAN,
220:
221: With raw partitions, one has a disklabel (or gpt) partition sized for
222: each virtual disk to be used by the domU. (If you are able to predict
223: how domU usage will evolve, please add an explanation to the HOWTO.
224: Seriously, needs tend to change over time.)
225:
1.27 jnemeth 226: One can use [lvm(8)](/guide/lvm/) to create logical devices to use
227: for domU disks. This is almost as efficient as raw disk partitions
228: and more flexible. Hence raw disk partitions should typically not
229: be used.
1.19 gdt 230:
231: One can use files in the dom0 filesystem, typically created by dd'ing
232: /dev/zero to create a specific size. This is somewhat less efficient,
233: but very convenient, as one can cp the files for backup, or move them
234: between dom0 hosts.
235:
236: Finally, in theory one can place the files backing the domU disks in a
237: SAN. (This is an invitation for someone who has done this to add a
238: HOWTO page.)
1.1 mspo 239:
1.19 gdt 240: Installation of Xen
241: -------------------
1.1 mspo 242:
1.20 gdt 243: In the dom0, install sysutils/xenkernel42 and sysutils/xentools42 from
244: pkgsrc (or another matching pair).
245: See [the pkgsrc
246: documentation](http://www.NetBSD.org/docs/pkgsrc/) for help with pkgsrc.
247:
248: For Xen 3.1, support for HVM guests is in sysutils/xentool3-hvm. More
249: recent versions have HVM support integrated in the main xentools
250: package. It is entirely reasonable to run only PV guests.
251:
252: Next you need to install the selected Xen kernel itself, which is
253: installed by pkgsrc as "/usr/pkg/xen*-kernel/xen.gz". Copy it to /.
254: For debugging, one may copy xen-debug.gz; this is conceptually similar
255: to DIAGNOSTIC and DEBUG in NetBSD. xen-debug.gz is basically only
256: useful with a serial console. Then, place a NetBSD XEN3_DOM0 kernel
257: in /, copied from releasedir/amd64/binary/kernel/netbsd-XEN3_DOM0.gz
258: of a NetBSD build. Both xen and NetBSD may be left compressed. (If
259: using i386, use releasedir/i386/binary/kernel/netbsd-XEN3PAE_DOM0.gz.)
260:
1.49 gdt 261: With Xen as the kernel, you must provide a dom0 NetBSD kernel to be
262: used as a module; place this in /. Suitable kernels are provided in
263: releasedir/binary/kernel:
264:
265: i386 XEN3_DOM0
266: i386 XEN3PAE_DOM0
267: amd64 XEN3_DOM0
268:
269: The first one is only for use with Xen 3.1 and i386-mode Xen (and you
270: should not do this). Current Xen always uses PAE on i386, but you
271: should generally use amd64 for the dom0. In a dom0 kernel, kernfs is
272: mandatory for xend to comunicate with the kernel, so ensure that /kern
273: is in fstab. TODO: Say this is default, or file a PR and give a
274: reference.
1.20 gdt 275:
276: Because you already installed NetBSD, you have a working boot setup
277: with an MBR bootblock, either bootxx_ffsv1 or bootxx_ffsv2 at the
278: beginning of your root filesystem, /boot present, and likely
279: /boot.cfg. (If not, fix before continuing!)
280:
281: See boot.cfg(5) for an example. The basic line is
282:
1.37 gdt 283: menu=Xen:load /netbsd-XEN3_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
1.20 gdt 284:
285: which specifies that the dom0 should have 256M, leaving the rest to be
1.37 gdt 286: allocated for domUs. In an attempt to add performance, one can also
287: add
288:
289: dom0_max_vcpus=1 dom0_vcpus_pin
290:
291: to force only one vcpu to be provided (since NetBSD dom0 can't use
292: more) and to pin that vcpu to a physical cpu. TODO: benchmark this.
1.20 gdt 293:
294: As with non-Xen systems, you should have a line to boot /netbsd (a
295: kernel that works without Xen) and fallback versions of the non-Xen
296: kernel, Xen, and the dom0 kernel.
1.1 mspo 297:
1.54 gdt 298: Using grub (historic)
299: ---------------------
300:
301: Before NetBSD's native bootloader could support Xen, the use of
302: grub was recommended. If necessary, see the
303: [old grub information](/ports/xen/howto-grub/).
304:
1.28 gdt 305: The [HowTo on Installing into
306: RAID-1](http://mail-index.NetBSD.org/port-xen/2006/03/01/0010.html)
307: explains how to set up booting a dom0 with Xen using grub with
308: NetBSD's RAIDframe. (This is obsolete with the use of NetBSD's native
309: boot.)
310:
1.21 gdt 311: Configuring Xen
312: ---------------
313:
1.53 gdt 314: Xen logs will be in /var/log/xen.
315:
1.21 gdt 316: Now, you have a system that will boot Xen and the dom0 kernel, and
317: just run the dom0 kernel. There will be no domUs, and none can be
1.31 gdt 318: started because you still have to configure the dom0 tools. The
319: daemons which should be run vary with Xen version and with whether one
320: is using xm or xl. Note that xend is for supporting "xm", and should
321: only be used if you plan on using "xm". Do NOT enable xend if you
322: plan on using "xl" as it will cause problems.
1.21 gdt 323:
1.43 gdt 324: The installation of NetBSD should already have created devices for xen
325: (xencons, xenevt), but if they are not present, create them:
326:
327: cd /dev && sh MAKEDEV xen
328:
1.31 gdt 329: TODO: Give 3.1 advice (or remove it from pkgsrc).
330:
331: For 3.3 (and thus xm), add to rc.conf (but note that you should have
332: installed 4.1 or 4.2):
333:
1.32 gdt 334: xend=YES
335: xenbackendd=YES
1.31 gdt 336:
1.33 gdt 337: For 4.1 (and thus xm; xl is believed not to work well), add to rc.conf:
1.31 gdt 338:
1.53 gdt 339: xencommons=YES
1.31 gdt 340: xend=YES
341:
1.56 gdt 342: (If you are using xentools41 from before 2014-12-26, change
343: rc.d/xendomains to use xm rather than xl.)
1.31 gdt 344:
1.33 gdt 345: For 4.2 with xm, add to rc.conf
346:
1.53 gdt 347: xencommons=YES
1.33 gdt 348: xend=YES
349:
350: For 4.2 with xl (preferred), add to rc.conf:
1.31 gdt 351:
1.53 gdt 352: xencommons=YES
1.31 gdt 353: TODO: explain if there is a xend replacement
354:
355: TODO: Recommend for/against xen-watchdog.
1.27 jnemeth 356:
1.53 gdt 357: After you have configured the daemons and either started them (in the
358: order given) or rebooted, run the following (or use xl) to inspect
359: Xen's boot messages, available resources, and running domains:
1.34 gdt 360:
1.43 gdt 361: # xm dmesg
362: [xen's boot info]
363: # xm info
364: [available memory, etc.]
365: # xm list
366: Name Id Mem(MB) CPU State Time(s) Console
367: Domain-0 0 64 0 r---- 58.1
1.33 gdt 368:
1.41 gdt 369: anita (for testing NetBSD)
370: --------------------------
371:
372: With the setup so far, one should be able to run anita (see
373: pkgsrc/sysutils/py-anita) to test NetBSD releases, by doing (as root,
374: because anita must create a domU):
375:
376: anita --vmm=xm test file:///usr/obj/i386/
377:
378: Alternatively, one can use --vmm=xl to use xl-based domU creation instead.
379: TODO: check this.
380:
1.40 gdt 381: Xen-specific NetBSD issues
382: --------------------------
383:
384: There are (at least) two additional things different about NetBSD as a
385: dom0 kernel compared to hardware.
386:
387: One is that modules are not usable in DOM0 kernels, so one must
388: compile in what's needed. It's not really that modules cannot work,
389: but that modules must be built for XEN3_DOM0 because some of the
390: defines change and the normal module builds don't do this. Basically,
391: enabling Xen changes the kernel ABI, and the module build system
392: doesn't cope with this.
393:
394: The other difference is that XEN3_DOM0 does not have exactly the same
395: options as GENERIC. While it is debatable whether or not this is a
396: bug, users should be aware of this and can simply add missing config
397: items if desired.
398:
1.15 gdt 399: Updating NetBSD in a dom0
400: -------------------------
401:
402: This is just like updating NetBSD on bare hardware, assuming the new
403: version supports the version of Xen you are running. Generally, one
404: replaces the kernel and reboots, and then overlays userland binaries
405: and adjusts /etc.
406:
407: Note that one must update both the non-Xen kernel typically used for
408: rescue purposes and the DOM0 kernel used with Xen.
409:
1.55 gdt 410: Converting from grub to /boot
411: -----------------------------
412:
413: These instructions were [TODO: will be] used to convert a system from
414: grub to /boot. The system was originally installed in February of
415: 2006 with a RAID1 setup and grub to boot Xen 2, and has been updated
416: over time. Before these commands, it was running NetBSD 6 i386, Xen
417: 4.1 and grub, much like the message linked earlier in the grub
418: section.
419:
420: # Install mbr bootblocks on both disks.
421: fdisk -i /dev/rwd0d
422: fdisk -i /dev/rwd1d
423: # Install NetBSD primary boot loader (/ is FFSv1) into RAID1 components.
424: installboot -v /dev/rwd0d /usr/mdec/bootxx_ffsv1
425: installboot -v /dev/rwd1d /usr/mdec/bootxx_ffsv1
426: # Install secondary boot loader
427: cp -p /usr/mdec/boot /
428: # Create boog.cfg following earlier guidance:
429: menu=Xen:load /netbsd-XEN3PAE_DOM0.gz console=pc;multiboot /xen.gz dom0_mem=256M
430: menu=Xen.ok:load /netbsd-XEN3PAE_DOM0.ok.gz console=pc;multiboot /xen.ok.gz dom0_mem=256M
431: menu=GENERIC:boot
432: menu=GENERIC single-user:boot -s
433: menu=GENERIC.ok:boot netbsd.ok
434: menu=GENERIC.ok single-user:boot netbsd.ok -s
435: menu=Drop to boot prompt:prompt
436: default=1
437: timeout=30
438:
439: TODO: actually do this and fix it if necessary.
1.22 gdt 440:
1.15 gdt 441: Updating Xen versions
442: ---------------------
443:
1.21 gdt 444: Updating Xen is conceptually not difficult, but can run into all the
445: issues found when installing Xen. Assuming migration from 4.1 to 4.2,
446: remove the xenkernel41 and xentools41 packages and install the
447: xenkernel42 and xentools42 packages. Copy the 4.2 xen.gz to /.
448:
449: Ensure that the contents of /etc/rc.d/xen* are correct. Enable the
450: correct set of daemons. Ensure that the domU config files are valid
451: for the new version.
1.15 gdt 452:
1.28 gdt 453:
454: Unprivileged domains (domU)
455: ===========================
456:
457: This section describes general concepts about domUs. It does not
1.33 gdt 458: address specific domU operating systems or how to install them. The
459: config files for domUs are typically in /usr/pkg/etc/xen, and are
460: typically named so that the file anme, domU name and the domU's host
461: name match.
462:
463: The domU is provided with cpu and memory by Xen, configured by the
464: dom0. The domU is provided with disk and network by the dom0,
465: mediated by Xen, and configured in the dom0.
466:
467: Entropy in domUs can be an issue; physical disks and network are on
468: the dom0. NetBSD's /dev/random system works, but is often challenged.
469:
1.48 gdt 470: Config files
471: ------------
472:
473: There is no good order to present config files and the concepts
474: surrounding what is being configured. We first show an example config
475: file, and then in the various sections give details.
476:
477: See (at least in xentools41) /usr/pkg/share/examples/xen/xmexample*,
478: for a large number of well-commented examples, mostly for running
479: GNU/Linux.
480:
481: The following is an example minimal domain configuration file
482: "/usr/pkg/etc/xen/foo". It is (with only a name change) an actual
483: known working config file on Xen 4.1 (NetBSD 5 amd64 dom0 and NetBSD 5
484: i386 domU). The domU serves as a network file server.
485:
486: # -*- mode: python; -*-
487:
488: kernel = "/netbsd-XEN3PAE_DOMU-i386-foo.gz"
489: memory = 1024
490: vif = [ 'mac=aa:00:00:d1:00:09,bridge=bridge0' ]
491: disk = [ 'file:/n0/xen/foo-wd0,0x0,w',
492: 'file:/n0/xen/foo-wd1,0x1,w' ]
493:
494: The domain will have the same name as the file. The kernel has the
495: host/domU name in it, so that on the dom0 one can update the various
496: domUs independently. The vif line causes an interface to be provided,
497: with a specific mac address (do not reuse MAC addresses!), in bridge
498: mode. Two disks are provided, and they are both writable; the bits
499: are stored in files and Xen attaches them to a vnd(4) device in the
500: dom0 on domain creation. The system treates xbd0 as the boot device
501: without needing explicit configuration.
502:
503: By default xm looks for domain config files in /usr/pkg/etc/xen. Note
504: that "xm create" takes the name of a config file, while other commands
505: take the name of a domain. To create the domain, connect to the
506: console, create the domain while attaching the console, shutdown the
507: domain, and see if it has finished stopping, do (or xl with Xen >=
508: 4.2):
509:
510: xm create foo
511: xm console foo
512: xm create -c foo
513: xm shutdown foo
514: xm list
515:
516: Typing ^] will exit the console session. Shutting down a domain is
517: equivalent to pushing the power button; a NetBSD domU will receive a
518: power-press event and do a clean shutdown. Shutting down the dom0
519: will trigger controlled shutdowns of all configured domUs.
520:
521: domU kernels
522: ------------
523:
524: On a physical computer, the BIOS reads sector 0, and a chain of boot
525: loaders finds and loads a kernel. Normally this comes from the root
526: filesystem. With Xen domUs, the process is totally different. The
527: normal path is for the domU kernel to be a file in the dom0's
528: filesystem. At the request of the dom0, Xen loads that kernel into a
529: new domU instance and starts execution. While domU kernels can be
530: anyplace, reasonable places to store domU kernels on the dom0 are in /
531: (so they are near the dom0 kernel), in /usr/pkg/etc/xen (near the
532: config files), or in /u0/xen (where the vdisks are).
533:
534: See the VPS section near the end for discussion of alternate ways to
535: obtain domU kernels.
536:
1.33 gdt 537: CPU and memory
538: --------------
539:
1.48 gdt 540: A domain is provided with some number of vcpus, less than the number
541: of cpus seen by the hypervisor. (For a dom0, this is controlled by
542: the boot argument "dom0_max_vcpus=1".) For a domU, it is controlled
543: from the config file by the "vcpus = N" directive.
544:
545: A domain is provided with memory; this is controlled in the config
546: file by "memory = N" (in megabytes). In the straightforward case, the
547: sum of the the memory allocated to the dom0 and all domUs must be less
1.33 gdt 548: than the available memory.
549:
550: Xen also provides a "balloon" driver, which can be used to let domains
551: use more memory temporarily. TODO: Explain better, and explain how
552: well it works with NetBSD.
1.28 gdt 553:
554: Virtual disks
555: -------------
556:
1.33 gdt 557: With the file/vnd style, typically one creates a directory,
558: e.g. /u0/xen, on a disk large enough to hold virtual disks for all
559: domUs. Then, for each domU disk, one writes zeros to a file that then
560: serves to hold the virtual disk's bits; a suggested name is foo-xbd0
561: for the first virtual disk for the domU called foo. Writing zeros to
562: the file serves two purposes. One is that preallocating the contents
563: improves performance. The other is that vnd on sparse files has
564: failed to work. TODO: give working/notworking NetBSD versions for
565: sparse vnd. Note that the use of file/vnd for Xen is not really
566: different than creating a file-backed virtual disk for some other
1.39 gdt 567: purpose, except that xentools handles the vnconfig commands. To
568: create an empty 4G virtual disk, simply do
569:
570: dd if=/dev/zero of=foo-xbd0 bs=1m count=4096
1.33 gdt 571:
572: With the lvm style, one creates logical devices. They are then used
1.48 gdt 573: similarly to vnds. TODO: Add an example with lvm.
574:
575: In domU config files, the disks are defined as a sequence of 3-tuples.
576: The first element is "method:/path/to/disk". Common methods are
577: "file:" for file-backed vnd. and "phy:" for something that is already
578: a (TODO: character or block) device.
579:
580: The second element is an artifact of how virtual disks are passed to
581: Linux, and a source of confusion with NetBSD Xen usage. Linux domUs
582: are given a device name to associate with the disk, and values like
583: "hda1" or "sda1" are common. In a NetBSD domU, the first disk appears
584: as xbd0, the second as xbd1, and so on. However, xm/xl demand a
585: second argument. The name given is converted to a major/minor by
1.49 gdt 586: calling stat(2) on the name in /dev and this is passed to the domU.
587: In the general case, the dom0 and domU can be different operating
1.48 gdt 588: systems, and it is an unwarranted assumption that they have consistent
589: numbering in /dev, or even that the dom0 OS has a /dev. With NetBSD
590: as both dom0 and domU, using values of 0x0 for the first disk and 0x1
1.49 gdt 591: for the second works fine and avoids this issue. For a GNU/Linux
592: guest, one can create /dev/hda1 in /dev, or to pass 0x301 for
593: /dev/hda1.
1.48 gdt 594:
595: The third element is "w" for writable disks, and "r" for read-only
596: disks.
1.28 gdt 597:
598: Virtual Networking
599: ------------------
600:
1.46 gdt 601: Xen provides virtual ethernets, each of which connects the dom0 and a
602: domU. For each virtual network, there is an interface "xvifN.M" in
603: the dom0, and in domU index N, a matching interface xennetM (NetBSD
604: name). The interfaces behave as if there is an Ethernet with two
605: adaptors connected. From this primitive, one can construct various
606: configurations. We focus on two common and useful cases for which
607: there are existing scripts: bridging and NAT.
1.28 gdt 608:
1.48 gdt 609: With bridging (in the example above), the domU perceives itself to be
610: on the same network as the dom0. For server virtualization, this is
611: usually best. Bridging is accomplished by creating a bridge(4) device
612: and adding the dom0's physical interface and the various xvifN.0
613: interfaces to the bridge. One specifies "bridge=bridge0" in the domU
614: config file. The bridge must be set up already in the dom0; an
615: example /etc/ifconfig.bridge0 is:
1.46 gdt 616:
617: create
618: up
619: !brconfig bridge0 add wm0
1.28 gdt 620:
621: With NAT, the domU perceives itself to be behind a NAT running on the
622: dom0. This is often appropriate when running Xen on a workstation.
1.48 gdt 623: TODO: NAT appears to be configured by "vif = [ '' ]".
1.28 gdt 624:
1.49 gdt 625: The MAC address specified is the one used for the interface in the new
1.53 gdt 626: domain. The interface in dom0 will use this address XOR'd with
1.49 gdt 627: 00:00:00:01:00:00. Random MAC addresses are assigned if not given.
628:
1.33 gdt 629: Sizing domains
630: --------------
631:
632: Modern x86 hardware has vast amounts of resources. However, many
633: virtual servers can function just fine on far less. A system with
634: 256M of RAM and a 4G disk can be a reasonable choice. Note that it is
635: far easier to adjust virtual resources than physical ones. For
636: memory, it's just a config file edit and a reboot. For disk, one can
637: create a new file and vnconfig it (or lvm), and then dump/restore,
638: just like updating physical disks, but without having to be there and
639: without those pesky connectors.
640:
1.48 gdt 641: Starting domains automatically
642: ------------------------------
1.28 gdt 643:
1.48 gdt 644: To start domains foo at bar at boot and shut them down cleanly on dom0
645: shutdown, in rc.conf add:
1.28 gdt 646:
1.48 gdt 647: xendomains="foo bar"
1.28 gdt 648:
649: TODO: Explain why 4.1 rc.d/xendomains has xl, when one should use xm
1.48 gdt 650: on 4.1. Or fix the xentools41 package to have xm
1.28 gdt 651:
652: Creating specific unprivileged domains (domU)
653: =============================================
1.14 gdt 654:
655: Creating domUs is almost entirely independent of operating system. We
1.49 gdt 656: have already presented the basics of config files. Note that you must
657: have already completed the dom0 setup so that "xl list" (or "xm list")
658: works.
1.14 gdt 659:
660: Creating an unprivileged NetBSD domain (domU)
661: ---------------------------------------------
1.1 mspo 662:
1.49 gdt 663: See the earlier config file, and adjust memory. Decide on how much
664: storage you will provide, and prepare it (file or lvm).
665:
666: While the kernel will be obtained from the dom0 filesystem, the same
667: file should be present in the domU as /netbsd so that tools like
668: savecore(8) can work. (This is helpful but not necessary.)
669:
670: The kernel must be specifically for Xen and for use as a domU. The
671: i386 and amd64 provide the following kernels:
672:
673: i386 XEN3_DOMU
674: i386 XEN3PAE_DOMU
675: amd64 XEN3_DOMU
1.5 mspo 676:
1.49 gdt 677: Unless using Xen 3.1 (and you shouldn't) with i386-mode Xen, you must
678: use the PAE version of the i386 kernel.
679:
680: This will boot NetBSD, but this is not that useful if the disk is
681: empty. One approach is to unpack sets onto the disk outside of xen
682: (by mounting it, just as you would prepare a physical disk for a
683: system you can't run the installer on).
684:
685: A second approach is to run an INSTALL kernel, which has a miniroot
686: and can load sets from the network. To do this, copy the INSTALL
687: kernel to / and change the kernel line in the config file to:
1.5 mspo 688:
1.49 gdt 689: kernel = "/home/bouyer/netbsd-INSTALL_XEN3_DOMU"
1.5 mspo 690:
1.49 gdt 691: Then, start the domain as "xl create -c configname".
1.1 mspo 692:
1.49 gdt 693: Alternatively, if you want to install NetBSD/Xen with a CDROM image, the following
694: line should be used in the config file.
1.1 mspo 695:
1.3 mspo 696: disk = [ 'phy:/dev/wd0e,0x1,w', 'phy:/dev/cd0a,0x2,r' ]
1.1 mspo 697:
698: After booting the domain, the option to install via CDROM may be
1.49 gdt 699: selected. The CDROM device should be changed to `xbd1d`.
1.1 mspo 700:
1.49 gdt 701: Once done installing, "halt -p" the new domain (don't reboot or halt,
702: it would reload the INSTALL_XEN3_DOMU kernel even if you changed the
703: config file), switch the config file back to the XEN3_DOMU kernel,
704: and start the new domain again. Now it should be able to use "root on
705: xbd0a" and you should have a, functional NetBSD domU.
1.1 mspo 706:
1.49 gdt 707: TODO: check if this is still accurate.
1.1 mspo 708: When the new domain is booting you'll see some warnings about *wscons*
709: and the pseudo-terminals. These can be fixed by editing the files
1.5 mspo 710: `/etc/ttys` and `/etc/wscons.conf`. You must disable all terminals in
711: `/etc/ttys`, except *console*, like this:
1.1 mspo 712:
1.3 mspo 713: console "/usr/libexec/getty Pc" vt100 on secure
714: ttyE0 "/usr/libexec/getty Pc" vt220 off secure
715: ttyE1 "/usr/libexec/getty Pc" vt220 off secure
716: ttyE2 "/usr/libexec/getty Pc" vt220 off secure
717: ttyE3 "/usr/libexec/getty Pc" vt220 off secure
1.1 mspo 718:
1.5 mspo 719: Finally, all screens must be commented out from `/etc/wscons.conf`.
1.1 mspo 720:
721: It is also desirable to add
722:
1.49 gdt 723: powerd=YES
1.1 mspo 724:
1.5 mspo 725: in rc.conf. This way, the domain will be properly shut down if
1.53 gdt 726: `xm shutdown -R` or `xm shutdown -H` is used on the dom0.
1.1 mspo 727:
728: Your domain should be now ready to work, enjoy.
729:
1.14 gdt 730: Creating an unprivileged Linux domain (domU)
1.5 mspo 731: --------------------------------------------
1.1 mspo 732:
733: Creating unprivileged Linux domains isn't much different from
734: unprivileged NetBSD domains, but there are some details to know.
735:
736: First, the second parameter passed to the disk declaration (the '0x1' in
737: the example below)
738:
1.3 mspo 739: disk = [ 'phy:/dev/wd0e,0x1,w' ]
1.1 mspo 740:
741: does matter to Linux. It wants a Linux device number here (e.g. 0x300
1.49 gdt 742: for hda). Linux builds device numbers as: (major \<\< 8 + minor).
743: So, hda1 which has major 3 and minor 1 on a Linux system will have
744: device number 0x301. Alternatively, devices names can be used (hda,
745: hdb, ...) as xentools has a table to map these names to devices
746: numbers. To export a partition to a Linux guest we can use:
1.1 mspo 747:
1.49 gdt 748: disk = [ 'phy:/dev/wd0e,0x300,w' ]
749: root = "/dev/hda1 ro"
1.1 mspo 750:
751: and it will appear as /dev/hda on the Linux system, and be used as root
752: partition.
753:
1.49 gdt 754: To install the Linux system on the partition to be exported to the
755: guest domain, the following method can be used: install
756: sysutils/e2fsprogs from pkgsrc. Use mke2fs to format the partition
757: that will be the root partition of your Linux domain, and mount it.
758: Then copy the files from a working Linux system, make adjustments in
759: `/etc` (fstab, network config). It should also be possible to extract
760: binary packages such as .rpm or .deb directly to the mounted partition
761: using the appropriate tool, possibly running under NetBSD's Linux
762: emulation. Once the filesystem has been populated, umount it. If
763: desirable, the filesystem can be converted to ext3 using tune2fs -j.
764: It should now be possible to boot the Linux guest domain, using one of
765: the vmlinuz-\*-xenU kernels available in the Xen binary distribution.
1.1 mspo 766:
767: To get the linux console right, you need to add:
768:
1.3 mspo 769: extra = "xencons=tty1"
1.1 mspo 770:
771: to your configuration since not all linux distributions auto-attach a
772: tty to the xen console.
773:
1.14 gdt 774: Creating an unprivileged Solaris domain (domU)
1.5 mspo 775: ----------------------------------------------
1.1 mspo 776:
1.50 gdt 777: See possibly outdated
778: [Solaris domU instructions](/ports/xen/howto-solaris/).
1.5 mspo 779:
1.1 mspo 780:
1.52 gdt 781: PCI passthrough: Using PCI devices in guest domains
782: ---------------------------------------------------
1.1 mspo 783:
1.53 gdt 784: The dom0 can give other domains access to selected PCI
1.52 gdt 785: devices. This can allow, for example, a non-privileged domain to have
786: access to a physical network interface or disk controller. However,
787: keep in mind that giving a domain access to a PCI device most likely
788: will give the domain read/write access to the whole physical memory,
789: as PCs don't have an IOMMU to restrict memory access to DMA-capable
1.53 gdt 790: device. Also, it's not possible to export ISA devices to non-dom0
1.52 gdt 791: domains, which means that the primary VGA adapter can't be exported.
792: A guest domain trying to access the VGA registers will panic.
793:
1.53 gdt 794: If the dom0 is NetBSD, it has to be running Xen 3.1, as support has
1.52 gdt 795: not been ported to later versions at this time.
796:
797: For a PCI device to be exported to a domU, is has to be attached to
798: the "pciback" driver in dom0. Devices passed to the dom0 via the
799: pciback.hide boot parameter will attach to "pciback" instead of the
800: usual driver. The list of devices is specified as "(bus:dev.func)",
1.5 mspo 801: where bus and dev are 2-digit hexadecimal numbers, and func a
802: single-digit number:
1.1 mspo 803:
1.52 gdt 804: pciback.hide=(00:0a.0)(00:06.0)
1.1 mspo 805:
1.52 gdt 806: pciback devices should show up in the dom0's boot messages, and the
1.5 mspo 807: devices should be listed in the `/kern/xen/pci` directory.
1.1 mspo 808:
1.52 gdt 809: PCI devices to be exported to a domU are listed in the "pci" array of
810: the domU's config file, with the format "0000:bus:dev.func".
1.1 mspo 811:
1.52 gdt 812: pci = [ '0000:00:06.0', '0000:00:0a.0' ]
1.1 mspo 813:
1.52 gdt 814: In the domU an "xpci" device will show up, to which one or more pci
815: busses will attach. Then the PCI drivers will attach to PCI busses as
816: usual. Note that the default NetBSD DOMU kernels do not have "xpci"
817: or any PCI drivers built in by default; you have to build your own
818: kernel to use PCI devices in a domU. Here's a kernel config example;
819: note that only the "xpci" lines are unusual.
820:
821: include "arch/i386/conf/XEN3_DOMU"
822:
823: # Add support for PCI busses to the XEN3_DOMU kernel
824: xpci* at xenbus ?
825: pci* at xpci ?
826:
827: # PCI USB controllers
828: uhci* at pci? dev ? function ? # Universal Host Controller (Intel)
829:
830: # USB bus support
831: usb* at uhci?
832:
833: # USB Hubs
834: uhub* at usb?
835: uhub* at uhub? port ? configuration ? interface ?
836:
837: # USB Mass Storage
838: umass* at uhub? port ? configuration ? interface ?
839: wd* at umass?
840: # SCSI controllers
841: ahc* at pci? dev ? function ? # Adaptec [23]94x, aic78x0 SCSI
842:
843: # SCSI bus support (for both ahc and umass)
844: scsibus* at scsi?
845:
846: # SCSI devices
847: sd* at scsibus? target ? lun ? # SCSI disk drives
848: cd* at scsibus? target ? lun ? # SCSI CD-ROM drives
1.1 mspo 849:
850:
1.28 gdt 851: NetBSD as a domU in a VPS
852: =========================
853:
854: The bulk of the HOWTO is about using NetBSD as a dom0 on your own
855: hardware. This section explains how to deal with Xen in a domU as a
856: virtual private server where you do not control or have access to the
857: dom0.
858:
1.52 gdt 859: VPS operators provide varying degrees of access and mechanisms for
860: configuration. The big issue is usually how one controls which kernel
861: is booted, because the kernel is nominally in the dom0 filesystem (to
862: which VPS users do not normally have acesss).
863:
864: A VPS user may want to compile a kernel for security updates, to run
865: npf, run IPsec, or any other reason why someone would want to change
866: their kernel.
867:
868: One approach is to have an adminstrative interface to upload a kernel,
869: or to select from a prepopulated list.
870:
871: Otehr approaches are pvgrub and py-grub, which are ways to start a
872: bootloader from the dom0 instead of the actual domU kernel, and for
873: that loader to then load a kernel from the domU filesystem. This is
874: closer to a regular physical computer, where someone who controls a
875: machine can replace the kernel.
876:
877: prmgr and pvgrub
878: ----------------
879:
1.28 gdt 880: TODO: Perhaps reference panix, prmgr, amazon as interesting examples.
1.52 gdt 881: Explain what prmgr does.
1.44 gdt 882:
883: Using npf
884: ---------
885:
886: In standard kernels, npf is a module, and thus cannot be loadeed in a
887: DOMU kernel.
888:
889: TODO: explain how to compile npf into a custom kernel, answering:
890: http://mail-index.netbsd.org/netbsd-users/2014/12/26/msg015576.html
CVSweb for NetBSD wikisrc <wikimaster@NetBSD.org> software: FreeBSD-CVSweb