VirtualBox

source: vbox/trunk/doc/manual/en_US/user_Troubleshooting.xml@ 54558

Last change on this file since 54558 was 54330, checked in by vboxsync, 10 years ago

Manual: Documented xHCI.

  • Property svn:mergeinfo set to (toggle deleted branches)
    /branches/VBox-4.0/doc/manual/en_US/user_Troubleshooting.xml71214
    /branches/VBox-4.2/doc/manual/en_US/user_Troubleshooting.xml91503-91504,​91506-91508,​91510,​91514-91515,​91521
    /branches/VBox-4.3/doc/manual/en_US/user_Troubleshooting.xml91223
    /branches/VBox-4.3/trunk/doc/manual/en_US/user_Troubleshooting.xml91223
    /branches/dsen/gui/doc/manual/en_US/user_Troubleshooting.xml79076-79078,​79089,​79109-79110,​79112-79113,​79127-79130,​79134,​79141,​79151,​79155,​79157-79159,​79193,​79197
    /branches/dsen/gui2/doc/manual/en_US/user_Troubleshooting.xml79224,​79228,​79233,​79235,​79258,​79262-79263,​79273,​79341,​79345,​79354,​79357,​79387-79388,​79559-79569,​79572-79573,​79578,​79581-79582,​79590-79591,​79598-79599,​79602-79603,​79605-79606,​79632,​79635,​79637,​79644
    /branches/dsen/gui3/doc/manual/en_US/user_Troubleshooting.xml79645-79692
File size: 70.4 KB
Line 
1<?xml version="1.0" encoding="UTF-8"?>
2<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.4//EN"
3"http://www.oasis-open.org/docbook/xml/4.4/docbookx.dtd">
4<chapter id="Troubleshooting">
5 <title>Troubleshooting</title>
6
7 <para>This chapter provides answers to commonly asked questions. In order to
8 improve your user experience with VirtualBox, it is recommended to read this
9 section to learn more about common pitfalls and get recommendations on how
10 to use the product.</para>
11
12 <sect1>
13 <title>Procedures and tools</title>
14
15 <sect2>
16 <title>Categorizing and isolating problems</title>
17
18 <para>More often than not, a virtualized guest behaves like a physical
19 system. Any problems that a physical machine would encounter, a virtual
20 machine will encounter as well. If, for example, Internet connectivity
21 is lost due to external issues, virtual machines will be affected just
22 as much as physical ones.</para>
23
24 <para>If a true VirtualBox problem is encountered, it helps to
25 categorize and isolate the problem first. Here are some of the questions
26 that should be answered before reporting a problem:<orderedlist>
27 <listitem>
28 <para>Is the problem specific to a certain guest OS? Specific
29 release of a guest OS? Especially with Linux guest related
30 problems, the issue may be specific to a certain distribution and
31 version of Linux.</para>
32 </listitem>
33
34 <listitem>
35 <para>Is the problem specific to a certain host OS? Problems are
36 usually not host OS specific (because most of the VirtualBox code
37 base is shared across all supported platforms), but especially in
38 the areas of networking and USB support, there are significant
39 differences between host platforms. Some GUI related issues are
40 also host specific.</para>
41 </listitem>
42
43 <listitem>
44 <para>Is the problem specific to certain host hardware? This
45 category of issues is typically related to the host CPU. Because
46 of significant differences between VT-x and AMD-V, problems may be
47 specific to one or the other technology. The exact CPU model may
48 also make a difference (even for software virtualization) because
49 different CPUs support different features, which may affect
50 certain aspects of guest CPU operation.</para>
51 </listitem>
52
53 <listitem>
54 <para>Is the problem specific to a certain virtualization mode?
55 Some problems may only occur in software virtualization mode,
56 others may be specific to hardware virtualization.</para>
57 </listitem>
58
59 <listitem>
60 <para>Is the problem specific to guest SMP? That is, is it related
61 to the number of virtual CPUs (VCPUs) in the guest? Using more
62 than one CPU usually significantly affects the internal operation
63 of a guest OS.</para>
64 </listitem>
65
66 <listitem>
67 <para>Is the problem specific to the Guest Additions? In some
68 cases, this is a given (e.g., a shared folders problem), in other
69 cases it may be less obvious (for example, display problems). And
70 if the problem is Guest Additions specific, is it also specific to
71 a certain version of the Additions?</para>
72 </listitem>
73
74 <listitem>
75 <para>Is the problem specific to a certain environment? Some
76 problems are related to a particular environment external to the
77 VM; this usually involves network setup. Certain configurations of
78 external servers such as DHCP or PXE may expose problems which do
79 not occur with other, similar servers.</para>
80 </listitem>
81
82 <listitem>
83 <para>Is the problem a regression? Knowing that an issue is a
84 regression usually makes it significantly easier to find the
85 solution. In this case, it is crucial to know which version is
86 affected and which is not.</para>
87 </listitem>
88 </orderedlist></para>
89 </sect2>
90
91 <sect2>
92 <title>Collecting debugging information</title>
93
94 <para>For problem determination, it is often important to collect
95 debugging information which can be analyzed by VirtualBox support. This
96 section contains information about what kind of information can be
97 obtained.</para>
98
99 <para>Every time VirtualBox starts up a VM, a so-called <emphasis
100 role="bold">"release log file"</emphasis> is created containing lots of
101 information about the VM configuration and runtime events. The log file
102 is called <computeroutput><literal>VBox.log</literal></computeroutput>
103 and resides in the VM log file folder. Typically this will be a
104 directory like this:<screen>$HOME/VirtualBox VMs/{machinename}/Logs</screen></para>
105
106 <para>When starting a VM, the configuration file of the last run will be
107 renamed to <computeroutput>.1</computeroutput>, up to
108 <computeroutput>.3</computeroutput>. Sometimes when there is a problem,
109 it is useful to have a look at the logs. Also when requesting support
110 for VirtualBox, supplying the corresponding log file is
111 mandatory.</para>
112
113 <para>For convenience, for each virtual machine, the VirtualBox main
114 window can show these logs in a window. To access it, select a virtual
115 machine from the list on the left and select "Show logs..." from the
116 "Machine" window.</para>
117
118 <para>The release log file (VBox.log) contains a wealth of diagnostic
119 information, such as Host OS type and version, VirtualBox version and
120 build (32-bit or 64-bit), a complete dump of the guest's configuration
121 (CFGM), detailed information about the host CPU type and supported
122 features, whether hardware virtualization is enabled, information about
123 VT-x/AMD-V setup, state transitions (creating, running, paused,
124 stopping, etc.), guest BIOS messages, Guest Additions messages,
125 device-specific log entries and, at the end of execution, final guest
126 state and condensed statistics.</para>
127
128 <para>In case of crashes, it is very important to collect <emphasis
129 role="bold">crash dumps</emphasis>. This is true for both host and guest
130 crashes. For information about enabling core dumps on Linux, Solaris,
131 and OS X systems, refer to the core dump article on the VirtualBox
132 website.<footnote>
133 <para><ulink
134 url="http://www.virtualbox.org/wiki/Core_dump">http://www.virtualbox.org/wiki/Core_dump</ulink>.</para>
135 </footnote></para>
136
137 <para>You can also use <computeroutput>VBoxManage
138 debugvm</computeroutput> to create a dump of a complete virtual machine;
139 see <xref linkend="vboxmanage-debugvm" />.</para>
140
141 <para>For network related problems, it is often helpful to capture a
142 trace of network traffic. If the traffic is routed through an adapter on
143 the host, it is possible to use Wireshark or a similar tool to capture
144 the traffic there. However, this often also includes a lot of traffic
145 unrelated to the VM.</para>
146
147 <para>VirtualBox provides an ability to capture network traffic only on
148 a specific VM's network adapter. Refer to the network tracing article on
149 the VirtualBox website<footnote>
150 <para><ulink
151 url="http://www.virtualbox.org/wiki/Network_tips">http://www.virtualbox.org/wiki/Network_tips</ulink>.</para>
152 </footnote> for information on enabling this capture. The trace files
153 created by VirtualBox are in <computeroutput>.pcap</computeroutput>
154 format and can be easily analyzed with Wireshark.</para>
155 </sect2>
156
157 <sect2 id="ts_debugger">
158 <title>The built-in VM debugger</title>
159
160 <para>VirtualBox includes a built-in VM debugger, which advanced users
161 may find useful. This debugger allows for examining and, to some extent,
162 controlling the VM state.<warning>
163 <para>Use the VM debugger at your own risk. There is no support for
164 it, and the following documentation is only made available for
165 advanced users with a very high level of familiarity with the
166 x86/AMD64 machine instruction set, as well as detailed knowledge of
167 the PC architecture. A degree of familiarity with the internals of
168 the guest OS in question may also be very helpful.</para>
169 </warning></para>
170
171 <para>The VM debugger is available in all regular production versions of
172 VirtualBox, but it is disabled by default because the average user will
173 have little use for it. There are two ways to access the
174 debugger:<itemizedlist>
175 <listitem>
176 <para>A debugger console window displayed alongside the VM</para>
177 </listitem>
178
179 <listitem>
180 <para>Via the <computeroutput>telnet</computeroutput> protocol at
181 port 5000</para>
182 </listitem>
183 </itemizedlist></para>
184
185 <para>The debugger can be enabled in three ways:<itemizedlist>
186 <listitem>
187 <para>Start the VM directly using <computeroutput>VirtualBox
188 --startvm</computeroutput>, with an additional
189 <computeroutput>--dbg</computeroutput>,
190 <computeroutput>--debug</computeroutput>, or
191 <computeroutput>--debug-command-line</computeroutput> argument.
192 See the VirtualBox usage help for details.</para>
193 </listitem>
194
195 <listitem>
196 <para>Set the
197 <computeroutput>VBOX_GUI_DBG_ENABLED</computeroutput> or
198 <computeroutput>VBOX_GUI_DBG_AUTO_SHOW</computeroutput>
199 environment variable to <computeroutput>true</computeroutput>
200 before launching the VirtualBox process. Setting these variables
201 (only their presence is checked) is effective even when the first
202 VirtualBox process is the VM selector window. VMs subsequently
203 launched from the selector will have the debugger enabled.</para>
204 </listitem>
205
206 <listitem>
207 <para>Set the <computeroutput>GUI/Dbg/Enabled</computeroutput>
208 extra data item to <computeroutput>true</computeroutput> before
209 launching the VM. This can be set globally or on a per VM
210 basis.</para>
211 </listitem>
212 </itemizedlist></para>
213
214 <para>A new 'Debug' menu entry will be added to the VirtualBox
215 application. This menu allows the user to open the debugger
216 console.</para>
217
218 <para>The VM debugger command syntax is loosely modeled on Microsoft and
219 IBM debuggers used on DOS, OS/2 and Windows. Users familiar with symdeb,
220 CodeView, or the OS/2 kernel debugger will find the VirtualBox VM
221 debugger familiar.</para>
222
223 <para>The most important command is
224 <computeroutput>help</computeroutput>. This will print brief usage help
225 for all debugger commands. The set of commands supported by the VM
226 debugger changes frequently and the
227 <computeroutput>help</computeroutput> command is always
228 up-to-date.</para>
229
230 <para>A brief summary of frequently used commands follows:<itemizedlist>
231 <listitem>
232 <para><computeroutput>stop</computeroutput> -- stops the VM
233 execution and enables single stepping</para>
234 </listitem>
235
236 <listitem>
237 <para><computeroutput>g</computeroutput> -- continue VM
238 execution</para>
239 </listitem>
240
241 <listitem>
242 <para><computeroutput>t</computeroutput> -- single step an
243 instruction</para>
244 </listitem>
245
246 <listitem>
247 <para><computeroutput>rg/rh/r</computeroutput> -- print the
248 guest/hypervisor/current registers</para>
249 </listitem>
250
251 <listitem>
252 <para><computeroutput>kg/kh/k</computeroutput> -- print the
253 guest/hypervisor/current call stack</para>
254 </listitem>
255
256 <listitem>
257 <para><computeroutput>da/db/dw/dd/dq</computeroutput> -- print
258 memory contents as ASCII/bytes/words/dwords/qwords</para>
259 </listitem>
260
261 <listitem>
262 <para><computeroutput>u</computeroutput> -- unassemble
263 memory</para>
264 </listitem>
265
266 <listitem>
267 <para><computeroutput>dg</computeroutput> -- print the guest's
268 GDT</para>
269 </listitem>
270
271 <listitem>
272 <para><computeroutput>di</computeroutput> -- print the guest's
273 IDT</para>
274 </listitem>
275
276 <listitem>
277 <para><computeroutput>dl</computeroutput> -- print the guest's
278 LDT</para>
279 </listitem>
280
281 <listitem>
282 <para><computeroutput>dt</computeroutput> -- print the guest's
283 TSS</para>
284 </listitem>
285
286 <listitem>
287 <para><computeroutput>dp*</computeroutput> -- print the guest's
288 page table structures</para>
289 </listitem>
290
291 <listitem>
292 <para><computeroutput>bp/br</computeroutput> -- set a
293 normal/recompiler breakpoint</para>
294 </listitem>
295
296 <listitem>
297 <para><computeroutput>bl</computeroutput> -- list
298 breakpoints</para>
299 </listitem>
300
301 <listitem>
302 <para><computeroutput>bc</computeroutput> -- clear a
303 breakpoint</para>
304 </listitem>
305
306 <listitem>
307 <para><computeroutput>writecore</computeroutput> -- writes a VM
308 core file to disk, refer <xref linkend="ts_guest-core-format" /></para>
309 </listitem>
310 </itemizedlist></para>
311
312 <para>See the built-in <computeroutput>help</computeroutput> for other
313 available commands.</para>
314
315 <para>The VM debugger supports symbolic debugging, although symbols for
316 guest code are often not available. For Solaris guests, the
317 <computeroutput>detect</computeroutput> command automatically determines
318 the guest OS version and locates kernel symbols in guest's memory.
319 Symbolic debugging is then available. For Linux guests, the
320 <computeroutput>detect</computeroutput> commands also determines the
321 guest OS version, but there are no symbols in the guest's memory. Kernel
322 symbols are available in the file
323 <computeroutput>/proc/kallsyms</computeroutput> on Linux guests. This
324 file must be copied to the host, for example using
325 <computeroutput>scp</computeroutput>. The
326 <computeroutput>loadmap</computeroutput> debugger command can be used to
327 make the symbol information available to the VM debugger. Note that the
328 <computeroutput>kallsyms</computeroutput> file contains the symbols for
329 the currently loaded modules; if the guest's configuration changes, the
330 symbols will change as well and must be updated.</para>
331
332 <para>For all guests, a simple way to verify that the correct symbols
333 are loaded is the <computeroutput>k</computeroutput> command. The guest
334 is normally idling and it should be clear from the symbolic information
335 that the guest operating system's idle loop is being executed.</para>
336
337 <para>Another group of debugger commands is the set of
338 <computeroutput>info</computeroutput> commands. Running
339 <computeroutput>info help</computeroutput> provides complete usage
340 information. The information commands provide ad-hoc data pertinent to
341 various emulated devices and aspects of the VMM. There is no general
342 guideline for using the <computeroutput>info</computeroutput> commands,
343 the right command to use depends entirely on the problem being
344 investigated. Some of the info commands are:<itemizedlist>
345 <listitem>
346 <para><computeroutput>cfgm</computeroutput> -- print a branch of
347 the configuration tree</para>
348 </listitem>
349
350 <listitem>
351 <para><computeroutput>cpuid</computeroutput> -- display the guest
352 CPUID leaves</para>
353 </listitem>
354
355 <listitem>
356 <para><computeroutput>ioport</computeroutput> -- print registered
357 I/O port ranges</para>
358 </listitem>
359
360 <listitem>
361 <para><computeroutput>mmio</computeroutput> -- print registered
362 MMIO ranges</para>
363 </listitem>
364
365 <listitem>
366 <para><computeroutput>mode</computeroutput> -- print the current
367 paging mode</para>
368 </listitem>
369
370 <listitem>
371 <para><computeroutput>pit</computeroutput> -- print the i8254 PIT
372 state</para>
373 </listitem>
374
375 <listitem>
376 <para><computeroutput>pic</computeroutput> -- print the i8259A PIC
377 state</para>
378 </listitem>
379
380 <listitem>
381 <para><computeroutput>ohci/ehci/xhci</computeroutput> -- print a subset
382 of the OHCI/EHCI/xHCI USB controller state</para>
383 </listitem>
384
385 <listitem>
386 <para><computeroutput>pcnet0</computeroutput> -- print the PCnet
387 state</para>
388 </listitem>
389
390 <listitem>
391 <para><computeroutput>vgatext</computeroutput> -- print the
392 contents of the VGA framebuffer formatted as standard text
393 mode</para>
394 </listitem>
395
396 <listitem>
397 <para><computeroutput>timers</computeroutput> -- print all VM
398 timers</para>
399 </listitem>
400 </itemizedlist></para>
401
402 <para>The output of the <computeroutput>info</computeroutput> commands
403 generally requires in-depth knowledge of the emulated device and/or
404 VirtualBox VMM internals. However, when used properly, the information
405 provided can be invaluable.</para>
406 </sect2>
407
408 <sect2 id="ts_guest-core-format">
409 <title>VM core format</title>
410
411 <para>VirtualBox uses the 64-bit ELF format for its VM core files
412 created by <computeroutput>VBoxManage debugvm</computeroutput>; see
413 <xref linkend="vboxmanage-debugvm" />. The VM core file contain the
414 memory and CPU dumps of the VM and can be useful for debugging your
415 guest OS. The 64-bit ELF object format specficiation can be obtained
416 here: <literal><ulink
417 url="http://downloads.openwatcom.org/ftp/devel/docs/elf-64-gen.pdf">http://downloads.openwatcom.org/ftp/devel/docs/elf-64-gen.pdf</ulink></literal>.</para>
418
419 <para>The overall layout of the VM core format is as follows:</para>
420
421 <para><screen>[ ELF 64 Header]
422[ Program Header, type PT_NOTE ]
423 -&gt; offset to COREDESCRIPTOR
424[ Program Header, type PT_LOAD ] - one for each contiguous physical memory range
425 -&gt; Memory offset of range
426 -&gt; File offset
427[ Note Header, type NT_VBOXCORE ]
428[ COREDESCRIPTOR ]
429 -&gt; Magic
430 -&gt; VM core file version
431 -&gt; VBox version
432 -&gt; Number of vCPUs etc.
433[ Note Header, type NT_VBOXCPU ] - one for each vCPU
434[ vCPU 1 Note Header ]
435 [ CPUMCTX - vCPU 1 dump ]
436[ Additional Notes + Data ] - currently unused
437[ Memory dump ]</screen></para>
438
439 <para>The memory descriptors contain physical addresses relative to the
440 guest and not virtual addresses. Regions of memory such as MMIO regions
441 are not included in the core file.</para>
442
443 <para>The relevant data structures and definitions can be found in the
444 VirtualBox sources under the following header files:
445 <computeroutput>include/VBox/dbgfcorefmt.h</computeroutput>,
446 <computeroutput>include/VBox/cpumctx.h</computeroutput> and
447 <computeroutput>src/VBox/Runtime/include/internal/ldrELFCommon.h</computeroutput>.</para>
448
449 <para>The VM core file can be inspected using
450 <computeroutput>elfdump</computeroutput> and GNU
451 <computeroutput>readelf</computeroutput> or other similar
452 utilities.</para>
453 </sect2>
454 </sect1>
455
456 <sect1>
457 <title>General</title>
458
459 <sect2 id="ts_config-periodic-flush">
460 <title>Guest shows IDE/SATA errors for file-based images on slow host
461 file system</title>
462
463 <para>Occasionally, some host file systems provide very poor writing
464 performance and as a consequence cause the guest to time out IDE/SATA
465 commands. This is normal behavior and should normally cause no real
466 problems, as the guest should repeat commands that have timed out.
467 However, some guests (e.g. some Linux versions) have severe problems if a
468 write to an image file takes longer than about 15 seconds. Some file
469 systems however require more than a minute to complete a single write,
470 if the host cache contains a large amount of data that needs to be
471 written.</para>
472
473 <para>The symptom for this problem is that the guest can no longer
474 access its files during large write or copying operations, usually
475 leading to an immediate hang of the guest.</para>
476
477 <para>In order to work around this problem (the true fix is to use a
478 faster file system that doesn't exhibit such unacceptable write
479 performance), it is possible to flush the image file after a certain
480 amount of data has been written. This interval is normally infinite, but
481 can be configured individually for each disk of a VM.</para>
482
483 <para>For IDE disks use the following command:</para>
484
485 <screen>VBoxManage setextradata "VM name"
486 "VBoxInternal/Devices/piix3ide/0/LUN#[x]/Config/FlushInterval" [b]</screen>
487
488 <para>For SATA disks use the following command:</para>
489
490 <screen>VBoxManage setextradata "VM name"
491 "VBoxInternal/Devices/ahci/0/LUN#[x]/Config/FlushInterval" [b]</screen>
492
493 <para>The value [x] that selects the disk for IDE is 0 for the master
494 device on the first channel, 1 for the slave device on the first
495 channel, 2 for the master device on the second channel or 3 for the
496 master device on the second channel. For SATA use values between 0 and
497 29. Only disks support this configuration option; it must not be set for
498 CD/DVD drives.</para>
499
500 <para>The unit of the interval [b] is the number of bytes written since
501 the last flush. The value for it must be selected so that the occasional
502 long write delays do not occur. Since the proper flush interval depends
503 on the performance of the host and the host filesystem, finding the
504 optimal value that makes the problem disappear requires some
505 experimentation. Values between 1000000 and 10000000 (1 to 10 megabytes)
506 are a good starting point. Decreasing the interval both decreases the
507 probability of the problem and the write performance of the guest.
508 Setting the value unnecessarily low will cost performance without
509 providing any benefits. An interval of 1 will cause a flush for each
510 write operation and should solve the problem in any case, but has a
511 severe write performance penalty.</para>
512
513 <para>Providing a value of 0 for [b] is treated as an infinite flush
514 interval, effectively disabling this workaround. Removing the extra data
515 key by specifying no value for [b] has the same effect.</para>
516 </sect2>
517
518 <sect2>
519 <title>Responding to guest IDE/SATA flush requests</title>
520
521 <para>If desired, the virtual disk images can be flushed when the guest
522 issues the IDE FLUSH CACHE command. Normally these requests are ignored
523 for improved performance. The parameters below are only accepted for
524 disk drives. They must not be set for DVD drives.</para>
525
526 <para>To enable flushing for IDE disks, issue the following
527 command:</para>
528
529 <screen>VBoxManage setextradata "VM name" "VBoxInternal/Devices/piix3ide/0/LUN#[x]/Config/IgnoreFlush" 0</screen>
530
531 <para>The value [x] that selects the disk is 0 for the master device on
532 the first channel, 1 for the slave device on the first channel, 2 for
533 the master device on the second channel or 3 for the master device on
534 the second channel.</para>
535
536 <para>To enable flushing for SATA disks, issue the following
537 command:</para>
538
539 <screen>VBoxManage setextradata "VM name" "VBoxInternal/Devices/ahci/0/LUN#[x]/Config/IgnoreFlush" 0</screen>
540
541 <para>The value [x] that selects the disk can be a value between 0 and
542 29.</para>
543
544 <para>Note that this doesn't affect the flushes performed according to
545 the configuration described in <xref linkend="ts_config-periodic-flush"
546 xrefstyle="template: %n" />. Restoring the default of ignoring flush
547 commands is possible by setting the value to 1 or by removing the
548 key.</para>
549 </sect2>
550
551 <sect2 id="ts_host-freq-boost">
552 <title>Performance variation with frequency boosting</title>
553
554 <para>
555 Many newer multi-core processors support some form of frequency
556 boosting, which means that if only one core is utilized, it can run
557 faster (possibly 50% faster or even more) than the rated CPU frequency.
558 This causes measured performance to vary somewhat as a function of the
559 momentary overall system load. The exact behavior depends strongly
560 on the specific processor model.
561 </para>
562
563 <para>
564 As a consequence, benchmarking on systems which utilize frequency
565 boosting may produce unstable and non-repeatable results, especially
566 if benchmark runs are short (on the order of seconds). To obtain stable
567 results, benchmarks must be run over longer periods of time and with a
568 constant system load apart from the VM being tested.
569 </para>
570 </sect2>
571
572 <sect2 id="ts_host-freq-scaling">
573 <title>Frequency scaling effect on CPU usage</title>
574
575 <para>
576 On some hardware platforms and operating systems, CPU frequency
577 scaling may cause CPU usage reporting to be highly misleading. This
578 happens in situations when the host CPU load is significant but not
579 heavy, such as 15-30% of the maximum.
580 </para>
581
582 <para>
583 Most operating systems determine CPU usage in terms of time spent,
584 measuring for example how many nanoseconds the systems or a process
585 was active within one second. However, in order to save energy, modern
586 systems can significantly scale down CPU speed when the system is not
587 fully loaded. Naturally, when the CPU is running at (for example) one
588 half of its maximum speed, the same number of instructions will take
589 roughly twice as long to execute compared to running at full speed.
590 </para>
591
592 <para>
593 Depending on the specific hardware and host OS, this effect can very
594 significantly skew the CPU usage reported by the OS; the reported CPU
595 usage can be several times higher than what it would have been had the
596 CPU been running at full speed. The effect can be observed both on
597 the host OS and in a guest OS.
598 </para>
599 </sect2>
600
601 <sect2 id="ts_win-cpu-usage-rept">
602 <title>Inaccurate Windows CPU usage reporting</title>
603
604 <para>
605 CPU usage reporting tools which come with Windows (Task Manager, Resource
606 Monitor) do not take the time spent processing hardware interrupts into
607 account. If the interrupt load is heavy (thousands of interrupts per second),
608 CPU usage may be significantly underreported.
609 </para>
610
611 <para>
612 This problem affects Windows as both host and guest OS. Sysinternals tools
613 (e.g. Process Explorer) do not suffer from this problem.
614 </para>
615 </sect2>
616
617 <sect2 id="ts_host-powermgmt">
618 <title>Poor performance caused by host power management</title>
619
620 <para>On some hardware platforms and operating systems, virtualization
621 performance is negatively affected by host CPU power management. The
622 symptoms may be choppy audio in the guest or erratic guest clock
623 behavior.</para>
624
625 <para>Some of the problems may be caused by firmware and/or host
626 operating system bugs. Therefore, updating the firmware and applying
627 operating systems fixes is recommended.</para>
628
629 <para>For optimal virtualization performance, the C1E power state
630 support in the system's BIOS should be disabled, if such a setting is
631 available (not all systems support the C1E power state). On Intel
632 systems the <computeroutput>Intel C State</computeroutput> setting
633 should be disabled. Disabling other power management settings
634 may also improve performance. However, a balance between performance
635 and power consumption must always be considered.</para>
636 </sect2>
637
638 <sect2 id="ts_gui-2d-grayed-out">
639 <title>GUI: 2D Video Acceleration option is grayed out</title>
640
641 <para>To use 2D Video Acceleration within VirtualBox, your host's video
642 card should support certain OpenGL extensions. On startup, VirtualBox
643 checks for those extensions, and, if the test fails, this option is
644 silently grayed out.</para>
645
646 <para>To find out why it has failed, you can manually execute the
647 following command:</para>
648
649 <screen>VBoxTestOGL --log "log_file_name" --test 2D</screen>
650
651 <para>It will list the required OpenGL extensions one by one and will
652 show you which one failed the test. This usually means that you are
653 running an outdated or misconfigured OpenGL driver on your host. It can
654 also mean that your video chip is lacking required functionality.</para>
655 </sect2>
656 </sect1>
657
658 <sect1 id="ts_win-guests">
659 <title>Windows guests</title>
660
661 <sect2>
662 <title>Windows bluescreens after changing VM configuration</title>
663
664 <para>Changing certain virtual machine settings can cause Windows guests
665 to fail during start up with a bluescreen. This may happen if you change
666 VM settings after installing Windows, or if you copy a disk image with
667 an already installed Windows to a newly created VM which has settings
668 that differ from the original machine.</para>
669
670 <para>This applies in particular to the following settings:<itemizedlist>
671 <listitem>
672 <para>The ACPI and I/O APIC settings should never be changed after
673 installing Windows. Depending on the presence of these hardware
674 features, the Windows installation program chooses special kernel
675 and device driver versions and will fail to startup should these
676 hardware features be removed. (Enabling them for a Windows VM
677 which was installed without them does not cause any harm. However,
678 Windows will not use these features in this case.)</para>
679 </listitem>
680
681 <listitem>
682 <para>Changing the storage controller hardware will cause bootup
683 failures as well. This might also apply to you if you copy a disk
684 image from an older version of VirtualBox to a virtual machine
685 created with a newer VirtualBox version; the default subtype of
686 IDE controller hardware was changed from PIIX3 to PIIX4 with
687 VirtualBox 2.2. Make sure these settings are identical.</para>
688 </listitem>
689 </itemizedlist></para>
690 </sect2>
691
692 <sect2>
693 <title>Windows 0x101 bluescreens with SMP enabled (IPI timeout)</title>
694
695 <para>If a VM is configured to have more than one processor (symmetrical
696 multiprocessing, SMP), some configurations of Windows guests crash with
697 an 0x101 error message, indicating a timeout for inter-processor
698 interrupts (IPIs). These interrupts synchronize memory management
699 between processors.</para>
700
701 <para>According to Microsoft, this is due to a race condition in
702 Windows. A hotfix is available.<footnote>
703 <para>See <ulink
704 url="http://support.microsoft.com/kb/955076">http://support.microsoft.com/kb/955076</ulink>.</para>
705 </footnote> If this does not help, please reduce the number of virtual
706 processors to 1.</para>
707 </sect2>
708
709 <sect2>
710 <title>Windows 2000 installation failures</title>
711
712 <para>When installing Windows 2000 guests, you might run into one of the
713 following issues:</para>
714
715 <itemizedlist>
716 <listitem>
717 <para>Installation reboots, usually during component
718 registration.</para>
719 </listitem>
720
721 <listitem>
722 <para>Installation fills the whole hard disk with empty log
723 files.</para>
724 </listitem>
725
726 <listitem>
727 <para>Installation complains about a failure installing
728 <literal>msgina.dll</literal>.</para>
729 </listitem>
730 </itemizedlist>
731
732 <para>These problems are all caused by a bug in the hard disk driver of
733 Windows 2000. After issuing a hard disk request, there is a race
734 condition in the Windows driver code which leads to corruption if the
735 operation completes too fast, i.e. the hardware interrupt from the IDE
736 controller arrives too soon. With physical hardware, there is a
737 guaranteed delay in most systems so the problem is usually hidden there
738 (however it should be possible to reproduce it on physical hardware as
739 well). In a virtual environment, it is possible for the operation to be
740 done immediately (especially on very fast systems with multiple CPUs)
741 and the interrupt is signaled sooner than on a physical system. The
742 solution is to introduce an artificial delay before delivering such
743 interrupts. This delay can be configured for a VM using the following
744 command:</para>
745
746 <screen>VBoxManage setextradata "VM name" "VBoxInternal/Devices/piix3ide/0/Config/IRQDelay" 1</screen>
747
748 <para>This sets the delay to one millisecond. In case this doesn't help,
749 increase it to a value between 1 and 5 milliseconds. Please note that
750 this slows down disk performance. After installation, you should be able
751 to remove the key (or set it to 0).</para>
752 </sect2>
753
754 <sect2>
755 <title>How to record bluescreen information from Windows guests</title>
756
757 <para>When Windows guests run into a kernel crash, they display the
758 infamous bluescreen. Depending on how Windows is configured, the
759 information will remain on the screen until the machine is restarted or
760 it will reboot automatically. During installation, Windows is usually
761 configured to reboot automatically. With automatic reboots, there is no
762 chance to record the bluescreen information which might be important for
763 problem determination.</para>
764
765 <para>VirtualBox provides a method of halting a guest when it wants to
766 perform a reset. In order to enable this feature, issue the following
767 command:</para>
768
769 <para><screen>VBoxManage setextradata "VM name" "VBoxInternal/PDM/HaltOnReset" 1</screen></para>
770 </sect2>
771
772 <sect2>
773 <title>No networking in Windows Vista guests</title>
774
775 <para>With Windows Vista, Microsoft dropped support for the AMD PCNet
776 card that VirtualBox used to provide as the default virtual network card
777 before version 1.6.0. For Windows Vista guests, VirtualBox now uses an
778 Intel E1000 card by default.</para>
779
780 <para>If, for some reason, you still want to use the AMD card, you need
781 to download the PCNet driver from the AMD website (available for 32-bit
782 Windows only). You can transfer it into the virtual machine using a
783 shared folder, see (see <xref linkend="sharedfolders" />).</para>
784 </sect2>
785
786 <sect2>
787 <title>Windows guests may cause a high CPU load</title>
788
789 <para>Several background applications of Windows guests, especially
790 virus scanners, are known to increases the CPU load notably even if the
791 guest appears to be idle. We recommend to deactivate virus scanners
792 within virtualized guests if possible.</para>
793 </sect2>
794
795 <sect2>
796 <title>Long delays when accessing shared folders</title>
797
798 <para>The performance for accesses to shared folders from a Windows
799 guest might be decreased due to delays during the resolution of the
800 VirtualBox shared folders name service. To fix these delays, add the
801 following entries to the file
802 <computeroutput>\windows\system32\drivers\etc\lmhosts</computeroutput>
803 of the Windows guest:</para>
804
805 <screen>255.255.255.255 VBOXSVR #PRE
806255.255.255.255 VBOXSRV #PRE</screen>
807
808 <para>After doing this change, a reboot of the guest is required.</para>
809 </sect2>
810
811 <sect2>
812 <title>USB tablet coordinates wrong in Windows 98 guests</title>
813
814 <para>If a Windows 98 VM is configured to use the emulated USB tablet
815 (absolute pointing device), the coordinate translation may be incorrect
816 and the pointer is restricted to the upper left quarter of the guest's
817 screen.
818 </para>
819
820 <para>The USB HID (Human Interface Device) drivers in Windows 98 are very
821 old and do not handle tablets the same way all more recent operating
822 systems do (Windows 2000 and later, Mac OS X, Solaris). To
823 work around the problem, issue the following command:
824 </para>
825
826 <para><screen>VBoxManage setextradata "VM name" "VBoxInternal/USB/HidMouse/0/Config/CoordShift" 0</screen></para>
827
828 <para>To restore the default behavior, remove the key or set its value
829 to 1.
830 </para>
831 </sect2>
832
833 <sect2>
834 <title>Windows guests are removed from an Active Directory domain after
835 restoring a snapshot</title>
836
837 <para>If a Windows guest is a member of an Active Directory domain and
838 the snapshot feature of VirtualBox is used, it could happen it loses
839 this status after you restore an older snapshot.
840 </para>
841
842 <para>The reason is the automatic machine password changing performed by
843 Windows in regular intervals for security purposes. You can disable
844 this feature by following the instruction of this <ulink
845 url="http://support.microsoft.com/kb/154501">http://support.microsoft.com/kb/154501</ulink>
846 article from Microsoft.
847 </para>
848 </sect2>
849
850 <sect2 id="ts_d3d8-d3d9-restore">
851 <title>Restoring d3d8.dll and d3d9.dll</title>
852
853 <para>VirtualBox Guest Additions for Windows prior to 4.1.8 did not properly
854 back up the original d3d8.dll and d3d9.dll system files when selecting
855 and installing the experimental Direct3D support. This process replaces
856 both system files with files from the VirtualBox Guest Additions so that
857 Direct3D calls can be handled correctly. Although this issue was fixed
858 with VirtualBox 4.1.8, there is no way the Windows Guest Additions
859 installer can repair these files.</para>
860
861 <para>Corruption of these files has no implications in case 3D acceleration
862 is enabled and basic Direct3D support is installed, that is, without WDDM
863 (on Windows Vista or higher) or on older Windows systems like Windows XP.
864 With the basic Direct3D support all Direct3D 8.0 and Direct3D 9.0
865 applications will utilize VirtualBox Direct3D files directly and thus
866 will run as expected.</para>
867
868 <para>For WDDM Direct3D support however, the originally shipped d3d8.dll and
869 d3d9.dll files are required in order to run Direct3D 8.0
870 and Direct3D 9.0 applications. As a result of the above mentioned system
871 files corruption these applications will not work anymore. See below for
872 a step-by-step guide for restoring the original d3d8.dll and d3d9.dll
873 system files in case the VirtualBox Guest Additions installer warned
874 about those incorrect files or when having trouble running Direct3D
875 applications.</para>
876
877 <note><para>Starting at Windows 7 the 3D desktop (aka Aero) uses DirectX 10
878 for rendering so that corrupted d3d8.dll and d3d9.dll system files will
879 have no effect on the actual rendering.</para></note>
880
881 <para>This is why such a detected file corruption is not considered as fatal
882 for the basic Direct3D installation on all supported Windows guests,
883 and for WDDM Direct3D installation on Windows 7 and later guests.</para>
884
885 <para>Extracting d3d8 and d3d9.dll from a Windows XP installation CD:</para>
886
887 <orderedlist>
888 <listitem>
889 <para>Download and install the latest version of 7-Zip File Manager <ulink
890 url="http//www.7-zip.org">http//www.7-zip.org</ulink></para>
891 </listitem>
892
893 <listitem>
894 <para>Browse into installation CD for example E:\i386 (or AMD64 for 64bit version)</para>
895 </listitem>
896
897 <listitem>
898 <para>Locate file d3d8.dl_ and d3d9.dl_, double click on it and Extract d3d8.dll and d3d9.dll</para>
899 </listitem>
900
901 <listitem>
902 <para>Reboot Windows in Safe mode</para>
903 </listitem>
904
905 <listitem>
906 <para>Copy extracted d3d8.dll and d3d9.dll to C:\Windows\system32 and C:\Windows\system32\dllcache</para>
907 </listitem>
908
909 <listitem>
910 <para>Reboot</para>
911 </listitem>
912 </orderedlist>
913
914 <para>Extracting d3d8 and d3d9.dll from Windows XP Service pack </para>
915
916 <orderedlist>
917 <listitem>
918 <para>1, 3-6 Same as installation CD</para>
919 </listitem>
920 <listitem>
921 <para>Use 'Open inside' to open WindowsXP-KB936929-SP3-x86.exe as archive and browse i386 directory.</para>
922 </listitem>
923 </orderedlist>
924
925 <para>Extracting d3d8 and d3d9.dll from Vista/Windows7 installation CD or Service Pack iso</para>
926
927 <orderedlist>
928 <listitem>
929 <para>Download and install the latest version of 7-Zip File Manager <ulink
930 url="http//www.7-zip.org">http//www.7-zip.org</ulink></para>
931 </listitem>
932
933 <listitem>
934 <para>Browse into installation CD for example E:\sources</para>
935 </listitem>
936
937 <listitem>
938 <para>Locate file install.wim and double click it. After 7-Zip utility opens the file, you'll get a few numbered folders. Each numeric subfolder represents a different version of Windows (Starter, Home Basic, and so on)</para>
939 </listitem>
940
941 <listitem>
942 <para>After entering into the one of the numeric folders, browse into Windows\System32 (or C:\Windows\SysWOW64 for 64 bit version) directory locate d3d8.dll and d3d9.dll and extract</para>
943 </listitem>
944
945 <listitem>
946 <para>Copy extracted d3d8.dll and d3d9.dll to C:\Windows\system32 or C:\Windows\SysWOW64 (files from system32 should go to system32, from SysWOW64 to SysWOW64)</para>
947 </listitem>
948
949 <listitem>
950 <para>Reboot</para>
951 </listitem>
952 </orderedlist>
953 </sect2>
954
955 </sect1>
956
957 <sect1 id="ts_lin-x11-guests">
958 <title>Linux and X11 guests</title>
959
960 <sect2>
961 <title>Linux guests may cause a high CPU load</title>
962
963 <para>Some Linux guests may cause a high CPU load even if the guest
964 system appears to be idle. This can be caused by a high timer frequency
965 of the guest kernel. Some Linux distributions, for example Fedora, ship
966 a Linux kernel configured for a timer frequency of <emphasis
967 role="bold"> 1000Hz</emphasis>. We recommend to recompile the guest
968 kernel and to select a timer frequency of 100Hz.</para>
969
970 <para>Linux kernels shipped with Red Hat Enterprise Linux (RHEL) as of
971 release 4.7 and 5.1 as well as kernels of related Linux distributions
972 (for instance CentOS and Oracle Linux) support a kernel
973 parameter <emphasis>divider=N</emphasis>. Hence, such kernels support a
974 lower timer frequency without recompilation. We suggest to add the
975 kernel parameter <emphasis>divider=10</emphasis> to select a guest
976 kernel timer frequency of 100Hz.</para>
977 </sect2>
978
979 <sect2>
980 <title>AMD Barcelona CPUs</title>
981
982 <para>Most Linux-based guests will fail with AMD Phenoms or
983 Barcelona-level Opterons due to a bug in the Linux kernel. Enable the
984 I/O-APIC to work around the problem (see <xref
985 linkend="settings-system" />).</para>
986 </sect2>
987
988 <sect2 id="ts_linux-buggy">
989 <title>Buggy Linux 2.6 kernel versions</title>
990
991 <para>The following bugs in Linux kernels prevent them from executing
992 correctly in VirtualBox, causing VM boot crashes:<itemizedlist>
993 <listitem>
994 <para>The Linux kernel version 2.6.18 (and some 2.6.17 versions)
995 introduced a race condition that can cause boot crashes in
996 VirtualBox. Please use a kernel version 2.6.19 or later.</para>
997 </listitem>
998
999 <listitem>
1000 <para>With hardware virtualization and the I/O APIC enabled,
1001 kernels before 2.6.24-rc6 may panic on boot with the following
1002 message:<screen>Kernel panic - not syncing: IO-APIC + timer doesn't work! Boot with
1003apic=debug and send a report. Then try booting with the 'noapic' option</screen></para>
1004
1005 <para>If you see this message, either disable hardware
1006 virtualization or the I/O APIC (see <xref
1007 linkend="settings-system" />), or upgrade the guest to a newer
1008 kernel.<footnote>
1009 <para>See <ulink
1010 url="http://www.mail-archive.com/git-commits-head@vger.kernel.org/msg30813.html">http://www.mail-archive.com/git-commits-head@vger.kernel.org/msg30813.html</ulink>
1011 for details about the kernel fix.</para>
1012 </footnote></para>
1013 </listitem>
1014 </itemizedlist></para>
1015 </sect2>
1016
1017 <sect2>
1018 <title>Shared clipboard, auto-resizing and seamless desktop in X11
1019 guests</title>
1020
1021 <para>Guest desktop services in guests running the X11 window system
1022 (Solaris, Linux and others) are provided by a guest service called
1023 <computeroutput>VBoxClient</computeroutput>, which runs under the ID of
1024 the user who started the desktop session and is automatically started
1025 using the following command lines <screen>VBoxClient --clipboard
1026VBoxClient --display
1027VBoxClient --seamless</screen> when your X11 user session is started if you
1028 are using a common desktop environment (Gnome, KDE and others). If a
1029 particular desktop service is not working correctly, it is worth
1030 checking whether the process which should provide it is running.</para>
1031
1032 <para>The <computeroutput>VBoxClient</computeroutput> processes create
1033 files in the user's home directory with names of the form
1034 <computeroutput>.vboxclient-*.pid</computeroutput> when they are running
1035 in order to prevent a given service from being started twice. It can
1036 happen due to misconfiguration that these files are created owned by
1037 root and not deleted when the services are stopped, which will prevent
1038 them from being started in future sessions. If the services cannot be
1039 started, you may wish to check whether these files still exist.</para>
1040 </sect2>
1041 </sect1>
1042
1043 <sect1 id="ts_sol-guests">
1044 <title>Solaris guests</title>
1045
1046 <sect2>
1047 <title>Older Solaris 10 releases crash in 64-bit mode</title>
1048
1049 <para>Solaris 10 releases up to and including Solaris 10 8/07 ("S10U4")
1050 incorrectly detect newer Intel processors produced since 2007. This
1051 problem leads to the 64-bit Solaris kernel crashing (and usually causing
1052 a triple fault) almost immediately during startup, in both virtualized
1053 and physical environments.
1054 </para>
1055 <para>
1056 The recommended solution is upgrading to at least Solaris 10 5/08
1057 ("S10U5"). Alternative solutions include forcing Solaris to always
1058 boot the 32-bit kernel or applying a patch for bug 6574102 (while
1059 Solaris is using the 32-bit kernel).
1060 </para>
1061
1062 </sect2>
1063
1064 <sect2>
1065 <title>Solaris 8 5/01 and earlier may crash on startup</title>
1066
1067 <para>
1068 Solaris 2.6, 7 and 8 releases up to and including Solaris 8 4/01 ("S8U4")
1069 incorrectly set up Machine Check Exception (MCE) MSRs on Pentium 4 and
1070 somene later Intel CPUs. The problem leads to the Solaris kernel crashing
1071 (and usually causing a triple fault) almost immediately during startup, in both
1072 virtualized and physical environments. Solaris 9 and later releases are
1073 not affected by this problem, and neither is Solaris 2.5.1 and earlier.
1074 </para>
1075 <para>
1076 The recommended solution is upgrading to at least Solaris 8 7/01
1077 ("S8U5"). Alternative solutions include applying a patch for bugs 4408508
1078 and 4414557 (on an unaffected system).
1079 </para>
1080
1081 </sect2>
1082 </sect1>
1083
1084 <sect1 id="ts_fbsd-guests">
1085 <title>FreeBSD guests</title>
1086
1087 <sect2>
1088 <title>FreeBSD 10.0 may hang with xHCI</title>
1089
1090 <para>
1091 If xHCI (USB 3.0) emulation is enabled for FreeBSD 10.0 guests, the guest
1092 OS will hang. This is caused by the guest OS incorrectly handling systems
1093 where MSIs (Message Signaled Interrupts) are not used with the xHCI device.
1094 </para>
1095 <para>
1096 The problem does not exist in earlier FreeBSD releases and was fixed in
1097 FreeBSD 10.1.
1098 </para>
1099 </sect2>
1100 </sect1>
1101
1102 <sect1 id="ts_win-hosts">
1103 <title>Windows hosts</title>
1104
1105 <sect2>
1106 <title>VBoxSVC out-of-process COM server issues</title>
1107
1108 <para>VirtualBox makes use of the Microsoft Component Object Model (COM)
1109 for inter- and intra-process communication. This allows VirtualBox to
1110 share a common configuration among different virtual machine processes
1111 and provide several user interface options based on a common
1112 architecture. All global status information and configuration is
1113 maintained by the process <computeroutput>VBoxSVC.exe</computeroutput>,
1114 which is an out-of-process COM server. Whenever a VirtualBox process is
1115 started, it requests access to the COM server and Windows automatically
1116 starts the process. Note that it should never be started by the end
1117 user.</para>
1118
1119 <para>When the last process disconnects from the COM server, it will
1120 terminate itself after some seconds. The VirtualBox configuration (XML
1121 files) is maintained and owned by the COM server and the files are
1122 locked whenever the server runs.</para>
1123
1124 <para>In some cases - such as when a virtual machine is terminated
1125 unexpectedly - the COM server will not notice that the client is
1126 disconnected and stay active for a longer period (10 minutes or so)
1127 keeping the configuration files locked. In other rare cases the COM
1128 server might experience an internal error and subsequently other
1129 processes fail to initialize it. In these situations, it is recommended
1130 to use the Windows task manager to kill the process
1131 <computeroutput>VBoxSVC.exe</computeroutput>.</para>
1132 </sect2>
1133
1134 <sect2>
1135 <title>CD/DVD changes not recognized</title>
1136
1137 <para>In case you have assigned a physical CD/DVD drive to a guest and
1138 the guest does not notice when the medium changes, make sure that the
1139 Windows media change notification (MCN) feature is not turned off. This
1140 is represented by the following key in the Windows registry:<screen><literal>HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Cdrom\Autorun</literal></screen>Certain
1141 applications may disable this key against Microsoft's advice. If it is
1142 set to 0, change it to 1 and reboot your system. VirtualBox relies on
1143 Windows notifying it of media changes.</para>
1144 </sect2>
1145
1146 <sect2>
1147 <title>Sluggish response when using Microsoft RDP client</title>
1148
1149 <para>If connecting to a Virtual Machine via the Microsoft RDP client
1150 (called Remote Desktop Connection), there can be large delays between
1151 input (moving the mouse over a menu is the most obvious situation) and
1152 output. This is because this RDP client collects input for a certain
1153 time before sending it to the RDP server.</para>
1154
1155 <para>The interval can be decreased by setting a Windows registry key to
1156 smaller values than the default of 100. The key does not exist initially
1157 and must be of type DWORD. The unit for its values is milliseconds.
1158 Values around 20 are suitable for low-bandwidth connections between the
1159 RDP client and server. Values around 4 can be used for a gigabit
1160 Ethernet connection. Generally values below 10 achieve a performance
1161 that is very close to that of the local input devices and screen of the
1162 host on which the Virtual Machine is running.</para>
1163
1164 <para>Depending whether the setting should be changed for an individual
1165 user or for the system, either</para>
1166
1167 <screen>HKEY_CURRENT_USER\Software\Microsoft\Terminal Server Client\Min Send Interval</screen>
1168
1169 <para>or</para>
1170
1171 <screen>HKEY_LOCAL_MACHINE\Software\Microsoft\Terminal Server Client\Min Send Interval</screen>
1172
1173 <para>can be set appropriately.</para>
1174 </sect2>
1175
1176 <sect2>
1177 <title>Running an iSCSI initiator and target on a single system</title>
1178
1179 <para>Deadlocks can occur on a Windows host when attempting to access an
1180 iSCSI target running in a guest virtual machine with an iSCSI initiator
1181 (e.g. Microsoft iSCSI Initiator) that is running on the host. This is
1182 caused by a flaw in the Windows cache manager component, and causes
1183 sluggish host system response for several minutes, followed by a
1184 "Delayed Write Failed" error message in the system tray or in a separate
1185 message window. The guest is blocked during that period and may show
1186 error messages or become unstable.</para>
1187
1188 <para>Setting the environment variable
1189 <computeroutput>VBOX_DISABLE_HOST_DISK_CACHE</computeroutput> to 1 will
1190 enable a workaround for this problem until Microsoft addresses the
1191 issue. For example, open a command prompt window and start VirtualBox
1192 like this:</para>
1193
1194 <screen>set VBOX_DISABLE_HOST_DISK_CACHE=1
1195VirtualBox</screen>
1196
1197 <para>While this will decrease guest disk performance (especially
1198 writes), it does not affect the performance of other applications
1199 running on the host.</para>
1200 </sect2>
1201
1202 <sect2>
1203 <title>Bridged networking adapters missing</title>
1204
1205 <para>If no bridged adapters show up in the "Networking" section of the
1206 VM settings, this typically means that the bridged networking driver was
1207 not installed properly on your host. This could be due to the following
1208 reasons: <itemizedlist>
1209 <listitem>
1210 <para>The maximum allowed filter count was reached on the host. In
1211 this case, the MSI log would mention the
1212 <computeroutput>0x8004a029</computeroutput> error code returned on
1213 NetFlt network component install:<screen>VBoxNetCfgWinInstallComponent: Install failed, hr (0x8004a029)</screen></para>
1214
1215 <para>You can try to increase the maximum filter count in the
1216 Windows registry at the following key:<screen>HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Network\MaxNumFilters</screen>The
1217 maximum number allowed is 14. After a reboot, try to re-install
1218 VirtualBox.</para>
1219 </listitem>
1220
1221 <listitem>
1222 <para>The INF cache is corrupt. In this case, the install log
1223 (<computeroutput>%windir%\inf\setupapi.log</computeroutput> on XP
1224 or <computeroutput>%windir%\inf\setupapi.dev.log</computeroutput>
1225 on Vista or later) would typically mention the failure to find a
1226 suitable driver package for either the
1227 <computeroutput>sun_VBoxNetFlt</computeroutput> or
1228 <computeroutput>sun_VBoxNetFltmp</computeroutput> components. The
1229 solution then is to uninstall VirtualBox, remove the INF cache
1230 (<computeroutput>%windir%\inf\INFCACHE.1</computeroutput>), reboot
1231 and try to re-install VirtualBox</para>
1232 </listitem>
1233 </itemizedlist></para>
1234 </sect2>
1235
1236 <sect2>
1237 <title>Host-only networking adapters cannot be created</title>
1238
1239 <para>If host-only adapter cannot be created (either via the Manager or
1240 VBoxManage), then the INF cache is probably corrupt. In this case, the
1241 install log (<computeroutput>%windir%\inf\setupapi.log</computeroutput>
1242 on XP or <computeroutput>%windir%\inf\setupapi.dev.log</computeroutput>
1243 on Vista or later) would typically mention the failure to find a
1244 suitable driver package for the
1245 <computeroutput>sun_VBoxNetAdp</computeroutput> component. Again, as
1246 with the bridged networking problem described above, the solution is to
1247 uninstall VirtualBox, remove the INF cache
1248 (<computeroutput>%windir%\inf\INFCACHE.1</computeroutput>), reboot and
1249 try to re-install VirtualBox.</para>
1250 </sect2>
1251 </sect1>
1252
1253 <sect1 id="ts_lin-hosts">
1254 <title>Linux hosts</title>
1255
1256 <sect2 id="ts_linux-kernelmodule-fails-to-load">
1257 <title>Linux kernel module refuses to load</title>
1258
1259 <para>If the VirtualBox kernel module
1260 (<computeroutput>vboxdrv</computeroutput>) refuses to load, i.e. you get
1261 an "Error inserting vboxdrv: Invalid argument", check (as root) the
1262 output of the <computeroutput>dmesg</computeroutput> command to find out
1263 why the load failed. Most probably the kernel disagrees with the version
1264 of the gcc used to compile the module. Make sure that you use the same
1265 compiler as used to build the kernel.</para>
1266 </sect2>
1267
1268 <sect2>
1269 <title>Linux host CD/DVD drive not found</title>
1270
1271 <para>If you have configured a virtual machine to use the host's CD/DVD
1272 drive, but this does not appear to work, make sure that the current user
1273 has permission to access the corresponding Linux device file
1274 (<computeroutput>/dev/hdc</computeroutput> or
1275 <computeroutput>/dev/scd0</computeroutput> or
1276 <computeroutput>/dev/cdrom</computeroutput> or similar). On most
1277 distributions, the user must be added to a corresponding group (usually
1278 called <computeroutput>cdrom</computeroutput> or
1279 <computeroutput>cdrw</computeroutput>).</para>
1280 </sect2>
1281
1282 <sect2>
1283 <title>Linux host CD/DVD drive not found (older distributions)</title>
1284
1285 <para>On older Linux distributions, if your CD/DVD device has a
1286 different name, VirtualBox may be unable to find it. On older Linux
1287 hosts, VirtualBox performs the following steps to locate your CD/DVD
1288 drives:</para>
1289
1290 <para><orderedlist>
1291 <listitem>
1292 <para>VirtualBox examines if the environment variable
1293 <computeroutput>VBOX_CDROM</computeroutput> is defined (see
1294 below). If so, VirtualBox omits all the following checks.</para>
1295 </listitem>
1296
1297 <listitem>
1298 <para>VirtualBox tests if
1299 <computeroutput>/dev/cdrom</computeroutput> works.</para>
1300 </listitem>
1301
1302 <listitem>
1303 <para>In addition, VirtualBox checks if any CD/DVD drives are
1304 currently mounted by checking
1305 <computeroutput>/etc/mtab</computeroutput>.</para>
1306 </listitem>
1307
1308 <listitem>
1309 <para>In addition, VirtualBox checks if any of the entries in
1310 <computeroutput>/etc/fstab</computeroutput> point to CD/DVD
1311 devices.</para>
1312 </listitem>
1313 </orderedlist></para>
1314
1315 <para>In other words, you can try to set VBOX_CDROM to contain a list of
1316 your CD/DVD devices, separated by colons, for example as follows:</para>
1317
1318 <para><screen>export VBOX_CDROM='/dev/cdrom0:/dev/cdrom1'</screen>On
1319 modern Linux distributions, VirtualBox uses the hardware abstraction
1320 layer (hal) to locate CD and DVD hardware.</para>
1321 </sect2>
1322
1323 <sect2>
1324 <title>Linux host floppy not found</title>
1325
1326 <para>The previous instructions (for CD and DVD drives) apply
1327 accordingly to floppy disks, except that on older distributions
1328 VirtualBox tests for <computeroutput>/dev/fd*</computeroutput> devices
1329 by default, and this can be overridden with the
1330 <computeroutput>VBOX_FLOPPY</computeroutput> environment
1331 variable.</para>
1332 </sect2>
1333
1334 <sect2>
1335 <title>Strange guest IDE error messages when writing to CD/DVD</title>
1336
1337 <para>If the experimental CD/DVD writer support is enabled with an
1338 incorrect VirtualBox, host or guest configuration, it is possible that
1339 any attempt to access the CD/DVD writer fails and simply results in
1340 guest kernel error messages (for Linux guests) or application error
1341 messages (for Windows guests). VirtualBox performs the usual consistency
1342 checks when a VM is powered up (in particular it aborts with an error
1343 message if the device for the CD/DVD writer is not writable by the user
1344 starting the VM), but it cannot detect all misconfigurations. The
1345 necessary host and guest OS configuration is not specific for
1346 VirtualBox, but a few frequent problems are listed here which occurred
1347 in connection with VirtualBox.</para>
1348
1349 <para>Special care must be taken to use the correct device. The
1350 configured host CD/DVD device file name (in most cases
1351 <literal>/dev/cdrom</literal>) must point to the device that allows
1352 writing to the CD/DVD unit. For CD/DVD writer units connected to a SCSI
1353 controller or to a IDE controller that interfaces to the Linux SCSI
1354 subsystem (common for some SATA controllers), this must refer to the
1355 SCSI device node (e.g. <literal>/dev/scd0</literal>). Even for IDE
1356 CD/DVD writer units this must refer to the appropriate SCSI CD-ROM
1357 device node (e.g. <literal>/dev/scd0</literal>) if the
1358 <literal>ide-scsi</literal> kernel module is loaded. This module is
1359 required for CD/DVD writer support with all Linux 2.4 kernels and some
1360 early 2.6 kernels. Many Linux distributions load this module whenever a
1361 CD/DVD writer is detected in the system, even if the kernel would
1362 support CD/DVD writers without the module. VirtualBox supports the use
1363 of IDE device files (e.g. <literal>/dev/hdc</literal>), provided the
1364 kernel supports this and the <literal>ide-scsi</literal> module is not
1365 loaded.</para>
1366
1367 <para>Similar rules (except that within the guest the CD/DVD writer is
1368 always an IDE device) apply to the guest configuration. Since this setup
1369 is very common, it is likely that the default configuration of the guest
1370 works as expected.</para>
1371 </sect2>
1372
1373 <sect2>
1374 <title>VBoxSVC IPC issues</title>
1375
1376 <para>On Linux, VirtualBox makes use of a custom version of Mozilla
1377 XPCOM (cross platform component object model) for inter- and
1378 intra-process communication (IPC). The process
1379 <computeroutput>VBoxSVC</computeroutput> serves as a communication hub
1380 between different VirtualBox processes and maintains the global
1381 configuration, i.e. the XML database. When starting a VirtualBox
1382 component, the processes <computeroutput>VBoxSVC</computeroutput> and
1383 <computeroutput>VBoxXPCOMIPCD</computeroutput> are started
1384 automatically. They are only accessible from the user account they are
1385 running under. <computeroutput>VBoxSVC</computeroutput> owns the
1386 VirtualBox configuration database which normally resides in
1387 <computeroutput>~/.config/VirtualBox</computeroutput>, or the appropriate configuration directory for your operating system. While it is running, the
1388 configuration files are locked. Communication between the various
1389 VirtualBox components and <computeroutput>VBoxSVC</computeroutput> is
1390 performed through a local domain socket residing in
1391 <computeroutput>/tmp/.vbox-&lt;username&gt;-ipc</computeroutput>. In
1392 case there are communication problems (i.e. a VirtualBox application
1393 cannot communicate with <computeroutput>VBoxSVC</computeroutput>),
1394 terminate the daemons and remove the local domain socket
1395 directory.</para>
1396 </sect2>
1397
1398 <sect2 id="ts_usb-linux">
1399 <title>USB not working</title>
1400
1401 <para>If USB is not working on your Linux host, make sure that the
1402 current user is a member of the
1403 <computeroutput>vboxusers</computeroutput> group. On older hosts, you
1404 need to make sure that the user has permission to access the USB
1405 filesystem (<computeroutput>usbfs</computeroutput>), which VirtualBox
1406 relies on to retrieve valid information about your host's USB devices.
1407 The rest of this section only applies to those older systems.</para>
1408
1409 <para>As <computeroutput>usbfs</computeroutput> is a virtual filesystem,
1410 a <computeroutput>chmod</computeroutput> on
1411 <computeroutput>/proc/bus/usb</computeroutput> has no effect. The
1412 permissions for <computeroutput>usbfs</computeroutput> can therefore
1413 <emphasis>only</emphasis> be changed by editing the
1414 <computeroutput>/etc/fstab</computeroutput> file.</para>
1415
1416 <para>For example, most Linux distributions have a user group called
1417 <computeroutput>usb</computeroutput> or similar, of which the current
1418 user must be a member. To give all users of that group access to usbfs,
1419 make sure the following line is present:<screen># 85 is the USB group
1420none /proc/bus/usb usbfs devgid=85,devmode=664 0 0</screen>Replace
1421 85 with the group ID that matches your system (search
1422 <computeroutput>/etc/group</computeroutput> for "usb" or similar).
1423 Alternatively, if you don't mind the security hole, give all users
1424 access to USB by changing "664" to "666".</para>
1425
1426 <para>The various distributions are very creative from which script the
1427 <computeroutput>usbfs</computeroutput> filesystem is mounted. Sometimes
1428 the command is hidden in unexpected places. For SuSE 10.0 the mount
1429 command is part of the udev configuration file
1430 <computeroutput>/etc/udev/rules.d/50-udev.rules</computeroutput>. As
1431 this distribution has no user group called
1432 <computeroutput>usb</computeroutput>, you may e.g. use the
1433 <computeroutput>vboxusers</computeroutput> group which was created by
1434 the VirtualBox installer. Since group numbers are allocated dynamically,
1435 the following example uses 85 as a placeholder. Modify the line
1436 containing (a linebreak has been inserted to improve
1437 readability)<screen>DEVPATH="/module/usbcore", ACTION=="add",
1438 RUN+="/bin/mount -t usbfs usbfs /proc/bus/usb"</screen> and add the
1439 necessary options (make sure that everything is in a single
1440 line):<screen>DEVPATH="/module/usbcore", ACTION=="add",
1441 RUN+="/bin/mount -t usbfs usbfs /proc/bus/usb -o devgid=85,devmode=664"</screen></para>
1442
1443 <para>Debian Etch has the mount command in
1444 <computeroutput>/etc/init.d/mountkernfs.sh</computeroutput>. Since that
1445 distribution has no group <computeroutput>usb</computeroutput>, it is
1446 also the easiest solution to allow all members of the group
1447 <computeroutput>vboxusers</computeroutput> to access the USB subsystem.
1448 Modify the line <screen>domount usbfs usbdevfs /proc/bus/usb -onoexec,nosuid,nodev</screen>
1449 so that it contains <screen>domount usbfs usbdevfs /proc/bus/usb -onoexec,nosuid,nodev,devgid=85,devmode=664</screen>
1450 As usual, replace the 85 with the actual group number which should get
1451 access to USB devices.</para>
1452
1453 <para>Other distributions do similar operations in scripts stored in the
1454 <computeroutput>/etc/init.d</computeroutput> directory.</para>
1455 </sect2>
1456
1457 <sect2>
1458 <title>PAX/grsec kernels</title>
1459
1460 <para>Linux kernels including the grsec patch (see <literal><ulink
1461 url="http://www.grsecurity.net/">http://www.grsecurity.net/</ulink></literal>)
1462 and derivates have to disable PAX_MPROTECT for the VBox binaries to be
1463 able to start a VM. The reason is that VBox has to create executable
1464 code on anonymous memory.</para>
1465 </sect2>
1466
1467 <sect2>
1468 <title>Linux kernel vmalloc pool exhausted</title>
1469
1470 <para>When running a large number of VMs with a lot of RAM on a Linux
1471 system (say 20 VMs with 1GB of RAM each), additional VMs might fail to
1472 start with a kernel error saying that the vmalloc pool is exhausted and
1473 should be extended. The error message also tells you to specify
1474 <computeroutput>vmalloc=256MB</computeroutput> in your kernel parameter
1475 list. If adding this parameter to your GRUB or LILO configuration makes
1476 the kernel fail to boot (with a weird error message such as "failed to
1477 mount the root partition"), then you have probably run into a memory
1478 conflict of your kernel and initial RAM disk. This can be solved by
1479 adding the following parameter to your GRUB configuration:</para>
1480
1481 <screen>uppermem 524288</screen>
1482 </sect2>
1483 </sect1>
1484
1485 <sect1 id="ts_sol-hosts">
1486 <title>Solaris hosts</title>
1487
1488 <sect2>
1489 <title>Cannot start VM, not enough contiguous memory</title>
1490
1491 <para>The ZFS file system is known to use nearly all available RAM as cache if
1492 the default system settings are not changed. This may lead to a heavy
1493 fragmentation of the host memory preventing VirtualBox VMs from being
1494 started. We recommend to limit the ZFS cache by adding a line<screen>set zfs:zfs_arc_max = xxxx</screen>
1495 to /etc/system where <computeroutput>xxxx</computeroutput> bytes is the
1496 amount of memory usable for the ZFS cache.</para>
1497 </sect2>
1498
1499 <sect2>
1500 <title>VM aborts with out of memory errors on Solaris 10 hosts</title>
1501
1502 <para>32-bit Solaris 10 hosts (bug 1225025) require swap space equal to,
1503 or greater than the host's physical memory size. For example, 8 GB
1504 physical memory would require at least 8 GB swap. This can be configured
1505 during a Solaris 10 install by choosing a 'custom install' and changing
1506 the default partitions.</para>
1507
1508 <note>
1509 <para>This restriction applies only to 32-bit Solaris hosts, 64-bit
1510 hosts are not affected!</para>
1511 </note>
1512
1513 <para>For existing Solaris 10 installs, an additional swap image needs
1514 to be mounted and used as swap. Hence if you have 1 GB swap and 8 GB of
1515 physical memory, you require to add 7 GB more swap. This can be done as
1516 follows:</para>
1517
1518 <para>For ZFS (as root user):</para>
1519
1520 <para><screen>zfs create -V 8gb /_&lt;ZFS volume&gt;_/swap
1521swap -a /dev/zvol/dsk/_&lt;ZFS volume&gt;_/swap</screen></para>
1522
1523 <para>To mount if after reboot, add the following line to
1524 /etc/vfstab:</para>
1525
1526 <screen>/dev/zvol/dsk/_&lt;ZFS volume&gt;_/swap - - swap - no -</screen>
1527
1528 <para>Alternatively, you could grow the existing swap using:</para>
1529
1530 <screen>zfs set volsize=8G rpool/swap</screen>
1531
1532 <para>And reboot the system for the changes to take effect.</para>
1533
1534 <para>For UFS (as root user):</para>
1535
1536 <screen>mkfile 7g /path/to/swapfile.img
1537swap -a /path/to/swapfile.img</screen>
1538
1539 <para>To mount it after reboot, add the following line to
1540 /etc/vfstab:</para>
1541
1542 <screen>/path/to/swap.img - - swap - no -</screen>
1543 </sect2>
1544 </sect1>
1545</chapter>
Note: See TracBrowser for help on using the repository browser.

© 2024 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette