source: vbox/trunk/doc/manual/en_US/user_Troubleshooting.xml@ 91099

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE chapter PUBLIC "-//OASIS//DTD DocBook XML V4.5//EN"
"http://www.oasis-open.org/docbook/xml/4.5/docbookx.dtd"[
<!ENTITY % all.entities SYSTEM "all-entities.ent">
%all.entities;
]>
<chapter id="Troubleshooting">

  <title>Troubleshooting</title>

  <para>
    This chapter provides answers to commonly asked questions. In order
    to improve your user experience with &product-name;, it is
    recommended to read this section to learn more about common pitfalls
    and get recommendations on how to use the product.
  </para>

  <sect1 id="ts_procs-tools">

    <title>Procedures and Tools</title>

    <sect2 id="ts_categorize-isolate">

      <title>Categorizing and Isolating Problems</title>

      <para>
        More often than not, a virtualized guest behaves like a physical
        system. Any problems that a physical machine would encounter, a
        virtual machine will encounter as well. If, for example,
        Internet connectivity is lost due to external issues, virtual
        machines will be affected just as much as physical ones.
      </para>

      <para>
        If a true &product-name; problem is encountered, it helps to
        categorize and isolate the problem first. Here are some of the
        questions that should be answered before reporting a problem:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            Is the problem specific to a certain guest OS? Or a specific
            release of a guest OS? Especially with Linux guest related
            problems, the issue may be specific to a certain
            distribution and version of Linux.
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem specific to a certain host OS? Problems are
            usually not host OS specific, because most of the
            &product-name; code base is shared across all supported
            platforms, but especially in the areas of networking and USB
            support, there are significant differences between host
            platforms. Some GUI related issues are also host specific.
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem specific to certain host hardware? This
            category of issues is typically related to the host CPU.
            Because of significant differences between VT-x and AMD-V,
            problems may be specific to one or the other technology. The
            exact CPU model may also make a difference because different
            CPUs support different features, which may affect certain
            aspects of guest CPU operation.
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem specific to guest SMP? That is, is it related
            to the number of virtual CPUs (VCPUs) in the guest? Using
            more than one CPU usually significantly affects the internal
            operation of a guest OS.
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem specific to the Guest Additions? In some
            cases, this is obvious, such as a shared folders problem. In
            other cases such as display problems, it may be less
            obvious. If the problem is Guest Additions specific, is it
            also specific to a certain version of the Guest Additions?
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem specific to a certain environment? Some
            problems are related to a particular environment external to
            the VM. This usually involves network setup. Certain
            configurations of external servers such as DHCP or PXE may
            expose problems which do not occur with other, similar
            servers.
          </para>
        </listitem>

        <listitem>
          <para>
            Is the problem a regression? Knowing that an issue is a
            regression usually makes it significantly easier to find the
            solution. In this case, it is crucial to know which version
            is affected and which is not.
          </para>
        </listitem>

      </itemizedlist>

    </sect2>

    <sect2 id="collect-debug-info">

      <title>Collecting Debugging Information</title>

      <para>
        For problem determination, it is often important to collect
        debugging information which can be analyzed by &product-name;
        support. This section contains information about what kind of
        information can be obtained.
      </para>

      <para>
        Every time &product-name; starts up a VM, a so-called
        <emphasis>release log file</emphasis> is created, containing
        lots of information about the VM configuration and runtime
        events. The log file is called <filename>VBox.log</filename> and
        resides in the VM log file folder, which is
        <filename>$HOME/VirtualBox
        VMs/<replaceable>VM-name</replaceable>/Logs</filename> by
        default.
      </para>

      <para>
        When starting a VM, the configuration file of the last run will
        be renamed to <filename>.1</filename>, up to
        <filename>.3</filename>. Sometimes when there is a problem, it
        is useful to have a look at the logs. Also when requesting
        support for &product-name;, supplying the corresponding log file
        is mandatory.
      </para>

      <para>
        For convenience, for each virtual machine, the VirtualBox
        Manager window can show these logs in a window. To access it,
        select a virtual machine from the list on the left and select
        <emphasis role="bold">Show Log</emphasis> from the
        <emphasis role="bold">Machine</emphasis> menu.
      </para>

      <para>
        The release log file, <filename>VBox.log</filename>, contains a
        wealth of diagnostic information, such as Host OS type and
        version, &product-name; version and build. It also includes a
        complete dump of the guest's configuration (CFGM), detailed
        information about the host CPU type and supported features,
        whether hardware virtualization is enabled, information about
        VT-x/AMD-V setup, state transitions (such as creating, running,
        paused, stopping), guest BIOS messages, Guest Additions
        messages, device-specific log entries and, at the end of
        execution, final guest state and condensed statistics.
      </para>

      <para>
        In case of crashes, it is very important to collect
        <emphasis>crash dumps</emphasis>. This is true for both host and
        guest crashes. For information about enabling core dumps on
        Linux, Oracle Solaris, and Mac OS X systems, refer to the
        following core dump article on the &product-name; website:
      </para>

      <para>
        <ulink url="http://www.virtualbox.org/wiki/Core_dump" />.
      </para>

      <para>
        You can also use <command>VBoxManage debugvm</command> to create
        a dump of a complete virtual machine. See
        <xref linkend="vboxmanage-debugvm" />.
      </para>

      <para>
        For network related problems, it is often helpful to capture a
        trace of network traffic. If the traffic is routed through an
        adapter on the host, it is possible to use Wireshark or a
        similar tool to capture the traffic there. However, this often
        also includes a lot of traffic unrelated to the VM.
      </para>

      <para>
        &product-name; provides an ability to capture network traffic
        only on a specific VM's network adapter. Refer to the following
        network tracing article on the &product-name; website for
        information on enabling this capture:
      </para>

      <para>
        <ulink url="http://www.virtualbox.org/wiki/Network_tips" />.
      </para>

      <para>
        The trace files created by &product-name; are in
        <filename>.pcap</filename> format and can be easily analyzed
        with Wireshark.
      </para>

    </sect2>

    <sect2 id="ts_vboxbugreport">

      <title>Using the VBoxBugReport Command to Collect Debug Information
        Automatically</title>

      <para>
        The <command>VBoxBugReport</command> command is used to collect
        debug information automatically for an &product-name;
        installation. This command can be useful when you need to gather
        information to send to Oracle Support.
      </para>

      <para>
        The following examples show how to use
        <command>VBoxBugReport</command>.
      </para>

      <para>
        By default, the command collects <command>VBoxSVC</command>
        process logs, device settings, and global configuration data for
        an &product-name; host.
      </para>

<screen>$ VBoxBugReport
  ...
  0% - collecting VBoxSVC.log.10...
  7% - collecting VBoxSVC.log.9...
  ...
 64% - collecting VBoxSVC.log.1...
 71% - collecting VBoxSVC.log...
 78% - collecting VirtualBox.xml...
 85% - collecting HostUsbDevices...
 92% - collecting HostUsbFilters...
100% - compressing...

Report was written to '2019-03-26-13-32-02-bugreport.tgz'</screen>

      <para>
        The results are saved as a compressed tar file archive in the
        same directory where the command is run.
      </para>

      <para>
        To specify a different output file location:
      </para>

<screen>$ VBoxBugReport --output ~/debug/bug004.tgz</screen>

      <para>
        To output all debug information to a single text file, rather
        than a <filename>tgz</filename> file:
      </para>

<screen>$ VBoxBugReport --text</screen>

      <para>
        To collect information for a specific VM, called
        <literal>Windows_10</literal>:
      </para>

<screen>$ VBoxBugReport Windows_10</screen>

      <para>
        This command collects machine settings, guest properties, and
        log files for the specified VM. Global configuration information
        for the host is also included.
      </para>

      <para>
        To collect information for several VMs, called
        <literal>Windows_7</literal>, <literal>Windows_8</literal>, and
        <literal>Windows_10</literal>:
      </para>

<screen>$ VBoxBugReport Windows_7 Windows_8 Windows_10</screen>

      <para>
        To collect information for all VMs:
      </para>

<screen>$ VBoxBugReport --all</screen>

      <para>
        To show a full list of the available command options, run
        <command>VBoxBugReport --help</command>.
      </para>

    </sect2>

    <sect2 id="ts_debugger">

      <title>The Built-In VM Debugger</title>

      <para>
        &product-name; includes a built-in VM debugger, which advanced
        users may find useful. This debugger enables you to examine and,
        to some extent, control the VM state.
      </para>

      <warning>
        <para>
          Use the VM debugger at your own risk. There is no support for
          it, and the following documentation is only made available for
          advanced users with a very high level of familiarity with the
          x86/AMD64 machine instruction set, as well as detailed
          knowledge of the PC architecture. A degree of familiarity with
          the internals of the guest OS in question may also be very
          helpful.
        </para>
      </warning>

      <para>
        The VM debugger is available in all regular production versions
        of &product-name;, but it is disabled by default because the
        average user will have little use for it. There are two ways to
        access the debugger:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            Using a debugger console window displayed alongside the VM
          </para>
        </listitem>

        <listitem>
          <para>
            Using the <command>telnet</command> protocol on port 5000
          </para>
        </listitem>

      </itemizedlist>

      <para>
        The debugger can be enabled in the following ways:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            Start the VM directly using <command>VirtualBoxVM
            --startvm</command>, with an additional
            <option>--dbg</option>, <option>--debug</option>, or
            <option>--debug-command-line</option> argument. See the
            <command>VirtualBoxVM --help</command> command usage help
            for details.
          </para>
        </listitem>

        <listitem>
          <para>
            Set the <literal>VBOX_GUI_DBG_ENABLED</literal> or
            <literal>VBOX_GUI_DBG_AUTO_SHOW</literal> environment
            variable to <literal>true</literal> before launching the
            &product-name; process. Setting these variables, only their
            presence is checked, is effective even when the first
            &product-name; process is the VM selector window. VMs
            subsequently launched from the selector will have the
            debugger enabled.
          </para>
        </listitem>

        <listitem>
          <para>
            Set the <literal>GUI/Dbg/Enabled</literal> extra data item
            to <literal>true</literal> before launching the VM. This can
            be set globally or on a per VM basis.
          </para>
        </listitem>

      </itemizedlist>

      <para>
        A new <emphasis role="bold">Debug</emphasis> menu entry is added
        to the &product-name; application. This menu enables the user to
        open the debugger console.
      </para>

      <para>
        The VM debugger command syntax is loosely modeled on Microsoft
        and IBM debuggers used on DOS, OS/2, and Windows. Users familiar
        with symdeb, CodeView, or the OS/2 kernel debugger will find the
        &product-name; VM debugger familiar.
      </para>

      <para>
        The most important command is <command>help</command>. This will
        print brief usage help for all debugger commands. The set of
        commands supported by the VM debugger changes frequently and the
        <command>help</command> command is always up-to-date.
      </para>

      <para>
        A brief summary of frequently used commands is as follows:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            <command>stop</command>: Stops the VM execution and enables
            single stepping
          </para>
        </listitem>

        <listitem>
          <para>
            <command>g</command>: Continue VM execution
          </para>
        </listitem>

        <listitem>
          <para>
            <command>t</command>: Single step an instruction
          </para>
        </listitem>

        <listitem>
          <para>
            <command>rg</command>, <command>rh</command>, and
            <command>r</command>: Print the guest, hypervisor, and
            current registers
          </para>
        </listitem>

        <listitem>
          <para>
            <command>kg</command>, <command>kh</command>, and
            <command>k</command>: Print the guest, hypervisor, and
            current call stack
          </para>
        </listitem>

        <listitem>
          <para>
            <command>da</command>, <command>db</command>,
            <command>dw</command>, <command>dd</command>,
            <command>dq</command>: Print memory contents as ASCII,
            bytes, words, dwords, and qwords
          </para>
        </listitem>

        <listitem>
          <para>
            <command>u</command>: Unassemble memory
          </para>
        </listitem>

        <listitem>
          <para>
            <command>dg</command>: Print the guest's GDT
          </para>
        </listitem>

        <listitem>
          <para>
            <command>di</command>: Print the guest's IDT
          </para>
        </listitem>

        <listitem>
          <para>
            <command>dl</command>: Print the guest's LDT
          </para>
        </listitem>

        <listitem>
          <para>
            <command>dt</command>: Print the guest's TSS
          </para>
        </listitem>

        <listitem>
          <para>
            <command>dp*</command>: Print the guest's page table
            structures
          </para>
        </listitem>

        <listitem>
          <para>
            <command>bp</command> and <command>br</command>: Set a
            normal and recompiler breakpoint
          </para>
        </listitem>

        <listitem>
          <para>
            <command>bl</command>: List breakpoints
          </para>
        </listitem>

        <listitem>
          <para>
            <command>bc</command>: Clear a breakpoint
          </para>
        </listitem>

        <listitem>
          <para>
            <command>writecore</command>: Write a VM core file to disk.
            See <xref linkend="ts_guest-core-format" />
          </para>
        </listitem>

      </itemizedlist>

      <para>
        See the built-in <command>help</command> for other available
        commands.
      </para>

      <para>
        The VM debugger supports symbolic debugging, although symbols
        for guest code are often not available. For Oracle Solaris
        guests, the <command>detect</command> command automatically
        determines the guest OS version and locates kernel symbols in
        guest's memory. Symbolic debugging is then available. For Linux
        guests, the <command>detect</command> commands also determines
        the guest OS version, but there are no symbols in the guest's
        memory. Kernel symbols are available in the file
        <filename>/proc/kallsyms</filename> on Linux guests. This file
        must be copied to the host, for example using
        <command>scp</command>. The <command>loadmap</command> debugger
        command can be used to make the symbol information available to
        the VM debugger. Note that the <filename>kallsyms</filename>
        file contains the symbols for the currently loaded modules. If
        the guest's configuration changes, the symbols will change as
        well and must be updated.
      </para>

      <para>
        For all guests, a simple way to verify that the correct symbols
        are loaded is the <command>k</command> command. The guest is
        normally idling and it should be clear from the symbolic
        information that the guest operating system's idle loop is being
        executed.
      </para>

      <para>
        Another group of debugger commands is the set of
        <command>info</command> commands. Running <command>info
        help</command> provides complete usage information. The
        information commands provide ad-hoc data pertinent to various
        emulated devices and aspects of the VMM. There is no general
        guideline for using the <command>info</command> commands, the
        right command to use depends entirely on the problem being
        investigated. Some of the <command>info</command> commands are
        as follows:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            <command>cfgm</command>: Print a branch of the configuration
            tree
          </para>
        </listitem>

        <listitem>
          <para>
            <command>cpuid</command>: Display the guest CPUID leaves
          </para>
        </listitem>

        <listitem>
          <para>
            <command>ioport</command>: Print registered I/O port ranges
          </para>
        </listitem>

        <listitem>
          <para>
            <command>mmio</command>: Print registered MMIO ranges
          </para>
        </listitem>

        <listitem>
          <para>
            <command>mode</command>: Print the current paging mode
          </para>
        </listitem>

        <listitem>
          <para>
            <command>pit</command>: Print the i8254 PIT state
          </para>
        </listitem>

        <listitem>
          <para>
            <command>pic</command>: Print the i8259A PIC state
          </para>
        </listitem>

        <listitem>
          <para>
            <command>ohci</command>, <command>ehci</command>,
            <command>xhci</command>: Print a subset of the OHCI, EHCI,
            and xHCI USB controller state
          </para>
        </listitem>

        <listitem>
          <para>
            <command>pcnet0</command>: Print the PCnet state
          </para>
        </listitem>

        <listitem>
          <para>
            <command>vgatext</command>: Print the contents of the VGA
            framebuffer formatted as standard text mode
          </para>
        </listitem>

        <listitem>
          <para>
            <command>timers</command>: Print all VM timers
          </para>
        </listitem>

      </itemizedlist>

      <para>
        The output of the <command>info</command> commands generally
        requires in-depth knowledge of the emulated device or
        &product-name; VMM internals. However, when used properly, the
        information provided can be invaluable.
      </para>

    </sect2>

    <sect2 id="ts_guest-core-format">

      <title>VM Core Format</title>

      <para>
        &product-name; uses the 64-bit ELF format for its VM core files
        created by <command>VBoxManage debugvm</command>, see
        <xref linkend="vboxmanage-debugvm" />. The VM core file contain
        the memory and CPU dumps of the VM and can be useful for
        debugging your guest OS. The 64-bit ELF object format
        specification can be obtained at:
      </para>

      <para>
        <ulink url="http://downloads.openwatcom.org/ftp/devel/docs/elf-64-gen.pdf" />.
      </para>

      <para>
        The overall layout of the VM core format is as follows:
      </para>

<screen>[ ELF 64 Header]
[ Program Header, type PT_NOTE ]
  &rarr; offset to COREDESCRIPTOR
[ Program Header, type PT_LOAD ] - one for each contiguous physical memory range
  &rarr; Memory offset of range
  &rarr; File offset
[ Note Header, type NT_VBOXCORE ]
[ COREDESCRIPTOR ]
  &rarr; Magic
  &rarr; VM core file version
  &rarr; VBox version
  &rarr; Number of vCPUs etc.
[ Note Header, type NT_VBOXCPU ] - one for each vCPU
[ vCPU 1 Note Header ]
  [ DBGFCORECPU - vCPU 1 dump ]
[ Additional Notes + Data ] - currently unused
[ Memory dump ]</screen>

      <para>
        The memory descriptors contain physical addresses relative to
        the guest and not virtual addresses. Regions of memory such as
        MMIO regions are not included in the core file.
      </para>

      <para>
        The relevant data structures and definitions can be found in the
        &product-name; sources under the following header files:
        <filename>include/VBox/dbgfcorefmt.h</filename>,
        <filename>include/iprt/x86.h</filename> and
        <filename>src/VBox/Runtime/include/internal/ldrELFCommon.h</filename>.
      </para>

      <para>
        The VM core file can be inspected using
        <command>elfdump</command> and GNU <command>readelf</command> or
        other similar utilities.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_general">

    <title>General Troubleshooting</title>

    <sect2 id="ts_config-periodic-flush">

      <title>Guest Shows IDE/SATA Errors for File-Based Images on Slow Host File
        System</title>

      <para>
        Occasionally, some host file systems provide very poor writing
        performance and as a consequence cause the guest to time out
        IDE/SATA commands. This is normal behavior and should normally
        cause no real problems, as the guest should repeat commands that
        have timed out. However, guests such as some Linux versions have
        severe problems if a write to an image file takes longer than
        about 15 seconds. Some file systems however require more than a
        minute to complete a single write, if the host cache contains a
        large amount of data that needs to be written.
      </para>

      <para>
        The symptom for this problem is that the guest can no longer
        access its files during large write or copying operations,
        usually leading to an immediate hang of the guest.
      </para>

      <para>
        In order to work around this problem, the true fix is to use a
        faster file system that does not exhibit such unacceptable write
        performance, it is possible to flush the image file after a
        certain amount of data has been written. This interval is
        normally infinite, but can be configured individually for each
        disk of a VM.
      </para>

      <para>
        For IDE disks use the following command:
      </para>

<screen>VBoxManage setextradata <replaceable>VM-name</replaceable>
"VBoxInternal/Devices/piix3ide/0/LUN#[<replaceable>x</replaceable>]/Config/FlushInterval" [<replaceable>b</replaceable>]</screen>

      <para>
        For SATA disks use the following command:
      </para>

<screen>VBoxManage setextradata <replaceable>VM-name</replaceable>
"VBoxInternal/Devices/ahci/0/LUN#[<replaceable>x</replaceable>]/Config/FlushInterval" [<replaceable>b</replaceable>]</screen>

      <para>
        <literal>[<replaceable>x</replaceable>]</literal> specifies the
        disk. For IDE, <literal>0</literal> represents device 0 on the
        primary channel, <literal>1</literal> represents device 1 on the
        primary channel, <literal>2</literal> represents device 0 on the
        secondary channel, and <literal>3</literal> represents device 1
        on the secondary channel. For SATA, use values between
        <literal>0</literal> and <literal>29</literal>. This
        configuration option applies to disks only. Do not use this
        option for CD or DVD drives.
      </para>

      <para>
        The unit of the interval
        (<literal>[<replaceable>b</replaceable>]</literal>) is the
        number of bytes written since the last flush. The value for it
        must be selected so that the occasional long write delays do not
        occur. Since the proper flush interval depends on the
        performance of the host and the host filesystem, finding the
        optimal value that makes the problem disappear requires some
        experimentation. Values between 1000000 and 10000000 (1 to 10
        megabytes) are a good starting point. Decreasing the interval
        both decreases the probability of the problem and the write
        performance of the guest. Setting the value unnecessarily low
        will cost performance without providing any benefits. An
        interval of 1 will cause a flush for each write operation and
        should solve the problem in any case, but has a severe write
        performance penalty.
      </para>

      <para>
        Providing a value of <literal>0</literal> for
        <literal>[<replaceable>b</replaceable>]</literal> is treated as
        an infinite flush interval, effectively disabling this
        workaround. Removing the extra data key by specifying no value
        for <literal>[<replaceable>b</replaceable>]</literal> has the
        same effect.
      </para>

    </sect2>

    <sect2 id="ts_ide-sata-flush">

      <title>Responding to Guest IDE/SATA Flush Requests</title>

      <para>
        If desired, the virtual disk images can be flushed when the
        guest issues the IDE FLUSH CACHE command. Normally these
        requests are ignored for improved performance. The parameters
        below are only accepted for disk drives. They must not be set
        for DVD drives.
      </para>

      <para>
        To enable flushing for IDE disks, issue the following command:
      </para>

<screen>$ VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxInternal/Devices/piix3ide/0/LUN#[<replaceable>x</replaceable>]/Config/IgnoreFlush" 0</screen>

      <para>
        <literal>[<replaceable>x</replaceable>]</literal> specifies the
        disk. Enter <literal>0</literal> for device 0 on the primary
        channel, <literal>1</literal> for device 1 on the primary
        channel, <literal>2</literal> for device 0 on the secondary
        channel, or <literal>3</literal> for device 1 on the secondary
        channel.
      </para>

      <para>
        To enable flushing for SATA disks, issue the following command:
      </para>

<screen>$ VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxInternal/Devices/ahci/0/LUN#[x]/Config/IgnoreFlush" 0</screen>

      <para>
        The value [x] that selects the disk can be a value between 0 and
        29.
      </para>

      <para>
        Note that this does not affect the flushes performed according
        to the configuration described in
        <xref linkend="ts_config-periodic-flush"/>. Restoring the
        default of ignoring flush commands is possible by setting the
        value to 1 or by removing the key.
      </para>

    </sect2>

    <sect2 id="ts_host-freq-boost">

      <title>Performance Variation with Frequency Boosting</title>

      <para>
        Many multicore processors support some form of frequency
        boosting, which means that if only one core is utilized, it can
        run possibly 50% faster or even more than the rated CPU
        frequency. This causes measured performance to vary somewhat as
        a function of the momentary overall system load. The exact
        behavior depends strongly on the specific processor model.
      </para>

      <para>
        As a consequence, benchmarking on systems which utilize
        frequency boosting may produce unstable and non-repeatable
        results. This is especially true if benchmark runs are short, of
        the order of seconds. To obtain stable results, benchmarks must
        be run over longer periods of time and with a constant system
        load apart from the VM being tested.
      </para>

    </sect2>

    <sect2 id="ts_host-freq-scaling">

      <title>Frequency Scaling Effect on CPU Usage</title>

      <para>
        On some hardware platforms and operating systems, CPU frequency
        scaling may cause CPU usage reporting to be highly misleading.
        This happens in situations when the host CPU load is significant
        but not heavy, such as between 15% to 30% of the maximum.
      </para>

      <para>
        Most operating systems determine CPU usage in terms of time
        spent, measuring for example how many nanoseconds the systems or
        a process was active within one second. However, in order to
        save energy, systems can significantly scale down CPU speed when
        the system is not fully loaded. When the CPU is running at for
        example one half of its maximum speed, the same number of
        instructions will take roughly twice as long to execute compared
        to running at full speed.
      </para>

      <para>
        Depending on the specific hardware and host OS, this effect can
        very significantly skew the CPU usage reported by the OS. The
        reported CPU usage can be several times higher than what it
        would have been had the CPU been running at full speed. The
        effect can be observed both on the host OS and in a guest OS.
      </para>

    </sect2>

    <sect2 id="ts_win-cpu-usage-rept">

      <title>Inaccurate Windows CPU Usage Reporting</title>

      <para>
        CPU usage reporting tools which come with Windows, such as Task
        Manager or Resource Monitor, do not take the time spent
        processing hardware interrupts into account. If the interrupt
        load is heavy, with thousands of interrupts per second, CPU
        usage may be significantly underreported.
      </para>

      <para>
        This problem affects Windows as both host and guest OS.
        Sysinternals tools, such as Process Explorer, do not suffer from
        this problem.
      </para>

    </sect2>

    <sect2 id="ts_host-powermgmt">

      <title>Poor Performance Caused by Host Power Management</title>

      <para>
        On some hardware platforms and operating systems, virtualization
        performance is negatively affected by host CPU power management.
        The symptoms may be choppy audio in the guest or erratic guest
        clock behavior.
      </para>

      <para>
        Some of the problems may be caused by firmware and/or host
        operating system bugs. Therefore, updating the firmware and
        applying operating systems fixes is recommended.
      </para>

      <para>
        For optimal virtualization performance, the C1E power state
        support in the system's BIOS should be disabled, if such a
        setting is available. Not all systems support the C1E power
        state. On Intel systems, the <literal>Intel C State</literal>
        setting should be disabled. Disabling other power management
        settings may also improve performance. However, a balance
        between performance and power consumption must always be
        considered.
      </para>

    </sect2>

    <sect2 id="ts_gui-2d-grayed-out">

      <title>GUI: 2D Video Acceleration Option is Grayed Out</title>

      <para>
        To use 2D Video Acceleration within &product-name;, your host's
        video card should support certain OpenGL extensions. On startup,
        &product-name; checks for those extensions, and, if the test
        fails, this option is silently grayed out.
      </para>

      <para>
        To find out why it has failed, you can manually execute the
        following command:
      </para>

<screen>$ VBoxTestOGL --log "log_file_name" --test 2D</screen>

      <para>
        It will list the required OpenGL extensions one by one and will
        show you which one failed the test. This usually means that you
        are running an outdated or misconfigured OpenGL driver on your
        host. It can also mean that your video chip is lacking required
        functionality.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_win-guests">

    <title>Windows Guests</title>

    <sect2 id="ts_win7-guest-usb3-support">

      <title>No USB 3.0 Support in Windows 7 Guests</title>

      <para>
        If a Windows 7 or Windows Server 2008 R2 guest is configured for
        USB 3.0 (xHCI) support, the guest OS will not have any USB
        support at all. This happens because Windows 7 predates USB 3.0
        and therefore does not ship with any xHCI drivers. Microsoft
        also does not offer any vendor-provided xHCI drivers through
        Windows Update.
      </para>

      <para>
        To solve this problem, it is necessary to download and install
        the Intel xHCI driver in the guest. Intel offers the driver as
        the USB 3.0 eXtensible Host Controller (xHCI) driver for Intel 7
        Series/C216 chipsets.
      </para>

      <para>
        Note that the driver only supports Windows 7 and Windows Server
        2008 R2. The driver package includes support for both 32-bit and
        64-bit OS variants.
      </para>

    </sect2>

    <sect2 id="ts_win-guest-bluescreen">

      <title>Windows Bluescreens After Changing VM Configuration</title>

      <para>
        Changing certain virtual machine settings can cause Windows
        guests to fail during start up with a bluescreen. This may
        happen if you change VM settings after installing Windows, or if
        you copy a disk image with an already installed Windows to a
        newly created VM which has settings that differ from the
        original machine.
      </para>

      <para>
        This applies in particular to the following settings:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            The ACPI and I/O APIC settings should never be changed after
            installing Windows. Depending on the presence of these
            hardware features, the Windows installation program chooses
            special kernel and device driver versions and will fail to
            startup should these hardware features be removed. Enabling
            them for a Windows VM which was installed without them does
            not cause any harm. However, Windows will not use these
            features in this case.
          </para>
        </listitem>

        <listitem>
          <para>
            Changing the storage controller hardware will cause bootup
            failures as well. This might also apply to you if you copy a
            disk image from an older version of &product-name; to a new
            virtual machine. The default subtype of IDE controller
            hardware used by &product-name; is PIIX4. Make sure that the
            storage controller settings are identical.
          </para>
        </listitem>

      </itemizedlist>

    </sect2>

    <sect2 id="ts_win-guest-bluescreen-smp">

      <title>Windows 0x101 Bluescreens with SMP Enabled (IPI Timeout)</title>

      <para>
        If a VM is configured to have more than one processor
        (symmetrical multiprocessing, SMP), some configurations of
        Windows guests crash with an 0x101 error message, indicating a
        timeout for interprocessor interrupts (IPIs). These interrupts
        synchronize memory management between processors.
      </para>

      <para>
        According to Microsoft, this is due to a race condition in
        Windows. A hotfix is available from Microsoft.
      </para>

      <para>
        If this does not help, please reduce the number of virtual
        processors to 1.
      </para>

    </sect2>

    <sect2 id="ts_win2k-guest-install">

      <title>Windows 2000 Installation Failures</title>

      <para>
        When installing Windows 2000 guests, you might run into one of
        the following issues:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            Installation reboots, usually during component registration.
          </para>
        </listitem>

        <listitem>
          <para>
            Installation fills the whole hard disk with empty log files.
          </para>
        </listitem>

        <listitem>
          <para>
            Installation complains about a failure installing
            <filename>msgina.dll</filename>.
          </para>
        </listitem>

      </itemizedlist>

      <para>
        These problems are all caused by a bug in the hard disk driver
        of Windows 2000. After issuing a hard disk request, there is a
        race condition in the Windows driver code which leads to
        corruption if the operation completes too fast. For example, the
        hardware interrupt from the IDE controller arrives too soon.
        With physical hardware, there is a guaranteed delay in most
        systems so the problem is usually hidden there. However, it
        should be possible to also reproduce it on physical hardware. In
        a virtual environment, it is possible for the operation to be
        done immediately, especially on very fast systems with multiple
        CPUs, and the interrupt is signaled sooner than on a physical
        system. The solution is to introduce an artificial delay before
        delivering such interrupts. This delay can be configured for a
        VM using the following command:
      </para>

<screen>$ VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxInternal/Devices/piix3ide/0/Config/IRQDelay" 1</screen>

      <para>
        This sets the delay to one millisecond. In case this does not
        help, increase it to a value between 1 and 5 milliseconds.
        Please note that this slows down disk performance. After
        installation, you should be able to remove the key, or set it to
        0.
      </para>

    </sect2>

    <sect2 id="ts_win-guest-bluescreen-record-info">

      <title>How to Record Bluescreen Information from Windows Guests</title>

      <para>
        When Windows guests run into a kernel crash, they display a
        bluescreen error. Depending on how Windows is configured, the
        information will remain on the screen until the machine is
        restarted or it will reboot automatically. During installation,
        Windows is usually configured to reboot automatically. With
        automatic reboots, there is no chance to record the bluescreen
        information which might be important for problem determination.
      </para>

      <para>
        &product-name; provides a method of halting a guest when it
        wants to perform a reset. In order to enable this feature, use
        the following command:
      </para>

<screen>$ VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxInternal/PDM/HaltOnReset" 1</screen>

    </sect2>

    <sect2 id="ts_win-vista-guest-networking">

      <title>No Networking in Windows Vista Guests</title>

      <para>
        With Windows Vista, Microsoft dropped support for the AMD PCNet
        card that legacy versions of &product-name; used to provide as
        the default virtual network card. For Windows Vista guests,
        &product-name; now uses an Intel E1000 card by default.
      </para>

      <para>
        If, for some reason, you still want to use the AMD card, you
        need to download the PCNet driver from the AMD website. This
        driver is available for 32-bit Windows only. You can transfer it
        into the virtual machine using a shared folder. See
        <xref linkend="sharedfolders" />.
      </para>

    </sect2>

    <sect2 id="ts_win-guest-high-cpu">

      <title>Windows Guests may Cause a High CPU Load</title>

      <para>
        Several background applications of Windows guests, especially
        virus scanners, are known to increase the CPU load notably even
        if the guest appears to be idle. We recommend to deactivate
        virus scanners within virtualized guests if possible.
      </para>

    </sect2>

    <sect2 id="ts_win-guest-shared-folders-access-delay">

      <title>Long Delays When Accessing Shared Folders</title>

      <para>
        The performance for accesses to shared folders from a Windows
        guest might be decreased due to delays during the resolution of
        the &product-name; shared folders name service. To fix these
        delays, add the following entries to the file
        <filename>\windows\system32\drivers\etc\lmhosts</filename> of
        the Windows guest:
      </para>

<screen>255.255.255.255        VBOXSVR #PRE
255.255.255.255        VBOXSRV #PRE</screen>

      <para>
        After doing this change, a reboot of the guest is required.
      </para>

    </sect2>

    <sect2 id="ts_win98-guest-usb-tablet-coordinates">

      <title>USB Tablet Coordinates Wrong in Windows 98 Guests</title>

      <para>
        If a Windows 98 VM is configured to use the emulated USB tablet
        (absolute pointing device), the coordinate translation may be
        incorrect and the pointer is restricted to the upper left
        quarter of the guest's screen.
      </para>

      <para>
        The USB HID (Human Interface Device) drivers in Windows 98 are
        very old and do not handle tablets in the same way as modern
        operating systems do. To work around the problem, use the
        following command:
      </para>

<screen>$ VBoxManage setextradata <replaceable>VM-name</replaceable> "VBoxInternal/USB/HidMouse/0/Config/CoordShift" 0</screen>

      <para>
        To restore the default behavior, remove the key or set its value
        to 1.
      </para>

    </sect2>

    <sect2 id="ts_win-guest-active-dir-domain">

      <title>Windows Guests are Removed From an Active Directory Domain After
        Restoring a Snapshot</title>

      <para>
        If a Windows guest is a member of an Active Directory domain and
        the snapshot feature of &product-name; is used, it could be
        removed from the Active Direcory domain after you restore an
        older snapshot.
      </para>

      <para>
        This is caused by automatic machine password changes performed
        by Windows at regular intervals for security purposes. You can
        disable this feature as shown in the following article from
        Microsoft:
        <ulink url="http://support.microsoft.com/kb/154501" />.
      </para>

    </sect2>

    <sect2 id="ts_win31-ram-limitations">

      <title>Windows 3.x Limited to 64 MB RAM</title>

      <para>
        Windows 3.x guests are typically limited to 64 MB RAM, even if a
        VM is assigned much more memory. While Windows 3.1 is
        theoretically capable of using up to 512 MB RAM, it only uses
        memory available through the XMS interface. Versions of
        HIMEM.SYS, the Microsoft XMS manager, shipped with MS-DOS and
        Microsoft Windows 3.x can only use up to 64 MB on standard PCs.
      </para>

      <para>
        This is a known HIMEM.SYS limitation. Windows 3.1 memory limits
        are described in detail in Microsoft Knowledge base article KB
        84388.
      </para>

      <para>
        It is possible for Windows 3.x guests to utilize more than 64 MB
        RAM if a different XMS provider is used. That could be a newer
        HIMEM.SYS version, such as that shipped with Windows 98, or a
        more capable third-party memory manager, such as QEMM.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_lin-x11-guests">

    <title>Linux and X11 Guests</title>

    <sect2 id="ts_linux-guest-high-cpu">

      <title>Linux Guests May Cause a High CPU load</title>

      <para>
        Some Linux guests may cause a high CPU load even if the guest
        system appears to be idle. This can be caused by a high timer
        frequency of the guest kernel. Some Linux distributions, for
        example Fedora, ship a Linux kernel configured for a timer
        frequency of 1000Hz. We recommend to recompile the guest kernel
        and to select a timer frequency of 100Hz.
      </para>

      <para>
        Linux kernels shipped with Red Hat Enterprise Linux, as well as
        kernels of related Linux distributions, such as CentOS and
        Oracle Linux, support a kernel parameter
        <emphasis>divider=N</emphasis>. Hence, such kernels support a
        lower timer frequency without recompilation. We suggest you add
        the kernel parameter <emphasis>divider=10</emphasis> to select a
        guest kernel timer frequency of 100Hz.
      </para>

    </sect2>

    <sect2 id="ts_linux-buggy">

      <title>Buggy Linux 2.6 Kernel Versions</title>

      <para>
        The following bugs in Linux kernels prevent them from executing
        correctly in &product-name;, causing VM boot crashes:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            The Linux kernel version 2.6.18, and some 2.6.17 versions,
            introduced a race condition that can cause boot crashes in
            &product-name;. Please use a kernel version 2.6.19 or later.
          </para>
        </listitem>

        <listitem>
          <para>
            With hardware virtualization and the I/O APIC enabled,
            kernels before 2.6.24-rc6 may panic on boot with the
            following message:
          </para>

<screen>Kernel panic - not syncing: IO-APIC + timer doesn't work!  Boot with
apic=debug and send a report.  Then try booting with the 'noapic' option</screen>

          <para>
            If you see this message, either disable hardware
            virtualization or the I/O APIC as described in
            <xref linkend="settings-system" />, or upgrade the guest to
            a newer kernel.
          </para>

          <para>
            See
            <ulink url="http://www.mail-archive.com/[email protected]/msg30813.html" />
            for details about the kernel fix.
          </para>
        </listitem>

      </itemizedlist>

    </sect2>

    <sect2 id="ts_linux-guest-x11-services">

      <title>Shared Clipboard, Auto-Resizing, and Seamless Desktop in X11 Guests</title>

      <para>
        Guest desktop services in guests running the X11 window system
        such as Oracle Solaris and Linux, are provided by a guest
        service called <command>VBoxClient</command>, which runs under
        the ID of the user who started the desktop session and is
        automatically started using the following command lines when
        your X11 user session is started if you are using a common
        desktop environment such as Gnome or KDE.
      </para>

<screen>$ VBoxClient --clipboard
$ VBoxClient --display
$ VBoxClient --seamless</screen>

      <para>
        If a particular desktop service is not working correctly, it is
        worth checking whether the process which should provide it is
        running.
      </para>

      <para>
        The <command>VBoxClient</command> processes create files in the
        user's home directory with names of the form
        <filename>.vboxclient-*.pid</filename> when they are running in
        order to prevent a given service from being started twice. It
        can happen due to misconfiguration that these files are created
        owned by root and not deleted when the services are stopped,
        which will prevent them from being started in future sessions.
        If the services cannot be started, you may wish to check whether
        these files still exist.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_sol-guests">

    <title>Oracle Solaris Guests</title>

    <sect2 id="ts_solaris-10-guest-slow-boot-smp">

      <title>Certain Oracle Solaris 10 Releases May Take a Long Time to Boot with SMP</title>

      <para>
        When using more than one CPU, Oracle Solaris 10 10/08, and
        Oracle Solaris 10 5/09 may take a long time to boot and may
        print warnings on the system console regarding failures to read
        from disk. This is a bug in Oracle Solaris 10 which affects
        specific physical and virtual configurations. It is caused by
        trying to read microcode updates from the boot disk when the
        disk interrupt is reassigned to a not yet fully initialized
        secondary CPU. Disk reads will time out and fail, triggering
        delays of about 45 seconds and warnings.
      </para>

      <para>
        The recommended solution is upgrading to at least Oracle Solaris
        10 10/09 which includes a fix for this problem. Alternative
        solutions include restricting the number of virtual CPUs to one
        or possibly using a different storage controller.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_win-hosts">

    <title>Windows Hosts</title>

    <sect2 id="ts_win-host-com-server">

      <title>VBoxSVC Out-of-Process COM Server Issues</title>

      <para>
        &product-name; makes use of the Microsoft Component Object Model
        (COM) for interprocess and intraprocess communication. This
        enables &product-name; to share a common configuration among
        different virtual machine processes and provide several user
        interface options based on a common architecture. All global
        status information and configuration is maintained by the
        process <filename>VBoxSVC.exe</filename>, which is an
        out-of-process COM server. Whenever an &product-name; process is
        started, it requests access to the COM server and Windows
        automatically starts the process. Note that it should never be
        started by the end user.
      </para>

      <para>
        When the last process disconnects from the COM server, it will
        terminate itself after some seconds. The &product-name;
        configuration XML files are maintained and owned by the COM
        server and the files are locked whenever the server runs.
      </para>

      <para>
        In some cases, such as when a virtual machine is terminated
        unexpectedly, the COM server will not notice that the client is
        disconnected and stay active for a longer period of 10 minutes
        or so, keeping the configuration files locked. In other rare
        cases the COM server might experience an internal error and
        subsequently other processes fail to initialize it. In these
        situations, it is recommended to use the Windows task manager to
        kill the process <filename>VBoxSVC.exe</filename>.
      </para>

    </sect2>

    <sect2 id="ts_win-host-cd-dvd-changes">

      <title>CD and DVD Changes Not Recognized</title>

      <para>
        In case you have assigned a physical CD or DVD drive to a guest
        and the guest does not notice when the medium changes, make sure
        that the Windows media change notification (MCN) feature is not
        turned off. This is represented by the following key in the
        Windows registry:
      </para>

<screen>HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services\Cdrom\Autorun</screen>

      <para>
        Certain applications may disable this key against Microsoft's
        advice. If it is set to 0, change it to 1 and reboot your
        system. &product-name; relies on Windows notifying it of media
        changes.
      </para>

    </sect2>

    <sect2 id="ts_win-host-rdp-client">

      <title>Sluggish Response When Using Microsoft RDP Client</title>

      <para>
        If connecting to a Virtual Machine using the Microsoft RDP
        client, called a Remote Desktop Connection, there can be large
        delays between input such as moving the mouse over a menu and
        output. This is because this RDP client collects input for a
        certain time before sending it to the RDP server.
      </para>

      <para>
        The interval can be decreased by setting a Windows registry key
        to smaller values than the default of 100. The key does not
        exist initially and must be of type DWORD. The unit for its
        values is milliseconds. Values around 20 are suitable for
        low-bandwidth connections between the RDP client and server.
        Values around 4 can be used for a gigabit Ethernet connection.
        Generally values below 10 achieve a performance that is very
        close to that of the local input devices and screen of the host
        on which the Virtual Machine is running.
      </para>

      <para>
        Depending whether the setting should be changed for an
        individual user or for the system, set either of the following.
      </para>

<screen>HKEY_CURRENT_USER\Software\Microsoft\Terminal Server Client\Min Send Interval</screen>

<screen>HKEY_LOCAL_MACHINE\Software\Microsoft\Terminal Server Client\Min Send Interval</screen>

    </sect2>

    <sect2 id="ts_win-host-iscsi">

      <title>Running an iSCSI Initiator and Target on a Single System</title>

      <para>
        Deadlocks can occur on a Windows host when attempting to access
        an iSCSI target running in a guest virtual machine with an iSCSI
        initiator, such as a Microsoft iSCSI Initiator, that is running
        on the host. This is caused by a flaw in the Windows cache
        manager component, and causes sluggish host system response for
        several minutes, followed by a "Delayed Write Failed" error
        message in the system tray or in a separate message window. The
        guest is blocked during that period and may show error messages
        or become unstable.
      </para>

      <para>
        Setting the <literal>VBOX_DISABLE_HOST_DISK_CACHE</literal>
        environment variable to <literal>1</literal> enables a
        workaround for this problem until Microsoft addresses the issue.
        For example, open a command prompt window and start
        &product-name; like this:
      </para>

<screen>set VBOX_DISABLE_HOST_DISK_CACHE=1
VirtualBox</screen>

      <para>
        While this will decrease guest disk performance, especially
        writes, it does not affect the performance of other applications
        running on the host.
      </para>

    </sect2>

    <sect2 id="ts_win-host-bridged-network-adapters">

      <title>Bridged Networking Adapters Missing</title>

      <para>
        If no bridged adapters show up in the
        <emphasis role="bold">Networking</emphasis> section of the VM
        settings, this typically means that the bridged networking
        driver was not installed properly on your host. This could be
        due to the following reasons:
      </para>

      <itemizedlist>

        <listitem>
          <para>
            The maximum allowed filter count was reached on the host. In
            this case, the MSI log would mention the
            <literal>0x8004a029</literal> error code returned on NetFlt
            network component install, as follows:
          </para>

<screen>VBoxNetCfgWinInstallComponent: Install failed, hr (0x8004a029)</screen>

          <para>
            You can try to increase the maximum filter count in the
            Windows registry using the following key:
          </para>

<screen>HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Network\MaxNumFilters</screen>

          <para>
            The maximum number allowed is 14. After a reboot, try to
            reinstall &product-name;.
          </para>
        </listitem>

        <listitem>
          <para>
            The INF cache is corrupt. In this case, the install log at
            <filename>%windir%\inf\setupapi.dev.log</filename> would
            typically mention the failure to find a suitable driver
            package for either the <command>sun_VBoxNetFlt</command> or
            <command>sun_VBoxNetFltmp</command> components. The solution
            then is to uninstall &product-name;, remove the INF cache
            (<filename>%windir%\inf\INFCACHE.1</filename>), reboot and
            try to reinstall &product-name;.
          </para>
        </listitem>

      </itemizedlist>

    </sect2>

    <sect2 id="ts_win-host-host-only-network-adapters">

      <title>Host-Only Networking Adapters Cannot be Created</title>

      <para>
        If a host-only adapter cannot be created, either with the
        VirtualBox Manager or the <command>VBoxManage</command> command,
        then the INF cache is probably corrupt. In this case, the
        install log at
        <filename>%windir%\inf\setupapi.dev.log</filename> would
        typically mention the failure to find a suitable driver package
        for the <filename>sun_VBoxNetAdp</filename> component. Again, as
        with the bridged networking problem described above, the
        solution is to uninstall &product-name;, remove the INF cache
        (<filename>%windir%\inf\INFCACHE.1</filename>), reboot and try
        to reinstall &product-name;.
      </para>

    </sect2>

  </sect1>

  <sect1 id="ts_lin-hosts">

    <title>Linux Hosts</title>

    <sect2 id="ts_linux-kernelmodule-fails-to-load">

      <title>Linux Kernel Module Refuses to Load</title>

      <para>
        If the &product-name; kernel module, <command>vboxdrv</command>,
        refuses to load you may see an <literal>Error inserting vboxdrv:
        Invalid argument</literal> message. As root, check the output of
        the <command>dmesg</command> command to find out why the load
        failed. Most probably the kernel disagrees with the version of
        <command>gcc</command> used to compile the module. Make sure
        that you use the same compiler that was used to build the
        kernel.
      </para>

    </sect2>

    <sect2 id="ts_linux-host-cd-dvd-not-found">

      <title>Linux Host CD/DVD or Floppy Disk Drive Not Found</title>

      <para>
        If you have configured a virtual machine to use the host's CD or
        DVD drive or floppy disk drive, but this does not appear to
        work, make sure that the current user has permission to access
        the corresponding Linux device file. For example, for a CD or
        DVD drive this may be <filename>/dev/hdc</filename>,
        <filename>/dev/scd0</filename>, <filename>/dev/cdrom</filename>
        or similar. On most distributions, the user must be added to a
        corresponding group, usually called <command>cdrom</command> or
        <command>cdrw</command> or <command>floppy</command>.
      </para>

      <para>
        On supported Linux distributions, &product-name; uses
        <command>udev</command> to locate hardware such as CD/DVD drives
        and floppy disk drives.
      </para>

    </sect2>

    <sect2 id="ts_linux-host-ide-messages">

      <title>Strange Guest IDE Error Messages When Writing to CD or DVD</title>

      <para>
        If the experimental CD or DVD writer support is enabled with an
        incorrect host or guest configuration, it is possible that any
        attempt to access the CD or DVD writer fails and simply results
        in guest kernel error messages for Linux guests or application
        error messages for Windows guests. &product-name; performs the
        usual consistency checks when a VM is powered up. In particular,
        it aborts with an error message if the device for the CD or DVD
        writer is not writable by the user starting the VM. But
        &product-name; cannot detect all misconfigurations. The
        necessary host and guest OS configuration is not specific for
        &product-name;, but a few frequent problems are listed here
        which occurred in connection with &product-name;.
      </para>

      <para>
        Special care must be taken to use the correct device. The
        configured host CD or DVD device file name, in most cases
        <filename>/dev/cdrom</filename>, must point to the device that
        allows writing to the CD or DVD unit. For CD or DVD writer units
        connected to a SCSI controller or to a IDE controller that
        interfaces to the Linux SCSI subsystem, common for some SATA
        controllers, this must refer to the SCSI device node, such as
        <filename>/dev/scd0</filename>. Even for IDE CD or DVD writer
        units this must refer to the appropriate SCSI CD-ROM device
        node, such as <filename>/dev/scd0</filename>, if the
        <command>ide-scsi</command> kernel module is loaded. This module
        is required for CD or DVD writer support with some early 2.6
        kernels. Many Linux distributions load this module whenever a CD
        or DVD writer is detected in the system, even if the kernel
        would support CD or DVD writers without the module.
        &product-name; supports the use of IDE device files, such as
        <filename>/dev/hdc</filename>, provided the kernel supports this
        and the <command>ide-scsi</command> module is not loaded.
      </para>

      <para>
        Similar rules, except that within the guest the CD or DVD writer
        is always an IDE device, apply to the guest configuration. Since
        this setup is very common, it is likely that the default
        configuration of the guest works as expected.
      </para>

    </sect2>

    <sect2 id="ts_linux-host-vboxsvc">

      <title>VBoxSVC IPC Issues</title>

      <para>
        On Linux, &product-name; makes use of a custom version of
        Mozilla XPCOM (cross platform component object model) for
        interprocess and intraprocess communication (IPC). The process
        <command>VBoxSVC</command> serves as a communication hub between
        different &product-name; processes and maintains the global
        configuration, such as the XML database. When starting an
        &product-name; component, the processes
        <command>VBoxSVC</command> and <command>VBoxXPCOMIPCD</command>
        are started automatically. They are only accessible from the
        user account they are running under. <command>VBoxSVC</command>
        owns the &product-name; configuration database which normally
        resides in <filename>~/.config/VirtualBox</filename>, or the
        appropriate configuration directory for your operating system.
        While it is running, the configuration files are locked.
        Communication between the various &product-name; components and
        <command>VBoxSVC</command> is performed through a local domain
        socket residing in
        <filename>/tmp/.vbox-<replaceable>username</replaceable>-ipc</filename>.
        In case there are communication problems, such as an
        &product-name; application cannot communicate with
        <command>VBoxSVC</command>, terminate the daemons and remove the
        local domain socket directory.
      </para>

    </sect2>

    <sect2 id="ts_usb-linux">

      <title>USB Not Working</title>

      <para>
        If USB is not working on your Linux host, make sure that the
        current user is a member of the <literal>vboxusers</literal>
        group. Please keep in mind that group membership does not take
        effect immediately but rather at the next login. If available,
        the <command>newgrp</command> command may avoid the need for a
        logout and login.
      </para>

    </sect2>

    <sect2 id="ts_linux-host-grsec">

      <title>PAX/grsec Kernels</title>

      <para>
        Linux kernels including the grsec patch, see
        <ulink url="http://www.grsecurity.net/" />, and derivates have
        to disable PAX_MPROTECT for the <command>VBox</command> binaries
        to be able to start a VM. The reason is that &product-name; has
        to create executable code on anonymous memory.
      </para>

    </sect2>

    <sect2 id="ts_linux-host-malloc">

      <title>Linux Kernel vmalloc Pool Exhausted</title>

      <para>
        When running a large number of VMs with a lot of RAM on a Linux
        system, say 20 VMs with 1 GB of RAM each, additional VMs might
        fail to start with a kernel error saying that the vmalloc pool
        is exhausted and should be extended. The error message also
        tells you to specify <literal>vmalloc=256MB</literal> in your
        kernel parameter list. If adding this parameter to your GRUB or
        LILO configuration makes the kernel fail to boot, with an error
        message such as <literal>failed to mount the root
        partition</literal>, then you have probably run into a memory
        conflict of your kernel and initial RAM disk. This can be solved
        by adding the following parameter to your GRUB configuration:
      </para>

<screen>uppermem 524288</screen>

    </sect2>

  </sect1>

  <sect1 id="ts_sol-hosts">

    <title>Oracle Solaris Hosts</title>

    <sect2 id="ts_sol-host-zfs">

      <title>Cannot Start VM, Not Enough Contiguous Memory</title>

      <para>
        The ZFS file system is known to use nearly all available RAM as
        cache if the default system settings are not changed. This may
        lead to a heavy fragmentation of the host memory preventing
        &product-name; VMs from being started. We recommend to limit the
        ZFS cache by adding the following line to
        <filename>/etc/system</filename>, where
        <replaceable>xxxx</replaceable> bytes is the amount of memory
        usable for the ZFS cache.
      </para>

<screen>set zfs:zfs_arc_max = xxxx</screen>

    </sect2>

  </sect1>

</chapter>