libvirt/docs/uri.html.in

331 lines
11 KiB
XML

<?xml version="1.0"?>
<html>
<body>
<h1 >Connection URIs</h1>
<ul id="toc"></ul>
<p>
Since libvirt supports many different kinds of virtualization
(often referred to as "drivers" or "hypervisors"), we need a
way to be able to specify which driver a connection refers to.
Additionally we may want to refer to a driver on a remote
machine over the network.
</p>
<p>
To this end, libvirt uses URIs as used on the Web and as defined in <a href="http://www.ietf.org/rfc/rfc2396.txt">RFC 2396</a>. This page
documents libvirt URIs.
</p>
<h2><a name="URI_libvirt">Specifying URIs to libvirt</a></h2>
<p>
The URI is passed as the <code>name</code> parameter to <a href="html/libvirt-libvirt.html#virConnectOpen"><code>virConnectOpen</code></a> or <a href="html/libvirt-libvirt.html#virConnectOpenReadOnly"><code>virConnectOpenReadOnly</code></a>. For example:
</p>
<pre>
virConnectPtr conn = virConnectOpenReadOnly (<b>"test:///default"</b>);
</pre>
<h2>
<a name="URI_config">Configuring URI aliases</a>
</h2>
<p>
To simplify life for administrators, it is possible to setup URI aliases in a
libvirt client configuration file. The configuration file is <code>/etc/libvirt/libvirt.conf</code>
for the root user, or <code>$XDG_CONFIG_HOME/libvirt/libvirt.conf</code> for any unprivileged user.
In this file, the following syntax can be used to setup aliases
</p>
<pre>
uri_aliases = [
"hail=qemu+ssh://root@hail.cloud.example.com/system",
"sleet=qemu+ssh://root@sleet.cloud.example.com/system",
]
</pre>
<p>
A URI alias should be a string made up from the characters
<code>a-Z, 0-9, _, -</code>. Following the <code>=</code>
can be any libvirt URI string, including arbitrary URI parameters.
URI aliases will apply to any application opening a libvirt
connection, unless it has explicitly passed the <code>VIR_CONNECT_NO_ALIASES</code>
parameter to <code>virConnectOpenAuth</code>. If the passed in
URI contains characters outside the allowed alias character
set, no alias lookup will be attempted.
</p>
<h2><a name="URI_default">Default URI choice</a></h2>
<p>
If the URI passed to <code>virConnectOpen*</code> is NULL, then libvirt will use the following
logic to determine what URI to use.
</p>
<ol>
<li>The environment variable <code>LIBVIRT_DEFAULT_URI</code></li>
<li>The client configuration file <code>uri_default</code> parameter</li>
<li>Probe each hypervisor in turn until one that works is found</li>
</ol>
<h2>
<a name="URI_virsh">Specifying URIs to virsh, virt-manager and virt-install</a>
</h2>
<p>
In virsh use the <code>-c</code> or <code>--connect</code> option:
</p>
<pre>
virsh <b>-c test:///default</b> list
</pre>
<p>
If virsh finds the environment variable
<code>VIRSH_DEFAULT_CONNECT_URI</code> set, it will try this URI by
default. Use of this environment variable is, however, deprecated
now that libvirt supports <code>LIBVIRT_DEFAULT_URI</code> itself.
</p>
<p>
When using the interactive virsh shell, you can also use the
<code>connect</code> <i>URI</i> command to reconnect to another
hypervisor.
</p>
<p>
In virt-manager use the <code>-c</code> or <code>--connect=</code><i>URI</i> option:
</p>
<pre>
virt-manager <b>-c test:///default</b>
</pre>
<p>
In virt-install use the <code>--connect=</code><i>URI</i> option:
</p>
<pre>
virt-install <b>--connect=test:///default</b> <i>[other options]</i>
</pre>
<h2>
<a name="URI_xen">xen:/// URI</a>
</h2>
<p>
<i>This section describes a feature which is new in libvirt &gt;
0.2.3. For libvirt &#x2264; 0.2.3 use <a href="#URI_legacy_xen"><code>"xen"</code></a>.</i>
</p>
<p>
To access a Xen hypervisor running on the local machine
use the URI <code>xen:///</code>.
</p>
<h2>
<a name="URI_qemu">qemu:///... QEMU and KVM URIs</a>
</h2>
<p>
To use QEMU support in libvirt you must be running the
<code>libvirtd</code> daemon (named <code>libvirt_qemud</code>
in releases prior to 0.3.0). The purpose of this
daemon is to manage qemu instances.
</p>
<p>
The <code>libvirtd</code> daemon should be started by the
init scripts when the machine boots. It should appear as
a process <code>libvirtd --daemon</code> running as root
in the background and will handle qemu instances on behalf
of all users of the machine (among other things). </p>
<p>
So to connect to the daemon, one of two different URIs is used:
</p>
<ul>
<li><code>qemu:///system</code> connects to a system mode daemon. </li>
<li><code>qemu:///session</code> connects to a session mode daemon. </li>
</ul>
<p>
(If you do <code>libvirtd --help</code>, the daemon will print
out the paths of the Unix domain socket(s) that it listens on in
the various different modes).
</p>
<p>
KVM URIs are identical. You select between qemu, qemu accelerated and
KVM guests in the <a href="format.html#KVM1">guest XML as described
here</a>.
</p>
<h2>
<a name="URI_remote">Remote URIs</a>
</h2>
<p>
Remote URIs are formed by taking ordinary local URIs and adding a
hostname and/or transport name. As a special case, using a URI
scheme of 'remote', will tell the remote libvirtd server to probe
for the optimal hypervisor driver. This is equivalent to passing
a NULL URI for a local connection. For example:
</p>
<table class="top_table">
<tr>
<th> Local URI </th>
<th> Remote URI </th>
<th> Meaning </th>
</tr>
<tr>
<td>
<code>xen:///</code>
</td>
<td>
<code>xen://oirase/</code>
</td>
<td> Connect to the Xen hypervisor running on host <code>oirase</code>
using TLS. </td>
</tr>
<tr>
<td>
<code>NULL</code>
</td>
<td>
<code>remote://oirase/</code>
</td>
<td> Connect to the "default" hypervisor running on host <code>oirase</code>
using TLS. </td>
</tr>
<tr>
<td>
<code>xen:///</code>
</td>
<td>
<code>xen+ssh://oirase/</code>
</td>
<td> Connect to the Xen hypervisor running on host <code>oirase</code>
by going over an <code>ssh</code> connection. </td>
</tr>
<tr>
<td>
<code>test:///default</code>
</td>
<td>
<code>test+tcp://oirase/default</code>
</td>
<td> Connect to the test driver on host <code>oirase</code>
using an unsecured TCP connection. </td>
</tr>
</table>
<p>
Remote URIs in libvirt offer a rich syntax and many features.
We refer you to <a href="remote.html#Remote_URI_reference">the libvirt
remote URI reference</a> and <a href="remote.html">full documentation
for libvirt remote support</a>.
</p>
<h2>
<a name="URI_test">test:///... Test URIs</a>
</h2>
<p>
The test driver is a dummy hypervisor for test purposes.
The URIs supported are:
</p>
<ul>
<li><code>test:///default</code> connects to a default set of
host definitions built into the driver. </li>
<li><code>test:///path/to/host/definitions</code> connects to
a set of host definitions held in the named file.
</li>
</ul>
<h2>
<a name="URI_legacy">Other &amp; legacy URI formats</a>
</h2>
<h3>
<a name="URI_NULL">NULL and empty string URIs</a>
</h3>
<p>
Libvirt allows you to pass a <code>NULL</code> pointer to
<code>virConnectOpen*</code>. Empty string (<code>""</code>) acts in
the same way. Traditionally this has meant
<q>connect to the local Xen hypervisor</q>. However in future this
may change to mean <q>connect to the best available hypervisor</q>.
</p>
<p>
The theory is that if, for example, Xen is unavailable but the
machine is running an OpenVZ kernel, then we should not try to
connect to the Xen hypervisor since that is obviously the wrong
thing to do.
</p>
<p>
In any case applications linked to libvirt can continue to pass
<code>NULL</code> as a default choice, but should always allow the
user to override the URI, either by constructing one or by allowing
the user to type a URI in directly (if that is appropriate). If your
application wishes to connect specifically to a Xen hypervisor, then
for future proofing it should choose a full <a href="#URI_xen"><code>xen:///</code> URI</a>.
</p>
<h3>
<a name="URI_file">File paths (xend-unix-server)</a>
</h3>
<p>
If XenD is running and configured in <code>/etc/xen/xend-config.sxp</code>:
</p>
<pre>
(xend-unix-server yes)
</pre>
<p>
then it listens on a Unix domain socket, usually at
<code>/var/lib/xend/xend-socket</code>. You may pass a different path
using a file URI such as:
</p>
<pre>
virsh -c ///var/run/xend/xend-socket
</pre>
<h3>
<a name="URI_http">Legacy: <code>http://...</code> (xend-http-server)</a>
</h3>
<p>
If XenD is running and configured in <code>/etc/xen/xend-config.sxp</code>:
</p>
<pre>
(xend-http-server yes)
</pre>
<p>
then it listens on TCP port 8000. libvirt allows you to
try to connect to xend running on remote machines by passing
<code>http://<i>hostname</i>[:<i>port</i>]/</code>, for example:
</p>
<pre>
virsh -c http://oirase/ list
</pre>
<p>
This method is unencrypted and insecure and is definitely not
recommended for production use. Instead use <a href="remote.html">libvirt's remote support</a>.
</p>
<p>
Notes:
</p>
<ol>
<li> The HTTP client does not fully support IPv6. </li>
<li> Many features do not work as expected across HTTP connections, in
particular, <a href="html/libvirt-libvirt.html#virConnectGetCapabilities">virConnectGetCapabilities</a>.
The <a href="remote.html">remote support</a> however does work
correctly. </li>
<li> XenD's new-style XMLRPC interface is not supported by
libvirt, only the old-style sexpr interface known in the Xen
documentation as "unix server" or "http server".</li>
</ol>
<h3>
<a name="URI_legacy_xen">Legacy: <code>"xen"</code></a>
</h3>
<p>
Another legacy URI is to specify name as the string
<code>"xen"</code>. This will continue to refer to the Xen
hypervisor. However you should prefer a full <a href="#URI_xen"><code>xen:///</code> URI</a> in all future code.
</p>
<h3>
<a name="URI_legacy_proxy">Legacy: Xen proxy</a>
</h3>
<p>
Libvirt continues to support connections to a separately running Xen
proxy daemon. This provides a way to allow non-root users to make a
safe (read-only) subset of queries to the hypervisor.
</p>
<p>
There is no specific "Xen proxy" URI. However if a Xen URI of any of
the ordinary or legacy forms is used (eg. <code>NULL</code>,
<code>""</code>, <code>"xen"</code>, ...) which fails, <i>and</i> the
user is not root, <i>and</i> the Xen proxy socket can be connected to
(<code>/tmp/libvirt_proxy_conn</code>), then libvirt will use a proxy
connection.
</p>
<p>
You should consider using <a href="remote.html">libvirt remote support</a>
in future. <span class="since">Since 0.8.6</span> libvirt doesn't contain
the Xen proxy anymore and you should use libvirtd instead.
</p>
</body>
</html>