drvesx.html.in 30.3 KB
Newer Older
1
<?xml version="1.0" encoding="UTF-8"?>
2
<!DOCTYPE html>
3 4
<html xmlns="http://www.w3.org/1999/xhtml">
  <body>
5 6 7
    <h1>VMware ESX hypervisor driver</h1>
    <ul id="toc"></ul>
    <p>
8
        The libvirt VMware ESX driver can manage VMware ESX/ESXi 3.5/4.x/5.x and
9
        VMware GSX 2.0, also called VMware Server 2.0, and possibly later
10
        versions. <span class="since">Since 0.8.3</span> the driver can also
11
        connect to a VMware vCenter 2.5/4.x/5.x (VPX).
12 13
    </p>

14
    <h2><a id="project">Project Links</a></h2>
15 16 17 18 19 20 21

    <ul>
      <li>
        The <a href="http://www.vmware.com/">VMware ESX and GSX</a>
        hypervisors
      </li>
    </ul>
22

23
    <h2><a id="prereq">Deployment pre-requisites</a></h2>
24
    <p>
25
        None. Any out-of-the-box installation of VPX/ESX(i)/GSX should work. No
26 27 28 29
        preparations are required on the server side, no libvirtd must be
        installed on the ESX server. The driver uses version 2.5 of the remote,
        SOAP based
        <a href="http://www.vmware.com/support/developer/vc-sdk/visdk25pubs/ReferenceGuide/">
30
        VMware Virtual Infrastructure API</a> (VI API) to communicate with the
31 32
        ESX server, like the VMware Virtual Infrastructure Client (VI client)
        does. Since version 4.0 this API is called
33 34 35 36
        <a href="http://www.vmware.com/support/developer/vc-sdk/visdk400pubs/ReferenceGuide/">
        VMware vSphere API</a>.
    </p>

37
    <h2><a id="uri">Connections to the VMware ESX driver</a></h2>
38 39 40 41
    <p>
        Some example remote connection URIs for the driver are:
    </p>
<pre>
42
vpx://example-vcenter.com/dc1/srv1     (VPX over HTTPS, select ESX server 'srv1' in datacenter 'dc1')
43 44 45 46
esx://example-esx.com                  (ESX over HTTPS)
gsx://example-gsx.com                  (GSX over HTTPS)
esx://example-esx.com/?transport=http  (ESX over HTTP)
esx://example-esx.com/?no_verify=1     (ESX over HTTPS, but doesn't verify the server's SSL certificate)
47
</pre>
48 49 50 51 52 53 54
    <p>
        <strong>Note</strong>: In contrast to other drivers, the ESX driver is
        a client-side-only driver. It connects to the ESX server using HTTP(S).
        Therefore, the <a href="remote.html">remote transport mechanism</a>
        provided by the remote driver and libvirtd will not work, and you
        cannot use URIs like <code>esx+ssh://example.com</code>.
    </p>
55 56


57
    <h3><a id="uriformat">URI Format</a></h3>
58
    <p>
59
        URIs have this general form (<code>[...]</code> marks an optional part).
60 61
    </p>
<pre>
62
type://[username@]hostname[:port]/[[folder/...]datacenter/[folder/...][cluster/]server][?extraparameters]
63 64 65
</pre>
    <p>
        The <code>type://</code> is either <code>esx://</code> or
66 67 68 69 70
        <code>gsx://</code> or <code>vpx://</code> <span class="since">since 0.8.3</span>.
        The driver selects the default port depending on the <code>type://</code>.
        For <code>esx://</code> and <code>vpx://</code> the default HTTPS port
        is 443, for <code>gsx://</code> it is 8333.
        If the port parameter is given, it overrides the default port.
71
    </p>
72 73 74 75 76 77 78 79 80 81 82 83 84
    <p>
        A <code>vpx://</code> connection is currently restricted to a single
        ESX server. This might be relaxed in the future. The path part of the
        URI is used to specify the datacenter and the ESX server in it. If the
        ESX server is part of a cluster then the cluster has to be specified too.
    </p>
    <p>
        An example: ESX server <code>example-esx.com</code> is managed by
        vCenter <code>example-vcenter.com</code> and part of cluster
        <code>cluster1</code>. This cluster is part of datacenter <code>dc1</code>.
    </p>
<pre>
vpx://example-vcenter.com/dc1/cluster1/example-esx.com
85 86 87 88 89 90 91 92
</pre>
    <p>
        Datacenters and clusters can be organized in folders, those have to be
        specified as well. The driver can handle folders
        <span class="since">since 0.9.7</span>.
    </p>
<pre>
vpx://example-vcenter.com/folder1/dc1/folder2/example-esx.com
93
</pre>
94 95


96
    <h4><a id="extraparams">Extra parameters</a></h4>
97 98
    <p>
        Extra parameters can be added to a URI as part of the query string
99 100 101 102 103 104 105 106 107
        (the part following <code>?</code>). A single parameter is formed by a
        <code>name=value</code> pair. Multiple parameters are separated by
        <code>&amp;</code>.
    </p>
<pre>
?<span style="color: #E50000">no_verify=1</span>&amp;<span style="color: #00B200">auto_answer=1</span>&amp;<span style="color: #0000E5">proxy=socks://example-proxy.com:23456</span>
</pre>
    <p>
        The driver understands the extra parameters shown below.
108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
    </p>
    <table class="top_table">
        <tr>
            <th>Name</th>
            <th>Values</th>
            <th>Meaning</th>
        </tr>
        <tr>
            <td>
                <code>transport</code>
            </td>
            <td>
                <code>http</code> or <code>https</code>
            </td>
            <td>
123 124 125
                Overrides the default HTTPS transport. For <code>esx://</code>
                and <code>vpx://</code> the default HTTP port is 80, for
                <code>gsx://</code> it is 8222.
126 127 128 129 130 131 132
            </td>
        </tr>
        <tr>
            <td>
                <code>vcenter</code>
            </td>
            <td>
133
                Hostname of a VMware vCenter or <code>*</code>
134 135 136
            </td>
            <td>
                In order to perform a migration the driver needs to know the
137 138
                VMware vCenter for the ESX server. If set to <code>*</code>,
                the driver connects to the vCenter known to the ESX server.
139
                This parameter in useful when connecting to an ESX server only.
140 141 142 143 144 145 146 147 148 149 150
            </td>
        </tr>
        <tr>
            <td>
                <code>no_verify</code>
            </td>
            <td>
                <code>0</code> or <code>1</code>
            </td>
            <td>
                If set to 1, this disables libcurl client checks of the server's
151
                SSL certificate. The default value is 0. See the
152 153
                <a href="#certificates">Certificates for HTTPS</a> section for
                details.
154 155
            </td>
        </tr>
156 157 158 159 160 161 162 163 164 165 166
        <tr>
            <td>
                <code>auto_answer</code>
            </td>
            <td>
                <code>0</code> or <code>1</code>
            </td>
            <td>
                If set to 1, the driver answers all
                <a href="#questions">questions</a> with the default answer.
                If set to 0, questions are reported as errors. The default
167
                value is 0. <span class="since">Since 0.7.5</span>.
M
Matthias Bolte 已提交
168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185
            </td>
        </tr>
        <tr>
            <td>
                <code>proxy</code>
            </td>
            <td>
                <code>[type://]hostname[:port]</code>
            </td>
            <td>
                Allows to specify a proxy for HTTP and HTTPS communication.
                <span class="since">Since 0.8.2</span>.
                The optional <code>type</code> part may be one of:
                <code>http</code>, <code>socks</code>, <code>socks4</code>,
                <code>socks4a</code> or <code>socks5</code>. The default is
                <code>http</code> and <code>socks</code> is synonymous for
                <code>socks5</code>. The optional <code>port</code> allows to
                override the default port 1080.
186 187
            </td>
        </tr>
188 189 190
    </table>


191
    <h3><a id="auth">Authentication</a></h3>
192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210
    <p>
        In order to perform any useful operation the driver needs to log into
        the ESX server. Therefore, only <code>virConnectOpenAuth</code> can be
        used to connect to an ESX server, <code>virConnectOpen</code> and
        <code>virConnectOpenReadOnly</code> don't work.
        To log into an ESX server or vCenter the driver will request
        credentials using the callback passed to the
        <code>virConnectOpenAuth</code> function. The driver passes the
        hostname as challenge parameter to the callback. This enables the
        callback to distinguish between requests for ESX server and vCenter.
    </p>
    <p>
        <strong>Note</strong>: During the ongoing driver development, testing
        is done using an unrestricted <code>root</code> account. Problems may
        occur if you use a restricted account. Detailed testing with restricted
        accounts has not been done yet.
    </p>


211
    <h3><a id="certificates">Certificates for HTTPS</a></h3>
212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246
    <p>
        By default the ESX driver uses HTTPS to communicate with an ESX server.
        Proper HTTPS communication requires correctly configured SSL
        certificates. This certificates are different from the ones libvirt
        uses for <a href="remote.html">secure communication over TLS</a> to a
        libvirtd one a remote server.
    </p>
    <p>
        By default the driver tries to verify the server's SSL certificate
        using the CA certificate pool installed on your client computer. With
        an out-of-the-box installed ESX server this won't work, because a newly
        installed ESX server uses auto-generated self-signed certificates.
        Those are singed by a CA certificate that is typically not known to your
        client computer and libvirt will report an error like this one:
    </p>
<pre>
error: internal error curl_easy_perform() returned an error: Peer certificate cannot be authenticated with known CA certificates (60)
</pre>
    <p>
        Where are two ways to solve this problem:
    </p>
    <ul>
        <li>
            Use the <code>no_verify=1</code> <a href="#extraparams">extra parameter</a>
            to disable server certificate verification.
        </li>
        <li>
            Generate new SSL certificates signed by a CA known to your client
            computer and replace the original ones on your ESX server. See the
            section <i>Replace a Default Certificate with a CA-Signed Certificate</i>
            in the <a href="http://www.vmware.com/pdf/vsphere4/r40/vsp_40_esx_server_config.pdf">ESX Configuration Guide</a>
        </li>
    </ul>


247
    <h3><a id="connproblems">Connection problems</a></h3>
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305
    <p>
        There are also other causes for connection problems than the
        <a href="#certificates">HTTPS certificate</a> related ones.
    </p>
    <ul>
        <li>
            As stated before the ESX driver doesn't need the
            <a href="remote.html">remote transport mechanism</a>
            provided by the remote driver and libvirtd, nor does the ESX driver
            support it. Therefore, using an URI including a transport in the
            scheme won't work. Only <a href="#uriformat">URIs as described</a>
            are supported by the ESX driver. Here's a collection of possible
            error messages:
<pre>
$ virsh -c esx+tcp://example.com/
error: unable to connect to libvirtd at 'example.com': Connection refused
</pre>
<pre>
$ virsh -c esx+tls://example.com/
error: Cannot access CA certificate '/etc/pki/CA/cacert.pem': No such file or directory
</pre>
<pre>
$ virsh -c esx+ssh://example.com/
error: cannot recv data: ssh: connect to host example.com port 22: Connection refused
</pre>
<pre>
$ virsh -c esx+ssh://example.com/
error: cannot recv data: Resource temporarily unavailable
</pre>
        </li>
        <li>
            <span class="since">Since 0.7.0</span> libvirt contains the ESX
            driver. Earlier versions of libvirt will report a misleading error
            about missing certificates when you try to connect to an ESX server.
<pre>
$ virsh -c esx://example.com/
error: Cannot access CA certificate '/etc/pki/CA/cacert.pem': No such file or directory
</pre>
            <p>
                Don't let this error message confuse you. Setting up certificates
                as described on the <a href="remote.html#Remote_certificates">remote transport mechanism</a> page
                does not help, as this is not a certificate related problem.
            </p>
            <p>
                To fix this problem you need to update your libvirt to 0.7.0 or newer.
                You may also see this error when you use a libvirt version that
                contains the ESX driver but you or your distro disabled the ESX
                driver during compilation. <span class="since">Since 0.8.3</span>
                the error message has been improved in this case:
            </p>
<pre>
$ virsh -c esx://example.com/
error: invalid argument in libvirt was built without the 'esx' driver
</pre>
        </li>
    </ul>


306
    <h2><a id="questions">Questions blocking tasks</a></h2>
307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
    <p>
        Some methods of the VI API start tasks, for example
        <code>PowerOnVM_Task()</code>. Such tasks may be blocked by questions
        if the ESX server detects an issue with the domain that requires user
        interaction. The ESX driver cannot prompt the user to answer a
        question, libvirt doesn't have an API for something like this.
    </p>
    <p>
        The VI API provides the <code>AnswerVM()</code> method to
        programmatically answer a questions. So the driver has two options
        how to handle such a situation: either answer the questions with the
        default answer or report the question as an error and cancel the
        blocked task if possible. The
        <a href="#uriformat"><code>auto_answer</code></a> query parameter
        controls the answering behavior.
    </p>


325
    <h2><a id="xmlspecial">Specialties in the domain XML config</a></h2>
326 327 328 329
    <p>
        There are several specialties in the domain XML config for ESX domains.
    </p>

330
    <h3><a id="restrictions">Restrictions</a></h3>
331 332 333 334 335 336 337 338 339 340 341 342 343
    <p>
        There are some restrictions for some values of the domain XML config.
        The driver will complain if this restrictions are violated.
    </p>
    <ul>
        <li>
            Memory size has to be a multiple of 4096
        </li>
        <li>
            Number of virtual CPU has to be 1 or a multiple of 2
        </li>
        <li>
            Valid MAC address prefixes are <code>00:0c:29</code> and
344 345
            <code>00:50:56</code>. <span class="since">Since 0.7.6</span>
            arbitrary <a href="#macaddresses">MAC addresses</a> are supported.
346 347 348 349
        </li>
    </ul>


350
    <h3><a id="datastore">Datastore references</a></h3>
351 352 353 354 355 356 357 358 359 360 361 362
    <p>
        Storage is managed in datastores. VMware uses a special path format to
        reference files in a datastore. Basically, the datastore name is put
        into squared braces in front of the path.
    </p>
<pre>
[datastore] directory/filename
</pre>
    <p>
        To define a new domain the driver converts the domain XML into a
        VMware VMX file and uploads it to a datastore known to the ESX server.
        Because multiple datastores may be known to an ESX server the driver
363
        needs to decide to which datastore the VMX file should be uploaded.
364 365 366 367 368
        The driver deduces this information from the path of the source of the
        first file-based harddisk listed in the domain XML.
    </p>


369
    <h3><a id="macaddresses">MAC addresses</a></h3>
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
    <p>
        VMware has registered two MAC address prefixes for domains:
        <code>00:0c:29</code> and <code>00:50:56</code>. These prefixes are
        split into ranges for different purposes.
    </p>
    <table class="top_table">
        <tr>
            <th>Range</th>
            <th>Purpose</th>
        </tr>
        <tr>
            <td>
                <code>00:0c:29:00:00:00</code> - <code>00:0c:29:ff:ff:ff</code>
            </td>
            <td>
                An ESX server autogenerates MAC addresses from this range if
                the VMX file doesn't contain a MAC address when trying to start
                a domain.
            </td>
        </tr>
        <tr>
            <td>
                <code>00:50:56:00:00:00</code> - <code>00:50:56:3f:ff:ff</code>
            </td>
            <td>
                MAC addresses from this range can by manually assigned by the
                user in the VI client.
            </td>
        </tr>
        <tr>
            <td>
                <code>00:50:56:80:00:00</code> - <code>00:50:56:bf:ff:ff</code>
            </td>
            <td>
                A VI client autogenerates MAC addresses from this range for
                newly defined domains.
            </td>
        </tr>
    </table>
    <p>
        The VMX files generated by the ESX driver always contain a MAC address,
        because libvirt generates a random one if an interface element in the
        domain XML file lacks a MAC address.
        <span class="since">Since 0.7.6</span> the ESX driver sets the prefix
        for generated MAC addresses to <code>00:0c:29</code>. Before 0.7.6
        the <code>00:50:56</code> prefix was used. Sometimes this resulted in
        the generation of out-of-range MAC address that were rejected by the
        ESX server.
    </p>
    <p>
        Also <span class="since">since 0.7.6</span> every MAC address outside
        this ranges can be used. For such MAC addresses the ESX server-side
        check is disabled in the VMX file to stop the ESX server from rejecting
        out-of-predefined-range MAC addresses.
    </p>
<pre>
ethernet0.checkMACAddress = "false"
</pre>


430
    <h3><a id="hardware">Available hardware</a></h3>
431 432 433 434 435 436 437
    <p>
        VMware ESX supports different models of SCSI controllers and network
        cards.
    </p>

    <h4>SCSI controller models</h4>
    <dl>
438 439 440 441 442 443 444 445 446
        <dt><code>auto</code></dt>
        <dd>
            This isn't a actual controller model. If specified the ESX driver
            tries to detect the SCSI controller model referenced in the
            <code>.vmdk</code> file and use it. Autodetection fails when a
            SCSI controller has multiple disks attached and the SCSI controller
            models referenced in the <code>.vmdk</code> files are inconsistent.
            <span class="since">Since 0.8.3</span>
        </dd>
447 448 449 450 451 452 453 454
        <dt><code>buslogic</code></dt>
        <dd>
            BusLogic SCSI controller for older guests.
        </dd>
        <dt><code>lsilogic</code></dt>
        <dd>
            LSI Logic SCSI controller for recent guests.
        </dd>
455 456
        <dt><code>lsisas1068</code></dt>
        <dd>
457 458 459 460 461 462 463
            LSI Logic SAS 1068 controller. <span class="since">Since 0.8.0</span>
        </dd>
        <dt><code>vmpvscsi</code></dt>
        <dd>
            Special VMware Paravirtual SCSI controller, requires VMware tools inside
            the guest. See <a href="http://kb.vmware.com/kb/1010398">VMware KB1010398</a>
            for details. <span class="since">Since 0.8.3</span>
464
        </dd>
465 466 467 468
    </dl>
    <p>
        Here a domain XML snippet:
    </p>
469
<pre>
470 471 472 473 474 475 476 477
...
&lt;disk type='file' device='disk'&gt;
  &lt;source file='[local-storage] Fedora11/Fedora11.vmdk'/&gt;
  &lt;target dev='sda' bus='scsi'/&gt;
  &lt;address type='drive' controller='0' bus='0' unit='0'/&gt;
&lt;/disk&gt;
&lt;controller type='scsi' index='0' model='<strong>lsilogic</strong>'/&gt;
...
478 479 480 481 482 483
</pre>
    <p>
        The controller element is supported <span class="since">since 0.8.2</span>.
        Prior to this <code>&lt;driver name='lsilogic'/&gt;</code> was abused to
        specify the SCSI controller model. This attribute usage is deprecated now.
    </p>
484
<pre>
485 486 487 488 489 490 491
...
&lt;disk type='file' device='disk'&gt;
  &lt;driver name='<strong>lsilogic</strong>'/&gt;
  &lt;source file='[local-storage] Fedora11/Fedora11.vmdk'/&gt;
  &lt;target dev='sda' bus='scsi'/&gt;
&lt;/disk&gt;
...
492 493 494 495 496 497 498 499 500
</pre>


    <h4>Network card models</h4>
    <dl>
        <dt><code>vlance</code></dt>
        <dd>
            AMD PCnet32 network card for older guests.
        </dd>
501
        <dt><code>vmxnet</code>, <code>vmxnet2</code>, <code>vmxnet3</code></dt>
502 503
        <dd>
            Special VMware VMXnet network card, requires VMware tools inside
504 505
            the guest. See <a href="http://kb.vmware.com/kb/1001805">VMware KB1001805</a>
            for details.
506 507 508 509 510 511 512 513 514 515
        </dd>
        <dt><code>e1000</code></dt>
        <dd>
            Intel E1000 network card for recent guests.
        </dd>
    </dl>
    <p>
        Here a domain XML snippet:
    </p>
<pre>
516 517 518 519 520 521 522
...
&lt;interface type='bridge'&gt;
  &lt;mac address='00:50:56:25:48:c7'/&gt;
  &lt;source bridge='VM Network'/&gt;
  &lt;model type='<strong>e1000</strong>'/&gt;
&lt;/interface&gt;
...
523 524 525
</pre>


526
    <h2><a id="importexport">Import and export of domain XML configs</a></h2>
527 528 529 530 531 532
    <p>
        The ESX driver currently supports a native config format known as
        <code>vmware-vmx</code> to handle VMware VMX configs.
    </p>


533
    <h3><a id="xmlimport">Converting from VMware VMX config to domain XML config</a></h3>
534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
    <p>
        The <code>virsh domxml-from-native</code> provides a way to convert an
        existing VMware VMX config into a domain XML config that can then be
        used by libvirt.
    </p>
<pre>
$ cat &gt; demo.vmx &lt;&lt; EOF
#!/usr/bin/vmware
config.version = "8"
virtualHW.version = "4"
floppy0.present = "false"
nvram = "Fedora11.nvram"
deploymentPlatform = "windows"
virtualHW.productCompatibility = "hosted"
tools.upgrade.policy = "useGlobal"
powerType.powerOff = "default"
powerType.powerOn = "default"
powerType.suspend = "default"
powerType.reset = "default"
displayName = "Fedora11"
extendedConfigFile = "Fedora11.vmxf"
scsi0.present = "true"
scsi0.sharedBus = "none"
scsi0.virtualDev = "lsilogic"
memsize = "1024"
scsi0:0.present = "true"
scsi0:0.fileName = "/vmfs/volumes/498076b2-02796c1a-ef5b-000ae484a6a3/Fedora11/Fedora11.vmdk"
scsi0:0.deviceType = "scsi-hardDisk"
ide0:0.present = "true"
ide0:0.clientDevice = "true"
ide0:0.deviceType = "cdrom-raw"
ide0:0.startConnected = "false"
ethernet0.present = "true"
ethernet0.networkName = "VM Network"
ethernet0.addressType = "vpx"
569
ethernet0.generatedAddress = "00:50:56:91:48:c7"
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611
chipset.onlineStandby = "false"
guestOSAltName = "Red Hat Enterprise Linux 5 (32-Bit)"
guestOS = "rhel5"
uuid.bios = "50 11 5e 16 9b dc 49 d7-f1 71 53 c4 d7 f9 17 10"
snapshot.action = "keep"
sched.cpu.min = "0"
sched.cpu.units = "mhz"
sched.cpu.shares = "normal"
sched.mem.minsize = "0"
sched.mem.shares = "normal"
toolScripts.afterPowerOn = "true"
toolScripts.afterResume = "true"
toolScripts.beforeSuspend = "true"
toolScripts.beforePowerOff = "true"
scsi0:0.redo = ""
tools.syncTime = "false"
uuid.location = "56 4d b5 06 a2 bd fb eb-ae 86 f7 d8 49 27 d0 c4"
sched.cpu.max = "unlimited"
sched.swap.derivedName = "/vmfs/volumes/498076b2-02796c1a-ef5b-000ae484a6a3/Fedora11/Fedora11-7de040d8.vswp"
tools.remindInstall = "TRUE"
EOF

$ virsh -c esx://example.com domxml-from-native vmware-vmx demo.vmx
Enter username for example.com [root]:
Enter root password for example.com:
&lt;domain type='vmware'&gt;
  &lt;name&gt;Fedora11&lt;/name&gt;
  &lt;uuid&gt;50115e16-9bdc-49d7-f171-53c4d7f91710&lt;/uuid&gt;
  &lt;memory&gt;1048576&lt;/memory&gt;
  &lt;currentMemory&gt;1048576&lt;/currentMemory&gt;
  &lt;vcpu&gt;1&lt;/vcpu&gt;
  &lt;os&gt;
    &lt;type arch='i686'&gt;hvm&lt;/type&gt;
  &lt;/os&gt;
  &lt;clock offset='utc'/&gt;
  &lt;on_poweroff&gt;destroy&lt;/on_poweroff&gt;
  &lt;on_reboot&gt;restart&lt;/on_reboot&gt;
  &lt;on_crash&gt;destroy&lt;/on_crash&gt;
  &lt;devices&gt;
    &lt;disk type='file' device='disk'&gt;
      &lt;source file='[local-storage] Fedora11/Fedora11.vmdk'/&gt;
      &lt;target dev='sda' bus='scsi'/&gt;
612
      &lt;address type='drive' controller='0' bus='0' unit='0'/&gt;
613
    &lt;/disk&gt;
614
    &lt;controller type='scsi' index='0' model='lsilogic'/&gt;
615 616 617 618 619 620 621 622 623
    &lt;interface type='bridge'&gt;
      &lt;mac address='00:50:56:91:48:c7'/&gt;
      &lt;source bridge='VM Network'/&gt;
    &lt;/interface&gt;
  &lt;/devices&gt;
&lt;/domain&gt;
</pre>


624
    <h3><a id="xmlexport">Converting from domain XML config to VMware VMX config</a></h3>
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643
    <p>
      The <code>virsh domxml-to-native</code> provides a way to convert a
      domain XML config into a VMware VMX config.
    </p>
<pre>
$ cat &gt; demo.xml &lt;&lt; EOF
&lt;domain type='vmware'&gt;
  &lt;name&gt;Fedora11&lt;/name&gt;
  &lt;uuid&gt;50115e16-9bdc-49d7-f171-53c4d7f91710&lt;/uuid&gt;
  &lt;memory&gt;1048576&lt;/memory&gt;
  &lt;currentMemory&gt;1048576&lt;/currentMemory&gt;
  &lt;vcpu&gt;1&lt;/vcpu&gt;
  &lt;os&gt;
    &lt;type arch='x86_64'&gt;hvm&lt;/type&gt;
  &lt;/os&gt;
  &lt;devices&gt;
    &lt;disk type='file' device='disk'&gt;
      &lt;source file='[local-storage] Fedora11/Fedora11.vmdk'/&gt;
      &lt;target dev='sda' bus='scsi'/&gt;
644
      &lt;address type='drive' controller='0' bus='0' unit='0'/&gt;
645
    &lt;/disk&gt;
646
    &lt;controller type='scsi' index='0' model='lsilogic'/&gt;
647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677
    &lt;interface type='bridge'&gt;
      &lt;mac address='00:50:56:25:48:c7'/&gt;
      &lt;source bridge='VM Network'/&gt;
    &lt;/interface&gt;
  &lt;/devices&gt;
&lt;/domain&gt;
EOF

$ virsh -c esx://example.com domxml-to-native vmware-vmx demo.xml
Enter username for example.com [root]:
Enter root password for example.com:
config.version = "8"
virtualHW.version = "4"
guestOS = "other-64"
uuid.bios = "50 11 5e 16 9b dc 49 d7-f1 71 53 c4 d7 f9 17 10"
displayName = "Fedora11"
memsize = "1024"
numvcpus = "1"
scsi0.present = "true"
scsi0.virtualDev = "lsilogic"
scsi0:0.present = "true"
scsi0:0.deviceType = "scsi-hardDisk"
scsi0:0.fileName = "/vmfs/volumes/local-storage/Fedora11/Fedora11.vmdk"
ethernet0.present = "true"
ethernet0.networkName = "VM Network"
ethernet0.connectionType = "bridged"
ethernet0.addressType = "static"
ethernet0.address = "00:50:56:25:48:C7"
</pre>


678
    <h2><a id="xmlconfig">Example domain XML configs</a></h2>
679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694

    <h3>Fedora11 on x86_64</h3>
<pre>
&lt;domain type='vmware'&gt;
  &lt;name&gt;Fedora11&lt;/name&gt;
  &lt;uuid&gt;50115e16-9bdc-49d7-f171-53c4d7f91710&lt;/uuid&gt;
  &lt;memory&gt;1048576&lt;/memory&gt;
  &lt;currentMemory&gt;1048576&lt;/currentMemory&gt;
  &lt;vcpu&gt;1&lt;/vcpu&gt;
  &lt;os&gt;
    &lt;type arch='x86_64'&gt;hvm&lt;/type&gt;
  &lt;/os&gt;
  &lt;devices&gt;
    &lt;disk type='file' device='disk'&gt;
      &lt;source file='[local-storage] Fedora11/Fedora11.vmdk'/&gt;
      &lt;target dev='sda' bus='scsi'/&gt;
695
      &lt;address type='drive' controller='0' bus='0' unit='0'/&gt;
696
    &lt;/disk&gt;
697
    &lt;controller type='scsi' index='0'/&gt;
698 699 700 701 702 703 704 705 706
    &lt;interface type='bridge'&gt;
      &lt;mac address='00:50:56:25:48:c7'/&gt;
      &lt;source bridge='VM Network'/&gt;
    &lt;/interface&gt;
  &lt;/devices&gt;
&lt;/domain&gt;
</pre>


707
    <h2><a id="migration">Migration</a></h2>
708 709
    <p>
        A migration cannot be initiated on an ESX server directly, a VMware
710 711 712 713 714 715
        vCenter is necessary for this. The <code>vcenter</code> query
        parameter must be set either to the hostname or IP address of the
        vCenter managing the ESX server or to <code>*</code>. Setting it
        to <code>*</code> causes the driver to connect to the vCenter known to
        the ESX server. If the ESX server is not managed by a vCenter an error
        is reported.
716 717 718 719 720
    </p>
<pre>
esx://example.com/?vcenter=example-vcenter.com
</pre>
    <p>
721
        Here's an example how to migrate the domain <code>Fedora11</code> from
722
        ESX server <code>example-src.com</code> to ESX server
723
        <code>example-dst.com</code> implicitly involving vCenter
724 725 726
        <code>example-vcenter.com</code> using <code>virsh</code>.
    </p>
<pre>
727
$ virsh -c esx://example-src.com/?vcenter=* migrate Fedora11 esx://example-dst.com/?vcenter=*
728 729 730 731 732 733 734 735
Enter username for example-src.com [root]:
Enter root password for example-src.com:
Enter username for example-vcenter.com [administrator]:
Enter administrator password for example-vcenter.com:
Enter username for example-dst.com [root]:
Enter root password for example-dst.com:
Enter username for example-vcenter.com [administrator]:
Enter administrator password for example-vcenter.com:
736 737 738 739 740 741 742 743 744 745 746 747 748
</pre>
    <p>
        <span class="since">Since 0.8.3</span> you can directly connect to a vCenter.
        This simplifies migration a bit. Here's the same migration as above but
        using <code>vpx://</code> connections and assuming both ESX server are in
        datacenter <code>dc1</code> and aren't part of a cluster.
    </p>
<pre>
$ virsh -c vpx://example-vcenter.com/dc1/example-src.com migrate Fedora11 vpx://example-vcenter.com/dc1/example-dst.com
Enter username for example-vcenter.com [administrator]:
Enter administrator password for example-vcenter.com:
Enter username for example-vcenter.com [administrator]:
Enter administrator password for example-vcenter.com:
749 750 751
</pre>


752
    <h2><a id="scheduler">Scheduler configuration</a></h2>
753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
    <p>
        The driver exposes the ESX CPU scheduler. The parameters listed below
        are available to control the scheduler.
    </p>
    <dl>
        <dt><code>reservation</code></dt>
        <dd>
            The amount of CPU resource in MHz that is guaranteed to be
            available to the domain. Valid values are 0 and greater.
        </dd>
        <dt><code>limit</code></dt>
        <dd>
            The CPU utilization of the domain will be
            limited to this value in MHz, even if more CPU resources are
            available. If the limit is set to -1, the CPU utilization of the
            domain is unlimited. If the limit is not set to -1, it must be
            greater than or equal to the reservation.
        </dd>
        <dt><code>shares</code></dt>
        <dd>
            Shares are used to determine relative CPU
            allocation between domains. In general, a domain with more shares
            gets proportionally more of the CPU resource. Valid values are 0
            and greater. The special values -1, -2 and -3 represent the
            predefined shares level <code>low</code>, <code>normal</code> and
            <code>high</code>.
        </dd>
    </dl>


783
    <h2><a id="tools">VMware tools</a></h2>
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798
    <p>
        Some actions require installed VMware tools. If the VMware tools are
        not installed in the guest and one of the actions below is to be
        performed the ESX server raises an error and the driver reports it.
    </p>
    <ul>
        <li>
            <code>virDomainReboot</code>
        </li>
        <li>
            <code>virDomainShutdown</code>
        </li>
    </ul>


799
    <h2><a id="links">Links</a></h2>
800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815
    <ul>
        <li>
            <a href="http://www.vmware.com/support/developer/vc-sdk/">
                VMware vSphere Web Services SDK Documentation
            </a>
        </li>
        <li>
            <a href="http://www.vmware.com/pdf/esx3_memory.pdf">
                The Role of Memory in VMware ESX Server 3
            </a>
        </li>
        <li>
            <a href="http://www.sanbarrow.com/vmx.html">
                VMware VMX config parameters
            </a>
        </li>
816 817 818 819 820
        <li>
            <a href="http://www.vmware.com/pdf/vsp_4_pvscsi_perf.pdf">
                VMware ESX 4.0 PVSCSI Storage Performance
            </a>
        </li>
821 822
    </ul>
</body></html>