1 <?xml version="1.0" encoding="utf-8"?>
2 <database title="Open vSwitch Configuration Database">
4 A database with this schema holds the configuration for one Open
5 vSwitch daemon. The top-level configuration for the daemon is the
6 <ref table="Open_vSwitch"/> table, which must have exactly one
7 record. Records in other tables are significant only when they
8 can be reached directly or indirectly from the <ref
9 table="Open_vSwitch"/> table. Records that are not reachable from
10 the <ref table="Open_vSwitch"/> table are automatically deleted
11 from the database, except for records in a few distinguished
12 ``root set'' tables noted below.
15 <table name="Open_vSwitch" title="Open vSwitch configuration.">
16 Configuration for an Open vSwitch daemon. There must be exactly
17 one record in the <ref table="Open_vSwitch"/> table.
19 <group title="Configuration">
20 <column name="bridges">
21 Set of bridges managed by the daemon.
25 SSL used globally by the daemon.
28 <column name="other_config">
29 Key-value pairs for configuring rarely used Open vSwitch features. The
30 currently defined key-value pairs are:
32 <dt><code>enable-statistics</code></dt>
34 Set to <code>true</code> to enable populating the <ref
35 column="statistics"/> column or <code>false</code> (the default)
36 disable populating it.
41 <column name="external_ids">
42 Key-value pairs for use by external frameworks that integrate
43 with Open vSwitch, rather than by Open vSwitch itself. System
44 integrators should either use the Open vSwitch development
45 mailing list to coordinate on common key-value definitions, or
46 choose key names that are likely to be unique. The currently
47 defined common key-value pairs are:
49 <dt><code>system-id</code></dt>
50 <dd>A unique identifier for the Open vSwitch's physical host.
51 The form of the identifier depends on the type of the host.
52 On a Citrix XenServer, this will likely be the same as
53 <ref column="external_ids" key="xs-system-uuid"/>.</dd>
54 <dt><code>xs-system-uuid</code></dt>
55 <dd>The Citrix XenServer universally unique identifier for the
56 physical host as displayed by <code>xe host-list</code>.</dd>
61 <group title="Status">
62 <column name="next_cfg">
63 Sequence number for client to increment. When a client modifies
64 any part of the database configuration and wishes to wait for
65 Open vSwitch to finish applying the changes, it may increment
69 <column name="cur_cfg">
70 Sequence number that Open vSwitch sets to the current value of
71 <ref column="next_cfg"/> after it finishes applying a set of
72 configuration changes.
75 <column name="capabilities">
76 Describes functionality supported by the hardware and software platform
77 on which this Open vSwitch is based. Clients should not modify this
78 column. See the <ref table="Capability"/> description for defined
79 capability categories and the meaning of associated
80 <ref table="Capability"/> records.
83 <column name="statistics">
85 Key-value pairs that report statistics about a system running an Open
86 vSwitch. These are updated periodically (currently, every 5
87 seconds). Key-value pairs that cannot be determined or that do not
88 apply to a platform are omitted.
92 Statistics are disabled unless <ref column="other-config"
93 key="enable-statistics"/> is set to <code>true</code>.
97 <dt><code>cpu</code></dt>
100 Number of CPU processors, threads, or cores currently online and
101 available to the operating system on which Open vSwitch is
102 running, as an integer. This may be less than the number
103 installed, if some are not online or if they are not available to
104 the operating system.
107 Open vSwitch userspace processes are not multithreaded, but the
108 Linux kernel-based datapath is.
112 <dt><code>load_average</code></dt>
115 A comma-separated list of three floating-point numbers,
116 representing the system load average over the last 1, 5, and 15
117 minutes, respectively.
121 <dt><code>memory</code></dt>
124 A comma-separated list of integers, each of which represents a
125 quantity of memory in kilobytes that describes the operating
126 system on which Open vSwitch is running. In respective order,
131 <li>Total amount of RAM allocated to the OS.</li>
132 <li>RAM allocated to the OS that is in use.</li>
133 <li>RAM that can be flushed out to disk or otherwise discarded
134 if that space is needed for another purpose. This number is
135 necessarily less than or equal to the previous value.</li>
136 <li>Total disk space allocated for swap.</li>
137 <li>Swap space currently in use.</li>
141 On Linux, all five values can be determined and are included. On
142 other operating systems, only the first two values can be
143 determined, so the list will only have two values.
147 <dt><code>process_</code><var>name</var></dt>
150 One such key-value pair will exist for each running Open vSwitch
151 daemon process, with <var>name</var> replaced by the daemon's
152 name (e.g. <code>process_ovs-vswitchd</code>). The value is a
153 comma-separated list of integers. The integers represent the
154 following, with memory measured in kilobytes and durations in
159 <li>The process's virtual memory size.</li>
160 <li>The process's resident set size.</li>
161 <li>The amount of user and system CPU time consumed by the
163 <li>The number of times that the process has crashed and been
164 automatically restarted by the monitor.</li>
165 <li>The duration since the process was started.</li>
166 <li>The duration for which the process has been running.</li>
170 The interpretation of some of these values depends on whether the
171 process was started with the <option>--monitor</option>. If it
172 was not, then the crash count will always be 0 and the two
173 durations will always be the same. If <option>--monitor</option>
174 was given, then the crash count may be positive; if it is, the
175 latter duration is the amount of time since the most recent crash
180 There will be one key-value pair for each file in Open vSwitch's
181 ``run directory'' (usually <code>/var/run/openvswitch</code>)
182 whose name ends in <code>.pid</code>, whose contents are a
183 process ID, and which is locked by a running process. The
184 <var>name</var> is taken from the pidfile's name.
188 Currently Open vSwitch is only able to obtain all of the above
189 detail on Linux systems. On other systems, the same key-value
190 pairs will be present but the values will always be the empty
195 <dt><code>file_systems</code></dt>
198 A space-separated list of information on local, writable file
199 systems. Each item in the list describes one file system and
200 consists in turn of a comma-separated list of the following:
204 <li>Mount point, e.g. <code>/</code> or <code>/var/log</code>.
205 Any spaces or commas in the mount point are replaced by
207 <li>Total size, in kilobytes, as an integer.</li>
208 <li>Amount of storage in use, in kilobytes, as an integer.</li>
212 This key-value pair is omitted if there are no local, writable
213 file systems or if Open vSwitch cannot obtain the needed
221 <group title="Version Reporting">
223 These columns report the types and versions of the hardware and
224 software running Open vSwitch. We recommend in general that software
225 should test whether specific features are supported instead of relying
226 on version number checks. These values are primarily intended for
227 reporting to human administrators.
230 <column name="ovs_version">
231 The Open vSwitch version number, e.g. <code>1.1.0</code>.
232 If Open vSwitch was configured with a build number, then it is
233 also included, e.g. <code>1.1.0+build6579</code>.
236 <column name="db_version">
238 The database schema version number in the form
239 <code><var>major</var>.<var>minor</var>.<var>tweak</var></code>,
240 e.g. <code>1.2.3</code>. Whenever the database schema is changed in
241 a non-backward compatible way (e.g. deleting a column or a table),
242 <var>major</var> is incremented. When the database schema is changed
243 in a backward compatible way (e.g. adding a new column),
244 <var>minor</var> is incremented. When the database schema is changed
245 cosmetically (e.g. reindenting its syntax), <var>tweak</var> is
250 The schema version is part of the database schema, so it can also be
251 retrieved by fetching the schema using the Open vSwitch database
256 <column name="system_type">
258 An identifier for the type of system on top of which Open vSwitch
259 runs, e.g. <code>XenServer</code> or <code>KVM</code>.
262 System integrators are responsible for choosing and setting an
263 appropriate value for this column.
267 <column name="system_version">
269 The version of the system identified by <ref column="system_type"/>,
270 e.g. <code>5.6.100-39265p</code> on XenServer 5.6.100 build 39265.
273 System integrators are responsible for choosing and setting an
274 appropriate value for this column.
280 <group title="Database Configuration">
282 These columns primarily configure the Open vSwitch database
283 (<code>ovsdb-server</code>), not the Open vSwitch switch
284 (<code>ovs-vswitchd</code>). The OVSDB database also uses the <ref
285 column="ssl"/> settings.
289 The Open vSwitch switch does read the database configuration to
290 determine remote IP addresses to which in-band control should apply.
293 <column name="manager_options">
294 Database clients to which the Open vSwitch database server should
295 connect or to which it should listen, along with options for how these
296 connection should be configured. See the <ref table="Manager"/> table
297 for more information.
302 <table name="Bridge">
304 Configuration for a bridge within an
305 <ref table="Open_vSwitch"/>.
308 A <ref table="Bridge"/> record represents an Ethernet switch with one or
309 more ``ports,'' which are the <ref table="Port"/> records pointed to by
310 the <ref table="Bridge"/>'s <ref column="ports"/> column.
313 <group title="Core Features">
315 Bridge identifier. Should be alphanumeric and no more than about 8
316 bytes long. Must be unique among the names of ports, interfaces, and
320 <column name="ports">
321 Ports included in the bridge.
324 <column name="mirrors">
325 Port mirroring configuration.
328 <column name="netflow">
329 NetFlow configuration.
332 <column name="sflow">
336 <column name="flood_vlans">
338 VLAN IDs of VLANs on which MAC address learning should be disabled,
339 so that packets are flooded instead of being sent to specific ports
340 that are believed to contain packets' destination MACs. This should
341 ordinarily be used to disable MAC learning on VLANs used for
342 mirroring (RSPAN VLANs). It may also be useful for debugging.
345 SLB bonding (see the <ref table="Port" column="bond_mode"/> column in
346 the <ref table="Port"/> table) is incompatible with
347 <code>flood_vlans</code>. Consider using another bonding mode or
348 a different type of mirror instead.
353 <group title="OpenFlow Configuration">
354 <column name="controller">
355 OpenFlow controller set. If unset, then no OpenFlow controllers
359 <column name="fail_mode">
360 <p>When a controller is configured, it is, ordinarily, responsible
361 for setting up all flows on the switch. Thus, if the connection to
362 the controller fails, no new network connections can be set up.
363 If the connection to the controller stays down long enough,
364 no packets can pass through the switch at all. This setting
365 determines the switch's response to such a situation. It may be set
366 to one of the following:
368 <dt><code>standalone</code></dt>
369 <dd>If no message is received from the controller for three
370 times the inactivity probe interval
371 (see <ref column="inactivity_probe"/>), then Open vSwitch
372 will take over responsibility for setting up flows. In
373 this mode, Open vSwitch causes the bridge to act like an
374 ordinary MAC-learning switch. Open vSwitch will continue
375 to retry connecting to the controller in the background
376 and, when the connection succeeds, it will discontinue its
377 standalone behavior.</dd>
378 <dt><code>secure</code></dt>
379 <dd>Open vSwitch will not set up flows on its own when the
380 controller connection fails or when no controllers are
381 defined. The bridge will continue to retry connecting to
382 any defined controllers forever.</dd>
385 <p>If this value is unset, the default is implementation-specific.</p>
386 <p>When more than one controller is configured,
387 <ref column="fail_mode"/> is considered only when none of the
388 configured controllers can be contacted.</p>
391 <column name="datapath_id">
392 Reports the OpenFlow datapath ID in use. Exactly 16 hex digits.
393 (Setting this column has no useful effect. Set <ref
394 column="other-config" key="datapath-id"/> instead.)
398 <group title="Other Features">
399 <column name="datapath_type">
400 Name of datapath provider. The kernel datapath has
401 type <code>system</code>. The userspace datapath has
402 type <code>netdev</code>.
405 <column name="external_ids">
406 Key-value pairs for use by external frameworks that integrate
407 with Open vSwitch, rather than by Open vSwitch itself. System
408 integrators should either use the Open vSwitch development
409 mailing list to coordinate on common key-value definitions, or
410 choose key names that are likely to be unique. The currently
411 defined key-value pairs are:
413 <dt><code>bridge-id</code></dt>
414 <dd>A unique identifier of the bridge. On Citrix XenServer this will
415 commonly be the same as
416 <ref column="external_ids" key="xs-network-uuids"/>.</dd>
417 <dt><code>xs-network-uuids</code></dt>
418 <dd>Semicolon-delimited set of universally unique identifier(s) for
419 the network with which this bridge is associated on a Citrix
420 XenServer host. The network identifiers are RFC 4122 UUIDs as
421 displayed by, e.g., <code>xe network-list</code>.</dd>
425 <column name="other_config">
426 Key-value pairs for configuring rarely used bridge
427 features. The currently defined key-value pairs are:
429 <dt><code>datapath-id</code></dt>
431 digits to set the OpenFlow datapath ID to a specific
432 value. May not be all-zero.</dd>
433 <dt><code>disable-in-band</code></dt>
434 <dd>If set to <code>true</code>, disable in-band control on
435 the bridge regardless of controller and manager settings.</dd>
436 <dt><code>hwaddr</code></dt>
437 <dd>An Ethernet address in the form
438 <var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>
439 to set the hardware address of the local port and influence the
441 <dt><code>in-band-queue</code></dt>
443 A queue ID as a nonnegative integer. This sets the OpenFlow queue
444 ID that will be used by flows set up by in-band control on this
445 bridge. If unset, or if the port used by an in-band control flow
446 does not have QoS configured, or if the port does not have a queue
447 with the specified ID, the default queue is used instead.
449 <dt><code>flow-eviction-threshold</code></dt>
451 A number of flows as a nonnegative integer. This sets number
452 of flows at which eviction from the kernel flow table will
454 If there are a large number of flows then increasing this
455 value to around the number of flows present
456 can result in reduced CPU usage and packet loss.
462 Values below 100 will be rounded up to 100.
464 <dt><code>forward-bpdu</code></dt>
466 Option to allow forwarding of BPDU frames when NORMAL
467 action if invoked. Frames with reserved Ethernet addresses
468 (e.g. STP BPDU) will be forwarded when this option is enabled.
469 If the Open vSwitch bridge is used to connect different
470 Ethernet networks, and if Open vSwitch node does not run STP,
471 then this option should be enabled.
472 Default is disabled, set to <code>true</code> to enable.
479 <table name="Port" table="Port or bond configuration.">
480 <p>A port within a <ref table="Bridge"/>.</p>
481 <p>Most commonly, a port has exactly one ``interface,'' pointed to by its
482 <ref column="interfaces"/> column. Such a port logically
483 corresponds to a port on a physical Ethernet switch. A port
484 with more than one interface is a ``bonded port'' (see
485 <ref group="Bonding Configuration"/>).</p>
486 <p>Some properties that one might think as belonging to a port are actually
487 part of the port's <ref table="Interface"/> members.</p>
490 Port name. Should be alphanumeric and no more than about 8
491 bytes long. May be the same as the interface name, for
492 non-bonded ports. Must otherwise be unique among the names of
493 ports, interfaces, and bridges on a host.
496 <column name="interfaces">
497 The port's interfaces. If there is more than one, this is a
501 <group title="VLAN Configuration">
502 <p>Bridge ports support the following types of VLAN configuration:</p>
507 A trunk port carries packets on one or more specified VLANs
508 specified in the <ref column="trunks"/> column (often, on every
509 VLAN). A packet that ingresses on a trunk port is in the VLAN
510 specified in its 802.1Q header, or VLAN 0 if the packet has no
511 802.1Q header. A packet that egresses through a trunk port will
512 have a 802.1Q header if it has a nonzero VLAN ID (or a nonzero
517 Any packet that ingresses on a trunk port tagged with a VLAN that
518 the port does not trunk is dropped.
525 An access port carries packets on exactly one VLAN specified in the
526 <ref column="tag"/> column. Packets ingressing and egressing on an
527 access port have no 802.1Q header.
531 Any packet with an 802.1Q header that ingresses on an access port
532 is dropped, regardless of whether the VLAN ID in the header is the
533 access port's VLAN ID.
537 <dt>native-tagged</dt>
539 A native-tagged port resembles a trunk port, with the exception that
540 a packet without an 802.1Q header that ingresses on a native-tagged
541 port is in the ``native VLAN'' (specified in the <ref column="tag"/>
545 <dt>native-untagged</dt>
547 A native-untagged port resembles a native-tagged port, with the
548 exception that a packet that egresses on a native-untagged port in
549 the native VLAN not have an 802.1Q header.
553 A packet will only egress through bridge ports that carry the VLAN of
554 the packet, as described by the rules above.
557 <column name="vlan_mode">
559 The VLAN mode of the port, as described above. When this column is
560 empty, a default mode is selected as follows:
564 If <ref column="tag"/> contains a value, the port is an access
565 port. The <ref column="trunks"/> column should be empty.
568 Otherwise, the port is a trunk port. The <ref column="trunks"/>
569 column value is honored if it is present.
576 For an access port, the port's implicitly tagged VLAN. For a
577 native-tagged or native-untagged port, the port's native VLAN. Must
578 be empty if this is a trunk port.
582 <column name="trunks">
584 For a trunk, native-tagged, or native-untagged port, the 802.1Q VLAN
585 or VLANs that this port trunks; if it is empty, then the port trunks
586 all VLANs. Must be empty if this is an access port.
589 A native-tagged or native-untagged port always trunks its native
590 VLAN, regardless of whether <ref column="trunks"/> includes that
596 <group title="Bonding Configuration">
597 <p>A port that has more than one interface is a ``bonded port.'' Bonding
598 allows for load balancing and fail-over. Some kinds of bonding will
599 work with any kind of upstream switch:</p>
602 <dt><code>balance-slb</code></dt>
604 Balances flows among slaves based on source MAC address and output
605 VLAN, with periodic rebalancing as traffic patterns change.
608 <dt><code>active-backup</code></dt>
610 Assigns all flows to one slave, failing over to a backup slave when
611 the active slave is disabled.
616 The following modes require the upstream switch to support 802.3ad with
617 successful LACP negotiation. If LACP negotiation fails then
618 <code>balance-slb</code> style flow hashing is used as a fallback:
622 <dt><code>balance-tcp</code></dt>
624 Balances flows among slaves based on L2, L3, and L4 protocol
625 information such as destination MAC address, IP address, and TCP
631 <dt><code>stable</code></dt>
633 <p>Attempts to always assign a given flow to the same slave
634 consistently. In an effort to maintain stability, no load
635 balancing is done. Uses a similar hashing strategy to
636 <code>balance-tcp</code>, always taking into account L3 and L4
637 fields even if LACP negotiations are unsuccessful. </p>
638 <p>Slave selection decisions are made based on <ref table="Interface"
639 column="other_config" key="bond-stable-id"/> if set. Otherwise,
640 OpenFlow port number is used. Decisions are consistent across all
641 <code>ovs-vswitchd</code> instances with equivalent
642 <ref table="Interface" column="other_config" key="bond-stable-id"/>
647 <p>These columns apply only to bonded ports. Their values are
648 otherwise ignored.</p>
650 <column name="bond_mode">
651 <p>The type of bonding used for a bonded port. Defaults to
652 <code>balance-slb</code> if unset.
656 <column name="bond_updelay">
657 <p>For a bonded port, the number of milliseconds for which carrier must
658 stay up on an interface before the interface is considered to be up.
659 Specify <code>0</code> to enable the interface immediately.</p>
660 <p>This setting is honored only when at least one bonded interface is
661 already enabled. When no interfaces are enabled, then the first bond
662 interface to come up is enabled immediately.</p>
665 <column name="bond_downdelay">
666 For a bonded port, the number of milliseconds for which carrier must
667 stay down on an interface before the interface is considered to be
668 down. Specify <code>0</code> to disable the interface immediately.
671 <column name="bond_fake_iface">
672 For a bonded port, whether to create a fake internal interface with the
673 name of the port. Use only for compatibility with legacy software that
678 <p>Configures LACP on this port. LACP allows directly connected
679 switches to negotiate which links may be bonded. LACP may be enabled
680 on non-bonded ports for the benefit of any switches they may be
681 connected to. <code>active</code> ports are allowed to initiate LACP
682 negotiations. <code>passive</code> ports are allowed to participate
683 in LACP negotiations initiated by a remote switch, but not allowed to
684 initiate such negotiations themselves. Defaults to <code>off</code>
690 <group title="Other Features">
692 Quality of Service configuration for this port.
696 The MAC address to use for this port for the purpose of choosing the
697 bridge's MAC address. This column does not necessarily reflect the
698 port's actual MAC address, nor will setting it change the port's actual
702 <column name="fake_bridge">
703 Does this port represent a sub-bridge for its tagged VLAN within the
704 Bridge? See ovs-vsctl(8) for more information.
707 <column name="external_ids">
709 Key-value pairs for use by external frameworks that integrate with
710 Open vSwitch, rather than by Open vSwitch itself. System integrators
711 should either use the Open vSwitch development mailing list to
712 coordinate on common key-value definitions, or choose key names that
713 are likely to be unique.
716 No key-value pairs native to <ref table="Port"/> are currently
717 defined. For fake bridges (see the <ref column="fake_bridge"/>
718 column), external IDs for the fake bridge are defined here by
719 prefixing a <ref table="Bridge"/> <ref table="Bridge"
720 column="external_ids"/> key with <code>fake-bridge-</code>,
721 e.g. <code>fake-bridge-xs-network-uuids</code>.
725 <column name="other_config">
726 Key-value pairs for configuring rarely used port features. The
727 currently defined key-value pairs are:
729 <dt><code>bond-rebalance-interval</code></dt>
730 <dd>For an SLB bonded port, the number of milliseconds between
731 successive attempts to rebalance the bond, that is, to
732 move source MACs and their flows from one interface on
733 the bond to another in an attempt to keep usage of each
734 interface roughly equal. The default is 10000 (10
735 seconds), and the minimum is 1000 (1 second).</dd>
736 <dt><code>bond-detect-mode</code></dt>
737 <dd> Sets the method used to detect link failures in a bonded port.
738 Options are <code>carrier</code> and <code>miimon</code>. Defaults
739 to <code>carrier</code> which uses each interface's carrier to detect
740 failures. When set to <code>miimon</code>, will check for failures
741 by polling each interface's MII. </dd>
742 <dt><code>bond-miimon-interval</code></dt>
743 <dd> The number of milliseconds between successive attempts to
744 poll each interface's MII. Only relevant on ports which use
745 <code>miimon</code> to detect failures. </dd>
746 <dt><code>bond-hash-basis</code></dt>
747 <dd> An integer hashed along with flows when choosing output slaves.
748 When changed, all flows will be assigned different hash values
749 possibly causing slave selection decisions to change.</dd>
750 <dt><code>lacp-system-id</code></dt>
751 <dd> The LACP system ID of this <ref table="Port"/>. The system ID
752 of a LACP bond is used to identify itself to its partners. Must
753 be a nonzero MAC address.</dd>
754 <dt><code>lacp-system-priority</code></dt>
755 <dd> The LACP system priority of this <ref table="Port"/>. In
756 LACP negotiations, link status decisions are made by the system
757 with the numerically lower priority. Must be a number between 1
759 <dt><code>lacp-time</code></dt>
761 <p>The LACP timing which should be used on this
762 <ref table="Port"/>. Possible values are <code>fast</code>,
763 <code>slow</code> and a positive number of milliseconds. By
764 default <code>slow</code> is used. When configured to be
765 <code>fast</code> LACP heartbeats are requested at a rate of once
766 per second causing connectivity problems to be detected more
767 quickly. In <code>slow</code> mode, heartbeats are requested at
768 a rate of once every 30 seconds.</p>
770 <p>Users may manually set a heartbeat transmission rate to increase
771 the fault detection speed further. When manually set, OVS
772 expects the partner switch to be configured with the same
773 transmission rate. Manually setting <code>lacp-time</code> to
774 something other than <code>fast</code> or <code>slow</code> is
775 not supported by the LACP specification.</p>
777 <dt><code>lacp-heartbeat</code></dt>
778 <dd> Treats LACP like a simple heartbeat protocol for link state
779 monitoring. Most features of the LACP protocol are disabled when
780 this mode is in use.</dd>
786 <table name="Interface" title="One physical network device in a Port.">
787 An interface within a <ref table="Port"/>.
789 <group title="Core Features">
791 Interface name. Should be alphanumeric and no more than about 8 bytes
792 long. May be the same as the port name, for non-bonded ports. Must
793 otherwise be unique among the names of ports, interfaces, and bridges
798 <p>Ethernet address to set for this interface. If unset then the
799 default MAC address is used:</p>
801 <li>For the local interface, the default is the lowest-numbered MAC
802 address among the other bridge ports, either the value of the
803 <ref table="Port" column="mac"/> in its <ref table="Port"/> record,
804 if set, or its actual MAC (for bonded ports, the MAC of its slave
805 whose name is first in alphabetical order). Internal ports and
806 bridge ports that are used as port mirroring destinations (see the
807 <ref table="Mirror"/> table) are ignored.</li>
808 <li>For other internal interfaces, the default MAC is randomly
810 <li>External interfaces typically have a MAC address associated with
813 <p>Some interfaces may not have a software-controllable MAC
817 <column name="ofport">
818 <p>OpenFlow port number for this interface. Unlike most columns, this
819 column's value should be set only by Open vSwitch itself. Other
820 clients should set this column to an empty set (the default) when
821 creating an <ref table="Interface"/>.</p>
822 <p>Open vSwitch populates this column when the port number becomes
823 known. If the interface is successfully added,
824 <ref column="ofport"/> will be set to a number between 1 and 65535
825 (generally either in the range 1 to 65279, inclusive, or 65534, the
826 port number for the OpenFlow ``local port''). If the interface
827 cannot be added then Open vSwitch sets this column
832 <group title="System-Specific Details">
834 The interface type, one of:
836 <dt><code>system</code></dt>
837 <dd>An ordinary network device, e.g. <code>eth0</code> on Linux.
838 Sometimes referred to as ``external interfaces'' since they are
839 generally connected to hardware external to that on which the Open
840 vSwitch is running. The empty string is a synonym for
841 <code>system</code>.</dd>
842 <dt><code>internal</code></dt>
843 <dd>A simulated network device that sends and receives traffic. An
844 internal interface whose <ref column="name"/> is the same as its
845 bridge's <ref table="Open_vSwitch" column="name"/> is called the
846 ``local interface.'' It does not make sense to bond an internal
847 interface, so the terms ``port'' and ``interface'' are often used
848 imprecisely for internal interfaces.</dd>
849 <dt><code>tap</code></dt>
850 <dd>A TUN/TAP device managed by Open vSwitch.</dd>
851 <dt><code>gre</code></dt>
852 <dd>An Ethernet over RFC 2890 Generic Routing Encapsulation over IPv4
853 tunnel. Each tunnel must be uniquely identified by the
854 combination of <ref column="options" key="remote_ip"/>,
855 <ref column="options" key="local_ip"/>, and
856 <ref column="options" key="in_key"/>. Note that if two ports
857 are defined that are the same except one has an optional
858 identifier and the other does not, the more specific one is
859 matched first. <ref column="options" key="in_key"/> is considered
860 more specific than <ref column="options" key="local_ip"/> if a port
861 defines one and another port defines the other. The following
862 options may be specified in the <ref column="options"/> column:
864 <dt><code>remote_ip</code></dt>
865 <dd>Required. The tunnel endpoint.</dd>
868 <dt><code>local_ip</code></dt>
869 <dd>Optional. The destination IP that received packets must
870 match. Default is to match all addresses.</dd>
873 <dt><code>in_key</code></dt>
874 <dd>Optional. The GRE key that received packets must contain.
875 It may either be a 32-bit number (no key and a key of 0 are
876 treated as equivalent) or the word <code>flow</code>. If
877 <code>flow</code> is specified then any key will be accepted
878 and the key will be placed in the <code>tun_id</code> field
879 for matching in the flow table. The ovs-ofctl manual page
880 contains additional information about matching fields in
881 OpenFlow flows. Default is no key.</dd>
884 <dt><code>out_key</code></dt>
885 <dd>Optional. The GRE key to be set on outgoing packets. It may
886 either be a 32-bit number or the word <code>flow</code>. If
887 <code>flow</code> is specified then the key may be set using
888 the <code>set_tunnel</code> Nicira OpenFlow vendor extension (0
889 is used in the absence of an action). The ovs-ofctl manual
890 page contains additional information about the Nicira OpenFlow
891 vendor extensions. Default is no key.</dd>
894 <dt><code>key</code></dt>
895 <dd>Optional. Shorthand to set <code>in_key</code> and
896 <code>out_key</code> at the same time.</dd>
899 <dt><code>tos</code></dt>
900 <dd>Optional. The value of the ToS bits to be set on the
901 encapsulating packet. It may also be the word
902 <code>inherit</code>, in which case the ToS will be copied from
903 the inner packet if it is IPv4 or IPv6 (otherwise it will be
904 0). Note that the ECN fields are always inherited. Default is
908 <dt><code>ttl</code></dt>
909 <dd>Optional. The TTL to be set on the encapsulating packet.
910 It may also be the word <code>inherit</code>, in which case the
911 TTL will be copied from the inner packet if it is IPv4 or IPv6
912 (otherwise it will be the system default, typically 64).
913 Default is the system default TTL.</dd>
916 <dt><code>csum</code></dt>
917 <dd>Optional. Compute GRE checksums on outgoing packets.
918 Checksums present on incoming packets will be validated
919 regardless of this setting. Note that GRE checksums
920 impose a significant performance penalty as they cover the
921 entire packet. As the contents of the packet is typically
922 covered by L3 and L4 checksums, this additional checksum only
923 adds value for the GRE and encapsulated Ethernet headers.
924 Default is disabled, set to <code>true</code> to enable.</dd>
927 <dt><code>df_inherit</code></dt>
928 <dd>Optional. If enabled, the Don't Fragment bit will be copied
929 from the inner IP headers (those of the encapsulated traffic)
930 to the outer (tunnel) headers. Default is disabled; set to
931 <code>true</code> to enable.</dd>
934 <dt><code>df_default</code></dt>
935 <dd>Optional. If enabled, the Don't Fragment bit will be set by
936 default on tunnel headers if the <code>df_inherit</code> option
937 is not set, or if the encapsulated packet is not IP. Default
938 is enabled; set to <code>false</code> to disable.</dd>
941 <dt><code>pmtud</code></dt>
942 <dd>Optional. Enable tunnel path MTU discovery. If enabled
943 ``ICMP Destination Unreachable - Fragmentation Needed''
944 messages will be generated for IPv4 packets with the DF bit set
945 and IPv6 packets above the minimum MTU if the packet size
946 exceeds the path MTU minus the size of the tunnel headers.
947 Note that this option causes behavior that is typically
948 reserved for routers and therefore is not entirely in
949 compliance with the IEEE 802.1D specification for bridges.
950 Default is enabled; set to <code>false</code> to disable.</dd>
953 <dt><code>header_cache</code></dt>
954 <dd>Optional. Enable caching of tunnel headers and the output
955 path. This can lead to a significant performance increase
956 without changing behavior. In general it should not be
957 necessary to adjust this setting. However, the caching can
958 bypass certain components of the IP stack (such as IP tables)
959 and it may be useful to disable it if these features are
960 required or as a debugging measure. Default is enabled, set to
961 <code>false</code> to disable.</dd>
964 <dt><code>ipsec_gre</code></dt>
965 <dd>An Ethernet over RFC 2890 Generic Routing Encapsulation
966 over IPv4 IPsec tunnel. Each tunnel (including those of type
967 <code>gre</code>) must be uniquely identified by the
968 combination of <ref column="options" key="remote_ip"/> and
969 <ref column="options" key="local_ip"/>. Note that if two ports are
970 defined that are the same except one has an optional identifier and
971 the other does not, the more specific one is matched first.
972 An authentication method of <ref column="options" key="peer_cert"/>
973 or <ref column="options" key="psk"/> must be defined. The
974 following options may be specified in the <ref column="options"/>
977 <dt><code>remote_ip</code></dt>
978 <dd>Required. The tunnel endpoint.</dd>
981 <dt><code>local_ip</code></dt>
982 <dd>Optional. The destination IP that received packets must
983 match. Default is to match all addresses.</dd>
986 <dt><code>peer_cert</code></dt>
987 <dd>Required for certificate authentication. A string
988 containing the peer's certificate in PEM format.
989 Additionally the host's certificate must be specified
990 with the <code>certificate</code> option.</dd>
993 <dt><code>certificate</code></dt>
994 <dd>Required for certificate authentication. The name of a
995 PEM file containing a certificate that will be presented
996 to the peer during authentication.</dd>
999 <dt><code>private_key</code></dt>
1000 <dd>Optional for certificate authentication. The name of
1001 a PEM file containing the private key associated with
1002 <code>certificate</code>. If <code>certificate</code>
1003 contains the private key, this option may be omitted.</dd>
1006 <dt><code>psk</code></dt>
1007 <dd>Required for pre-shared key authentication. Specifies a
1008 pre-shared key for authentication that must be identical on
1009 both sides of the tunnel.</dd>
1012 <dt><code>in_key</code></dt>
1013 <dd>Optional. The GRE key that received packets must contain.
1014 It may either be a 32-bit number (no key and a key of 0 are
1015 treated as equivalent) or the word <code>flow</code>. If
1016 <code>flow</code> is specified then any key will be accepted
1017 and the key will be placed in the <code>tun_id</code> field
1018 for matching in the flow table. The ovs-ofctl manual page
1019 contains additional information about matching fields in
1020 OpenFlow flows. Default is no key.</dd>
1023 <dt><code>out_key</code></dt>
1024 <dd>Optional. The GRE key to be set on outgoing packets. It may
1025 either be a 32-bit number or the word <code>flow</code>. If
1026 <code>flow</code> is specified then the key may be set using
1027 the <code>set_tunnel</code> Nicira OpenFlow vendor extension (0
1028 is used in the absence of an action). The ovs-ofctl manual
1029 page contains additional information about the Nicira OpenFlow
1030 vendor extensions. Default is no key.</dd>
1033 <dt><code>key</code></dt>
1034 <dd>Optional. Shorthand to set <code>in_key</code> and
1035 <code>out_key</code> at the same time.</dd>
1038 <dt><code>tos</code></dt>
1039 <dd>Optional. The value of the ToS bits to be set on the
1040 encapsulating packet. It may also be the word
1041 <code>inherit</code>, in which case the ToS will be copied from
1042 the inner packet if it is IPv4 or IPv6 (otherwise it will be
1043 0). Note that the ECN fields are always inherited. Default is
1047 <dt><code>ttl</code></dt>
1048 <dd>Optional. The TTL to be set on the encapsulating packet.
1049 It may also be the word <code>inherit</code>, in which case the
1050 TTL will be copied from the inner packet if it is IPv4 or IPv6
1051 (otherwise it will be the system default, typically 64).
1052 Default is the system default TTL.</dd>
1055 <dt><code>csum</code></dt>
1056 <dd>Optional. Compute GRE checksums on outgoing packets.
1057 Checksums present on incoming packets will be validated
1058 regardless of this setting. Note that GRE checksums
1059 impose a significant performance penalty as they cover the
1060 entire packet. As the contents of the packet is typically
1061 covered by L3 and L4 checksums, this additional checksum only
1062 adds value for the GRE and encapsulated Ethernet headers.
1063 Default is disabled, set to <code>true</code> to enable.</dd>
1066 <dt><code>df_inherit</code></dt>
1067 <dd>Optional. If enabled, the Don't Fragment bit will be copied
1068 from the inner IP headers (those of the encapsulated traffic)
1069 to the outer (tunnel) headers. Default is disabled; set to
1070 <code>true</code> to enable.</dd>
1073 <dt><code>df_default</code></dt>
1074 <dd>Optional. If enabled, the Don't Fragment bit will be set by
1075 default on tunnel headers if the <code>df_inherit</code> option
1076 is not set, or if the encapsulated packet is not IP. Default
1077 is enabled; set to <code>false</code> to disable.</dd>
1080 <dt><code>pmtud</code></dt>
1081 <dd>Optional. Enable tunnel path MTU discovery. If enabled
1082 ``ICMP Destination Unreachable - Fragmentation Needed''
1083 messages will be generated for IPv4 packets with the DF bit set
1084 and IPv6 packets above the minimum MTU if the packet size
1085 exceeds the path MTU minus the size of the tunnel headers.
1086 Note that this option causes behavior that is typically
1087 reserved for routers and therefore is not entirely in
1088 compliance with the IEEE 802.1D specification for bridges.
1089 Default is enabled; set to <code>false</code> to disable.</dd>
1092 <dt><code>capwap</code></dt>
1093 <dd>Ethernet tunneling over the UDP transport portion of CAPWAP
1094 (RFC 5415). This allows interoperability with certain switches
1095 where GRE is not available. Note that only the tunneling component
1096 of the protocol is implemented. Due to the non-standard use of
1097 CAPWAP, UDP ports 58881 and 58882 are used as the source and
1098 destination ports respectively. Each tunnel must be uniquely
1099 identified by the combination of
1100 <ref column="options" key="remote_ip"/> and
1101 <ref column="options" key="local_ip"/>. If two ports are defined
1102 that are the same except one includes
1103 <ref column="options" key="local_ip"/> and the other does not, the
1104 more specific one is matched first. CAPWAP support is not
1105 available on all platforms. Currently it is only supported in the
1106 Linux kernel module with kernel versions >= 2.6.25. The following
1107 options may be specified in the <ref column="options"/> column:
1109 <dt><code>remote_ip</code></dt>
1110 <dd>Required. The tunnel endpoint.</dd>
1113 <dt><code>local_ip</code></dt>
1114 <dd>Optional. The destination IP that received packets must
1115 match. Default is to match all addresses.</dd>
1118 <dt><code>tos</code></dt>
1119 <dd>Optional. The value of the ToS bits to be set on the
1120 encapsulating packet. It may also be the word
1121 <code>inherit</code>, in which case the ToS will be copied from
1122 the inner packet if it is IPv4 or IPv6 (otherwise it will be
1123 0). Note that the ECN fields are always inherited. Default is
1127 <dt><code>ttl</code></dt>
1128 <dd>Optional. The TTL to be set on the encapsulating packet.
1129 It may also be the word <code>inherit</code>, in which case the
1130 TTL will be copied from the inner packet if it is IPv4 or IPv6
1131 (otherwise it will be the system default, typically 64).
1132 Default is the system default TTL.</dd>
1135 <dt><code>in_key</code></dt>
1136 <dd>Optional. The WSI key that received packets must contain.
1137 It may either be a 64-bit number (no key and a key of 0 are
1138 treated as equivalent) or the word <code>flow</code>. If
1139 <code>flow</code> is specified then any key will be accepted
1140 and the key will be placed in the <code>tun_id</code> field
1141 for matching in the flow table. The ovs-ofctl manual page
1142 contains additional information about matching fields in
1143 OpenFlow flows. Default is no key.</dd>
1146 <dt><code>out_key</code></dt>
1147 <dd>Optional. The WSI key to be set on outgoing packets. It may
1148 either be a 64-bit number or the word <code>flow</code>. If
1149 <code>flow</code> is specified then the key may be set using
1150 the <code>set_tunnel</code> Nicira OpenFlow vendor extension (0
1151 is used in the absence of an action). The ovs-ofctl manual
1152 page contains additional information about the Nicira OpenFlow
1153 vendor extensions. Default is no key.</dd>
1156 <dt><code>key</code></dt>
1157 <dd>Optional. Shorthand to set <code>in_key</code> and
1158 <code>out_key</code> at the same time.</dd>
1161 <dt><code>df_inherit</code></dt>
1162 <dd>Optional. If enabled, the Don't Fragment bit will be copied
1163 from the inner IP headers (those of the encapsulated traffic)
1164 to the outer (tunnel) headers. Default is disabled; set to
1165 <code>true</code> to enable.</dd>
1168 <dt><code>df_default</code></dt>
1169 <dd>Optional. If enabled, the Don't Fragment bit will be set by
1170 default on tunnel headers if the <code>df_inherit</code> option
1171 is not set, or if the encapsulated packet is not IP. Default
1172 is enabled; set to <code>false</code> to disable.</dd>
1175 <dt><code>pmtud</code></dt>
1176 <dd>Optional. Enable tunnel path MTU discovery. If enabled
1177 ``ICMP Destination Unreachable - Fragmentation Needed''
1178 messages will be generated for IPv4 packets with the DF bit set
1179 and IPv6 packets above the minimum MTU if the packet size
1180 exceeds the path MTU minus the size of the tunnel headers.
1181 Note that this option causes behavior that is typically
1182 reserved for routers and therefore is not entirely in
1183 compliance with the IEEE 802.1D specification for bridges.
1184 Default is enabled; set to <code>false</code> to disable.</dd>
1187 <dt><code>header_cache</code></dt>
1188 <dd>Optional. Enable caching of tunnel headers and the output
1189 path. This can lead to a significant performance increase
1190 without changing behavior. In general it should not be
1191 necessary to adjust this setting. However, the caching can
1192 bypass certain components of the IP stack (such as IP tables)
1193 and it may be useful to disable it if these features are
1194 required or as a debugging measure. Default is enabled, set to
1195 <code>false</code> to disable.</dd>
1198 <dt><code>patch</code></dt>
1201 A pair of virtual devices that act as a patch cable. The <ref
1202 column="options"/> column must have the following key-value pair:
1205 <dt><code>peer</code></dt>
1207 The <ref column="name"/> of the <ref table="Interface"/> for
1208 the other side of the patch. The named <ref
1209 table="Interface"/>'s own <code>peer</code> option must specify
1210 this <ref table="Interface"/>'s name. That is, the two patch
1211 interfaces must have reversed <ref column="name"/> and
1212 <code>peer</code> values.
1216 <dt><code>null</code></dt>
1217 <dd>An ignored interface.</dd>
1221 <column name="options">
1222 Configuration options whose interpretation varies based on
1223 <ref column="type"/>.
1227 <group title="Interface Status">
1229 Status information about interfaces attached to bridges, updated every
1230 5 seconds. Not all interfaces have all of these properties; virtual
1231 interfaces don't have a link speed, for example. Non-applicable
1232 columns will have empty values.
1234 <column name="admin_state">
1236 The administrative state of the physical network link.
1240 <column name="link_state">
1242 The observed state of the physical network link. This is ordinarily
1243 the link's carrier status. If the interface's <ref table="Port"/> is
1244 a bond configured for miimon monitoring, it is instead the network
1245 link's miimon status.
1249 <column name="link_speed">
1251 The negotiated speed of the physical network link.
1252 Valid values are positive integers greater than 0.
1256 <column name="duplex">
1258 The duplex mode of the physical network link.
1264 The MTU (maximum transmission unit); i.e. the largest
1265 amount of data that can fit into a single Ethernet frame.
1266 The standard Ethernet MTU is 1500 bytes. Some physical media
1267 and many kinds of virtual interfaces can be configured with
1271 This column will be empty for an interface that does not
1272 have an MTU as, for example, some kinds of tunnels do not.
1276 <column name="status">
1278 Key-value pairs that report port status. Supported status values are
1279 <ref column="type"/>-dependent; some interfaces may not have a valid
1280 <ref column="status" key="driver_name"/>, for example.
1282 <p>The currently defined key-value pairs are:</p>
1284 <dt><code>driver_name</code></dt>
1285 <dd>The name of the device driver controlling the network
1289 <dt><code>driver_version</code></dt>
1290 <dd>The version string of the device driver controlling the
1291 network adapter.</dd>
1294 <dt><code>firmware_version</code></dt>
1295 <dd>The version string of the network adapter's firmware, if
1299 <dt><code>source_ip</code></dt>
1300 <dd>The source IP address used for an IPv4 tunnel end-point,
1301 such as <code>gre</code> or <code>capwap</code>.</dd>
1304 <dt><code>tunnel_egress_iface</code></dt>
1305 <dd>Egress interface for tunnels. Currently only relevant for GRE
1306 and CAPWAP tunnels. On Linux systems, this column will show
1307 the name of the interface which is responsible for routing
1308 traffic destined for the configured
1309 <ref column="options" key="remote_ip"/>. This could be an
1310 internal interface such as a bridge port.</dd>
1313 <dt><code>tunnel_egress_iface_carrier</code></dt>
1314 <dd>Whether a carrier is detected on
1315 <ref column="status" key="tunnel_egress_iface"/>. Valid values
1316 are <code>down</code> and <code>up</code>.</dd>
1321 <group title="Ingress Policing">
1323 These settings control ingress policing for packets received on this
1324 interface. On a physical interface, this limits the rate at which
1325 traffic is allowed into the system from the outside; on a virtual
1326 interface (one connected to a virtual machine), this limits the rate at
1327 which the VM is able to transmit.
1330 Policing is a simple form of quality-of-service that simply drops
1331 packets received in excess of the configured rate. Due to its
1332 simplicity, policing is usually less accurate and less effective than
1333 egress QoS (which is configured using the <ref table="QoS"/> and <ref
1334 table="Queue"/> tables).
1337 Policing is currently implemented only on Linux. The Linux
1338 implementation uses a simple ``token bucket'' approach:
1342 The size of the bucket corresponds to <ref
1343 column="ingress_policing_burst"/>. Initially the bucket is full.
1346 Whenever a packet is received, its size (converted to tokens) is
1347 compared to the number of tokens currently in the bucket. If the
1348 required number of tokens are available, they are removed and the
1349 packet is forwarded. Otherwise, the packet is dropped.
1352 Whenever it is not full, the bucket is refilled with tokens at the
1353 rate specified by <ref column="ingress_policing_rate"/>.
1357 Policing interacts badly with some network protocols, and especially
1358 with fragmented IP packets. Suppose that there is enough network
1359 activity to keep the bucket nearly empty all the time. Then this token
1360 bucket algorithm will forward a single packet every so often, with the
1361 period depending on packet size and on the configured rate. All of the
1362 fragments of an IP packets are normally transmitted back-to-back, as a
1363 group. In such a situation, therefore, only one of these fragments
1364 will be forwarded and the rest will be dropped. IP does not provide
1365 any way for the intended recipient to ask for only the remaining
1366 fragments. In such a case there are two likely possibilities for what
1367 will happen next: either all of the fragments will eventually be
1368 retransmitted (as TCP will do), in which case the same problem will
1369 recur, or the sender will not realize that its packet has been dropped
1370 and data will simply be lost (as some UDP-based protocols will do).
1371 Either way, it is possible that no forward progress will ever occur.
1373 <column name="ingress_policing_rate">
1375 Maximum rate for data received on this interface, in kbps. Data
1376 received faster than this rate is dropped. Set to <code>0</code>
1377 (the default) to disable policing.
1381 <column name="ingress_policing_burst">
1382 <p>Maximum burst size for data received on this interface, in kb. The
1383 default burst size if set to <code>0</code> is 1000 kb. This value
1384 has no effect if <ref column="ingress_policing_rate"/>
1385 is <code>0</code>.</p>
1387 Specifying a larger burst size lets the algorithm be more forgiving,
1388 which is important for protocols like TCP that react severely to
1389 dropped packets. The burst size should be at least the size of the
1390 interface's MTU. Specifying a value that is numerically at least as
1391 large as 10% of <ref column="ingress_policing_rate"/> helps TCP come
1392 closer to achieving the full rate.
1397 <group title="Connectivity Fault Management">
1399 802.1ag Connectivity Fault Management (CFM) allows a group of
1400 Maintenance Points (MPs) called a Maintenance Association (MA) to
1401 detect connectivity problems with each other. MPs within a MA should
1402 have complete and exclusive interconnectivity. This is verified by
1403 occasionally broadcasting Continuity Check Messages (CCMs) at a
1404 configurable transmission interval.
1408 According to the 802.1ag specification, each Maintenance Point should
1409 be configured out-of-band with a list of Remote Maintenance Points it
1410 should have connectivity to. Open vSwitch differs from the
1411 specification in this area. It simply assumes the link is faulted if
1412 no Remote Maintenance Points are reachable, and considers it not
1416 <column name="cfm_mpid">
1417 A Maintenance Point ID (MPID) uniquely identifies each endpoint within
1418 a Maintenance Association. The MPID is used to identify this endpoint
1419 to other Maintenance Points in the MA. Each end of a link being
1420 monitored should have a different MPID. Must be configured to enable
1421 CFM on this <ref table="Interface"/>.
1424 <column name="cfm_fault">
1426 Indicates a connectivity fault triggered by an inability to receive
1427 heartbeats from any remote endpoint. When a fault is triggered on
1428 <ref table="Interface"/>s participating in bonds, they will be
1432 Faults can be triggered for several reasons. Most importantly they
1433 are triggered when no CCMs are received for a period of 3.5 times the
1434 transmission interval. Faults are also triggered when any CCMs
1435 indicate that a Remote Maintenance Point is not receiving CCMs but
1436 able to send them. Finally, a fault is triggered if a CCM is
1437 received which indicates unexpected configuration. Notably, this
1438 case arises when a CCM is received which advertises the local MPID.
1442 <column name="cfm_remote_mpids">
1443 When CFM is properly configured, Open vSwitch will occasionally
1444 receive CCM broadcasts. These broadcasts contain the MPID of the
1445 sending Maintenance Point. The list of MPIDs from which this
1446 <ref table="Interface"/> is receiving broadcasts from is regularly
1447 collected and written to this column.
1451 <group title="Other Features">
1453 <column name="lacp_current">
1454 Boolean value indicating LACP status for this interface. If true, this
1455 interface has current LACP information about its LACP partner. This
1456 information may be used to monitor the health of interfaces in a LACP
1457 enabled port. This column will be empty if LACP is not enabled.
1460 <column name="external_ids">
1461 Key-value pairs for use by external frameworks that integrate
1462 with Open vSwitch, rather than by Open vSwitch itself. System
1463 integrators should either use the Open vSwitch development
1464 mailing list to coordinate on common key-value definitions, or
1465 choose key names that are likely to be unique. The currently
1466 defined common key-value pairs are:
1468 <dt><code>attached-mac</code></dt>
1470 The MAC address programmed into the ``virtual hardware'' for this
1471 interface, in the form
1472 <var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>:<var>xx</var>.
1473 For Citrix XenServer, this is the value of the <code>MAC</code>
1474 field in the VIF record for this interface.</dd>
1475 <dt><code>iface-id</code></dt>
1476 <dd>A system-unique identifier for the interface. On XenServer,
1477 this will commonly be the same as
1478 <ref column="external_ids" key="xs-vif-uuid"/>.</dd>
1481 Additionally the following key-value pairs specifically
1482 apply to an interface that represents a virtual Ethernet interface
1483 connected to a virtual machine. These key-value pairs should not be
1484 present for other types of interfaces. Keys whose names end
1485 in <code>-uuid</code> have values that uniquely identify the entity
1486 in question. For a Citrix XenServer hypervisor, these values are
1487 UUIDs in RFC 4122 format. Other hypervisors may use other
1490 <p>The currently defined key-value pairs for XenServer are:</p>
1492 <dt><code>xs-vif-uuid</code></dt>
1493 <dd>The virtual interface associated with this interface.</dd>
1494 <dt><code>xs-network-uuid</code></dt>
1495 <dd>The virtual network to which this interface is attached.</dd>
1496 <dt><code>xs-vm-uuid</code></dt>
1497 <dd>The VM to which this interface belongs.</dd>
1501 <column name="other_config">
1502 Key-value pairs for rarely used interface features.
1504 <dt><code>cfm_interval</code></dt>
1505 <dd> The transmission interval of CFM heartbeats in milliseconds.
1506 Three missed heartbeat receptions indicate a connectivity fault.
1507 Defaults to 1000ms. </dd>
1508 <dt><code>cfm_extended</code></dt>
1509 <dd> When true, the CFM module operates in extended mode. This causes
1510 it to use a nonstandard destination address to avoid conflicting
1511 with compliant implementations which may be running concurrently on
1512 the network. Furthermore, extended mode increases the accuracy of
1513 the <code>cfm_interval</code> configuration parameter by breaking
1514 wire compatibility with 802.1ag compliant implementations.
1515 Defaults to false.</dd>
1516 <dt><code>bond-stable-id</code></dt>
1517 <dd> A positive integer using in <code>stable</code> bond mode to
1518 make slave selection decisions. Allocating
1519 <ref column="other_config" key="bond-stable-id"/> values
1520 consistently across interfaces participating in a bond will
1521 guarantee consistent slave selection decisions across
1522 <code>ovs-vswitchd</code> instances when using <code>stable</code>
1524 <dt><code>lacp-port-id</code></dt>
1525 <dd> The LACP port ID of this <ref table="Interface"/>. Port IDs are
1526 used in LACP negotiations to identify individual ports
1527 participating in a bond. Must be a number between 1 and
1529 <dt><code>lacp-port-priority</code></dt>
1530 <dd> The LACP port priority of this <ref table="Interface"/>. In
1531 LACP negotiations <ref table="Interface"/>s with numerically lower
1532 priorities are preferred for aggregation. Must be a number between
1534 <dt><code>lacp-aggregation-key</code></dt>
1535 <dd> The LACP aggregation key of this <ref table="Interface"/>.
1536 <ref table="Interface"/>s with different aggregation keys may not
1537 be active within a given <ref table="Port"/> at the same time. Must
1538 be a number between 1 and 65535.</dd>
1542 <column name="statistics">
1544 Key-value pairs that report interface statistics. The current
1545 implementation updates these counters periodically. In the future,
1546 we plan to, instead, update them when an interface is created, when
1547 they are queried (e.g. using an OVSDB <code>select</code> operation),
1548 and just before an interface is deleted due to virtual interface
1549 hot-unplug or VM shutdown, and perhaps at other times, but not on any
1550 regular periodic basis.</p>
1552 The currently defined key-value pairs are listed below. These are
1553 the same statistics reported by OpenFlow in its <code>struct
1554 ofp_port_stats</code> structure. If an interface does not support a
1555 given statistic, then that pair is omitted.</p>
1558 Successful transmit and receive counters:
1560 <dt><code>rx_packets</code></dt>
1561 <dd>Number of received packets.</dd>
1562 <dt><code>rx_bytes</code></dt>
1563 <dd>Number of received bytes.</dd>
1564 <dt><code>tx_packets</code></dt>
1565 <dd>Number of transmitted packets.</dd>
1566 <dt><code>tx_bytes</code></dt>
1567 <dd>Number of transmitted bytes.</dd>
1573 <dt><code>rx_dropped</code></dt>
1574 <dd>Number of packets dropped by RX.</dd>
1575 <dt><code>rx_frame_err</code></dt>
1576 <dd>Number of frame alignment errors.</dd>
1577 <dt><code>rx_over_err</code></dt>
1578 <dd>Number of packets with RX overrun.</dd>
1579 <dt><code>rx_crc_err</code></dt>
1580 <dd>Number of CRC errors.</dd>
1581 <dt><code>rx_errors</code></dt>
1583 Total number of receive errors, greater than or equal
1584 to the sum of the above.
1591 <dt><code>tx_dropped</code></dt>
1592 <dd>Number of packets dropped by TX.</dd>
1593 <dt><code>collisions</code></dt>
1594 <dd>Number of collisions.</dd>
1595 <dt><code>tx_errors</code></dt>
1597 Total number of transmit errors, greater
1598 than or equal to the sum of the above.
1607 <table name="QoS" title="Quality of Service configuration">
1608 <p>Quality of Service (QoS) configuration for each Port that
1611 <column name="type">
1612 <p>The type of QoS to implement. The <ref table="Open_vSwitch"
1613 column="capabilities"/> column in the <ref table="Open_vSwitch"/> table
1614 identifies the types that a switch actually supports. The currently
1615 defined types are listed below:</p>
1617 <dt><code>linux-htb</code></dt>
1619 Linux ``hierarchy token bucket'' classifier. See tc-htb(8) (also at
1620 <code>http://linux.die.net/man/8/tc-htb</code>) and the HTB manual
1621 (<code>http://luxik.cdi.cz/~devik/qos/htb/manual/userg.htm</code>)
1622 for information on how this classifier works and how to configure it.
1626 <dt><code>linux-hfsc</code></dt>
1628 Linux "Hierarchical Fair Service Curve" classifier.
1629 See <code>http://linux-ip.net/articles/hfsc.en/</code> for
1630 information on how this classifier works.
1635 <column name="queues">
1636 <p>A map from queue numbers to <ref table="Queue"/> records. The
1637 supported range of queue numbers depend on <ref column="type"/>. The
1638 queue numbers are the same as the <code>queue_id</code> used in
1639 OpenFlow in <code>struct ofp_action_enqueue</code> and other
1640 structures. Queue 0 is used by OpenFlow output actions that do not
1641 specify a specific queue.</p>
1644 <column name="other_config">
1645 <p>Key-value pairs for configuring QoS features that depend on
1646 <ref column="type"/>.</p>
1647 <p>The <code>linux-htb</code> and <code>linux-hfsc</code> classes support
1648 the following key-value pairs:</p>
1650 <dt><code>max-rate</code></dt>
1651 <dd>Maximum rate shared by all queued traffic, in bit/s.
1652 Optional. If not specified, for physical interfaces, the
1653 default is the link rate. For other interfaces or if the
1654 link rate cannot be determined, the default is currently 100
1659 <column name="external_ids">
1660 Key-value pairs for use by external frameworks that integrate with Open
1661 vSwitch, rather than by Open vSwitch itself. System integrators should
1662 either use the Open vSwitch development mailing list to coordinate on
1663 common key-value definitions, or choose key names that are likely to be
1664 unique. No common key-value pairs are currently defined.
1668 <table name="Queue" title="QoS output queue.">
1669 <p>A configuration for a port output queue, used in configuring Quality of
1670 Service (QoS) features. May be referenced by <ref column="queues"
1671 table="QoS"/> column in <ref table="QoS"/> table.</p>
1673 <column name="other_config">
1674 <p>Key-value pairs for configuring the output queue. The supported
1675 key-value pairs and their meanings depend on the <ref column="type"/>
1676 of the <ref column="QoS"/> records that reference this row.</p>
1677 <p>The key-value pairs defined for <ref table="QoS"/> <ref table="QoS"
1678 column="type"/> of <code>min-rate</code> are:</p>
1680 <dt><code>min-rate</code></dt>
1681 <dd>Minimum guaranteed bandwidth, in bit/s. Required. The
1682 floor value is 1500 bytes/s (12,000 bit/s).</dd>
1684 <p>The key-value pairs defined for <ref table="QoS"/> <ref table="QoS"
1685 column="type"/> of <code>linux-htb</code> are:</p>
1687 <dt><code>min-rate</code></dt>
1688 <dd>Minimum guaranteed bandwidth, in bit/s.</dd>
1689 <dt><code>max-rate</code></dt>
1690 <dd>Maximum allowed bandwidth, in bit/s. Optional. If specified, the
1691 queue's rate will not be allowed to exceed the specified value, even
1692 if excess bandwidth is available. If unspecified, defaults to no
1694 <dt><code>burst</code></dt>
1695 <dd>Burst size, in bits. This is the maximum amount of ``credits''
1696 that a queue can accumulate while it is idle. Optional. Details of
1697 the <code>linux-htb</code> implementation require a minimum burst
1698 size, so a too-small <code>burst</code> will be silently
1700 <dt><code>priority</code></dt>
1701 <dd>A nonnegative 32-bit integer. Defaults to 0 if
1702 unspecified. A queue with a smaller <code>priority</code>
1703 will receive all the excess bandwidth that it can use before
1704 a queue with a larger value receives any. Specific priority
1705 values are unimportant; only relative ordering matters.</dd>
1707 <p>The key-value pairs defined for <ref table="QoS"/> <ref table="QoS"
1708 column="type"/> of <code>linux-hfsc</code> are:</p>
1710 <dt><code>min-rate</code></dt>
1711 <dd>Minimum guaranteed bandwidth, in bit/s.</dd>
1712 <dt><code>max-rate</code></dt>
1713 <dd>Maximum allowed bandwidth, in bit/s. Optional. If specified, the
1714 queue's rate will not be allowed to exceed the specified value, even
1715 if excess bandwidth is available. If unspecified, defaults to no
1720 <column name="external_ids">
1721 Key-value pairs for use by external frameworks that integrate with Open
1722 vSwitch, rather than by Open vSwitch itself. System integrators should
1723 either use the Open vSwitch development mailing list to coordinate on
1724 common key-value definitions, or choose key names that are likely to be
1725 unique. No common key-value pairs are currently defined.
1729 <table name="Mirror" title="Port mirroring (SPAN/RSPAN/ERSPAN).">
1730 <p>A port mirror within a <ref table="Bridge"/>.</p>
1731 <p>A port mirror configures a bridge to send selected frames to special
1732 ``mirrored'' ports, in addition to their normal destinations. Mirroring
1733 traffic may also be referred to as SPAN, RSPAN, or ERSPAN, depending on how
1734 the mirrored traffic is sent.</p>
1736 <column name="name">
1737 Arbitrary identifier for the <ref table="Mirror"/>.
1740 <group title="Selecting Packets for Mirroring">
1742 To be selected for mirroring, a given packet must enter or leave the
1743 bridge through a selected port and it must also be in one of the
1747 <column name="select_all">
1748 If true, every packet arriving or departing on any port is
1749 selected for mirroring.
1752 <column name="select_dst_port">
1753 Ports on which departing packets are selected for mirroring.
1756 <column name="select_src_port">
1757 Ports on which arriving packets are selected for mirroring.
1760 <column name="select_vlan">
1761 VLANs on which packets are selected for mirroring. An empty set
1762 selects packets on all VLANs.
1766 <group title="Mirroring Destination Configuration">
1768 These columns are mutually exclusive. Exactly one of them must be
1772 <column name="output_port">
1773 <p>Output port for selected packets, if nonempty.</p>
1774 <p>Specifying a port for mirror output reserves that port exclusively
1775 for mirroring. No frames other than those selected for mirroring
1776 will be forwarded to the port, and any frames received on the port
1777 will be discarded.</p>
1779 The output port may be any kind of port supported by Open vSwitch.
1780 It may be, for example, a physical port (sometimes called SPAN), or a
1781 GRE tunnel (sometimes called ERSPAN).
1785 <column name="output_vlan">
1786 <p>Output VLAN for selected packets, if nonempty.</p>
1787 <p>The frames will be sent out all ports that trunk
1788 <ref column="output_vlan"/>, as well as any ports with implicit VLAN
1789 <ref column="output_vlan"/>. When a mirrored frame is sent out a
1790 trunk port, the frame's VLAN tag will be set to
1791 <ref column="output_vlan"/>, replacing any existing tag; when it is
1792 sent out an implicit VLAN port, the frame will not be tagged. This
1793 type of mirroring is sometimes called RSPAN.</p>
1795 The following destination MAC addresses will not be mirrored to a
1796 VLAN to avoid confusing switches that interpret the protocols that
1800 <dt><code>01:80:c2:00:00:00</code></dt>
1801 <dd>IEEE 802.1D Spanning Tree Protocol (STP).</dd>
1803 <dt><code>01:80:c2:00:00:01</code></dt>
1804 <dd>IEEE Pause frame.</dd>
1806 <dt><code>01:80:c2:00:00:0<var>x</var></code></dt>
1807 <dd>Other reserved protocols.</dd>
1809 <dt><code>01:00:0c:cc:cc:cc</code></dt>
1811 Cisco Discovery Protocol (CDP), VLAN Trunking Protocol (VTP),
1812 Dynamic Trunking Protocol (DTP), Port Aggregation Protocol (PAgP),
1816 <dt><code>01:00:0c:cc:cc:cd</code></dt>
1817 <dd>Cisco Shared Spanning Tree Protocol PVSTP+.</dd>
1819 <dt><code>01:00:0c:cd:cd:cd</code></dt>
1820 <dd>Cisco STP Uplink Fast.</dd>
1822 <dt><code>01:00:0c:00:00:00</code></dt>
1823 <dd>Cisco Inter Switch Link.</dd>
1825 <p><em>Please note:</em> Mirroring to a VLAN can disrupt a network that
1826 contains unmanaged switches. Consider an unmanaged physical switch
1827 with two ports: port 1, connected to an end host, and port 2,
1828 connected to an Open vSwitch configured to mirror received packets
1829 into VLAN 123 on port 2. Suppose that the end host sends a packet on
1830 port 1 that the physical switch forwards to port 2. The Open vSwitch
1831 forwards this packet to its destination and then reflects it back on
1832 port 2 in VLAN 123. This reflected packet causes the unmanaged
1833 physical switch to replace the MAC learning table entry, which
1834 correctly pointed to port 1, with one that incorrectly points to port
1835 2. Afterward, the physical switch will direct packets destined for
1836 the end host to the Open vSwitch on port 2, instead of to the end
1837 host on port 1, disrupting connectivity. If mirroring to a VLAN is
1838 desired in this scenario, then the physical switch must be replaced
1839 by one that learns Ethernet addresses on a per-VLAN basis. In
1840 addition, learning should be disabled on the VLAN containing mirrored
1841 traffic. If this is not done then intermediate switches will learn
1842 the MAC address of each end host from the mirrored traffic. If
1843 packets being sent to that end host are also mirrored, then they will
1844 be dropped since the switch will attempt to send them out the input
1845 port. Disabling learning for the VLAN will cause the switch to
1846 correctly send the packet out all ports configured for that VLAN. If
1847 Open vSwitch is being used as an intermediate switch, learning can be
1848 disabled by adding the mirrored VLAN to <ref column="flood_vlans"/>
1849 in the appropriate <ref table="Bridge"/> table or tables.</p>
1851 Mirroring to a GRE tunnel has fewer caveats than mirroring to a
1852 VLAN and should generally be preferred.
1857 <group title="Other Features">
1858 <column name="external_ids">
1859 Key-value pairs for use by external frameworks that integrate with Open
1860 vSwitch, rather than by Open vSwitch itself. System integrators should
1861 either use the Open vSwitch development mailing list to coordinate on
1862 common key-value definitions, or choose key names that are likely to be
1863 unique. No common key-value pairs are currently defined.
1868 <table name="Controller" title="OpenFlow controller configuration.">
1869 <p>An OpenFlow controller.</p>
1872 Open vSwitch supports two kinds of OpenFlow controllers:
1876 <dt>Primary controllers</dt>
1879 This is the kind of controller envisioned by the OpenFlow 1.0
1880 specification. Usually, a primary controller implements a network
1881 policy by taking charge of the switch's flow table.
1885 Open vSwitch initiates and maintains persistent connections to
1886 primary controllers, retrying the connection each time it fails or
1887 drops. The <ref table="Bridge" column="fail_mode"/> column in the
1888 <ref table="Bridge"/> table applies to primary controllers.
1892 Open vSwitch permits a bridge to have any number of primary
1893 controllers. When multiple controllers are configured, Open
1894 vSwitch connects to all of them simultaneously. Because
1895 OpenFlow 1.0 does not specify how multiple controllers
1896 coordinate in interacting with a single switch, more than
1897 one primary controller should be specified only if the
1898 controllers are themselves designed to coordinate with each
1899 other. (The Nicira-defined <code>NXT_ROLE</code> OpenFlow
1900 vendor extension may be useful for this.)
1903 <dt>Service controllers</dt>
1906 These kinds of OpenFlow controller connections are intended for
1907 occasional support and maintenance use, e.g. with
1908 <code>ovs-ofctl</code>. Usually a service controller connects only
1909 briefly to inspect or modify some of a switch's state.
1913 Open vSwitch listens for incoming connections from service
1914 controllers. The service controllers initiate and, if necessary,
1915 maintain the connections from their end. The <ref table="Bridge"
1916 column="fail_mode"/> column in the <ref table="Bridge"/> table does
1917 not apply to service controllers.
1921 Open vSwitch supports configuring any number of service controllers.
1927 The <ref column="target"/> determines the type of controller.
1930 <group title="Core Features">
1931 <column name="target">
1932 <p>Connection method for controller.</p>
1934 The following connection methods are currently supported for primary
1938 <dt><code>ssl:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
1940 <p>The specified SSL <var>port</var> (default: 6633) on the host at
1941 the given <var>ip</var>, which must be expressed as an IP address
1942 (not a DNS name). The <ref table="Open_vSwitch" column="ssl"/>
1943 column in the <ref table="Open_vSwitch"/> table must point to a
1944 valid SSL configuration when this form is used.</p>
1945 <p>SSL support is an optional feature that is not always built as
1946 part of Open vSwitch.</p>
1948 <dt><code>tcp:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
1949 <dd>The specified TCP <var>port</var> (default: 6633) on the host at
1950 the given <var>ip</var>, which must be expressed as an IP address
1951 (not a DNS name).</dd>
1954 The following connection methods are currently supported for service
1958 <dt><code>pssl:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
1961 Listens for SSL connections on the specified TCP <var>port</var>
1962 (default: 6633). If <var>ip</var>, which must be expressed as an
1963 IP address (not a DNS name), is specified, then connections are
1964 restricted to the specified local IP address.
1967 The <ref table="Open_vSwitch" column="ssl"/> column in the <ref
1968 table="Open_vSwitch"/> table must point to a valid SSL
1969 configuration when this form is used.
1971 <p>SSL support is an optional feature that is not always built as
1972 part of Open vSwitch.</p>
1974 <dt><code>ptcp:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
1976 Listens for connections on the specified TCP <var>port</var>
1977 (default: 6633). If <var>ip</var>, which must be expressed as an
1978 IP address (not a DNS name), is specified, then connections are
1979 restricted to the specified local IP address.
1982 <p>When multiple controllers are configured for a single bridge, the
1983 <ref column="target"/> values must be unique. Duplicate
1984 <ref column="target"/> values yield unspecified results.</p>
1987 <column name="connection_mode">
1988 <p>If it is specified, this setting must be one of the following
1989 strings that describes how Open vSwitch contacts this OpenFlow
1990 controller over the network:</p>
1993 <dt><code>in-band</code></dt>
1994 <dd>In this mode, this controller's OpenFlow traffic travels over the
1995 bridge associated with the controller. With this setting, Open
1996 vSwitch allows traffic to and from the controller regardless of the
1997 contents of the OpenFlow flow table. (Otherwise, Open vSwitch
1998 would never be able to connect to the controller, because it did
1999 not have a flow to enable it.) This is the most common connection
2000 mode because it is not necessary to maintain two independent
2002 <dt><code>out-of-band</code></dt>
2003 <dd>In this mode, OpenFlow traffic uses a control network separate
2004 from the bridge associated with this controller, that is, the
2005 bridge does not use any of its own network devices to communicate
2006 with the controller. The control network must be configured
2007 separately, before or after <code>ovs-vswitchd</code> is started.
2011 <p>If not specified, the default is implementation-specific.</p>
2015 <group title="Controller Failure Detection and Handling">
2016 <column name="max_backoff">
2017 Maximum number of milliseconds to wait between connection attempts.
2018 Default is implementation-specific.
2021 <column name="inactivity_probe">
2022 Maximum number of milliseconds of idle time on connection to
2023 controller before sending an inactivity probe message. If Open
2024 vSwitch does not communicate with the controller for the specified
2025 number of seconds, it will send a probe. If a response is not
2026 received for the same additional amount of time, Open vSwitch
2027 assumes the connection has been broken and attempts to reconnect.
2028 Default is implementation-specific. A value of 0 disables
2033 <group title="OpenFlow Rate Limiting">
2034 <column name="controller_rate_limit">
2035 <p>The maximum rate at which packets in unknown flows will be
2036 forwarded to the OpenFlow controller, in packets per second. This
2037 feature prevents a single bridge from overwhelming the controller.
2038 If not specified, the default is implementation-specific.</p>
2039 <p>In addition, when a high rate triggers rate-limiting, Open
2040 vSwitch queues controller packets for each port and transmits
2041 them to the controller at the configured rate. The number of
2042 queued packets is limited by
2043 the <ref column="controller_burst_limit"/> value. The packet
2044 queue is shared fairly among the ports on a bridge.</p><p>Open
2045 vSwitch maintains two such packet rate-limiters per bridge.
2046 One of these applies to packets sent up to the controller
2047 because they do not correspond to any flow. The other applies
2048 to packets sent up to the controller by request through flow
2049 actions. When both rate-limiters are filled with packets, the
2050 actual rate that packets are sent to the controller is up to
2051 twice the specified rate.</p>
2054 <column name="controller_burst_limit">
2055 In conjunction with <ref column="controller_rate_limit"/>,
2056 the maximum number of unused packet credits that the bridge will
2057 allow to accumulate, in packets. If not specified, the default
2058 is implementation-specific.
2062 <group title="Additional In-Band Configuration">
2063 <p>These values are considered only in in-band control mode (see
2064 <ref column="connection_mode"/>).</p>
2066 <p>When multiple controllers are configured on a single bridge, there
2067 should be only one set of unique values in these columns. If different
2068 values are set for these columns in different controllers, the effect
2071 <column name="local_ip">
2072 The IP address to configure on the local port,
2073 e.g. <code>192.168.0.123</code>. If this value is unset, then
2074 <ref column="local_netmask"/> and <ref column="local_gateway"/> are
2078 <column name="local_netmask">
2079 The IP netmask to configure on the local port,
2080 e.g. <code>255.255.255.0</code>. If <ref column="local_ip"/> is set
2081 but this value is unset, then the default is chosen based on whether
2082 the IP address is class A, B, or C.
2085 <column name="local_gateway">
2086 The IP address of the gateway to configure on the local port, as a
2087 string, e.g. <code>192.168.0.1</code>. Leave this column unset if
2088 this network has no gateway.
2092 <group title="Other Features">
2093 <column name="external_ids">
2094 Key-value pairs for use by external frameworks that integrate with Open
2095 vSwitch, rather than by Open vSwitch itself. System integrators should
2096 either use the Open vSwitch development mailing list to coordinate on
2097 common key-value definitions, or choose key names that are likely to be
2098 unique. No common key-value pairs are currently defined.
2102 <group title="Controller Status">
2103 <column name="is_connected">
2104 <code>true</code> if currently connected to this controller,
2105 <code>false</code> otherwise.
2108 <column name="role">
2109 <p>The level of authority this controller has on the associated
2110 bridge. Possible values are:</p>
2112 <dt><code>other</code></dt>
2113 <dd>Allows the controller access to all OpenFlow features.</dd>
2114 <dt><code>master</code></dt>
2115 <dd>Equivalent to <code>other</code>, except that there may be at
2116 most one master controller at a time. When a controller configures
2117 itself as <code>master</code>, any existing master is demoted to
2118 the <code>slave</code>role.</dd>
2119 <dt><code>slave</code></dt>
2120 <dd>Allows the controller read-only access to OpenFlow features.
2121 Attempts to modify the flow table will be rejected with an
2122 error. Slave controllers do not receive OFPT_PACKET_IN or
2123 OFPT_FLOW_REMOVED messages, but they do receive OFPT_PORT_STATUS
2128 <column name="status">
2129 <p>Key-value pairs that report controller status.</p>
2131 <dt><code>last_error</code></dt>
2132 <dd>A human-readable description of the last error on the connection
2133 to the controller; i.e. <code>strerror(errno)</code>. This key
2134 will exist only if an error has occurred.</dd>
2135 <dt><code>state</code></dt>
2136 <dd>The state of the connection to the controller. Possible values
2137 are: <code>VOID</code> (connection is disabled),
2138 <code>BACKOFF</code> (attempting to reconnect at an increasing
2139 period), <code>CONNECTING</code> (attempting to connect),
2140 <code>ACTIVE</code> (connected, remote host responsive), and
2141 <code>IDLE</code> (remote host idle, sending keep-alive). These
2142 values may change in the future. They are provided only for human
2144 <dt><code>sec_since_connect</code></dt>
2145 <dd>The amount of time since this controller last successfully
2146 connected to the switch (in seconds). Value is empty if controller
2147 has never successfully connected.</dd>
2148 <dt><code>sec_since_disconnect</code></dt>
2149 <dd>The amount of time since this controller last disconnected from
2150 the switch (in seconds). Value is empty if controller has never
2157 <table name="Manager" title="OVSDB management connection.">
2159 Configuration for a database connection to an Open vSwitch database
2164 This table primarily configures the Open vSwitch database
2165 (<code>ovsdb-server</code>), not the Open vSwitch switch
2166 (<code>ovs-vswitchd</code>). The switch does read the table to determine
2167 what connections should be treated as in-band.
2171 The Open vSwitch database server can initiate and maintain active
2172 connections to remote clients. It can also listen for database
2176 <group title="Core Features">
2177 <column name="target">
2178 <p>Connection method for managers.</p>
2180 The following connection methods are currently supported:
2183 <dt><code>ssl:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
2186 The specified SSL <var>port</var> (default: 6632) on the host at
2187 the given <var>ip</var>, which must be expressed as an IP address
2188 (not a DNS name). The <ref table="Open_vSwitch" column="ssl"/>
2189 column in the <ref table="Open_vSwitch"/> table must point to a
2190 valid SSL configuration when this form is used.
2193 SSL support is an optional feature that is not always built as
2194 part of Open vSwitch.
2198 <dt><code>tcp:<var>ip</var></code>[<code>:<var>port</var></code>]</dt>
2200 The specified TCP <var>port</var> (default: 6632) on the host at
2201 the given <var>ip</var>, which must be expressed as an IP address
2204 <dt><code>pssl:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
2207 Listens for SSL connections on the specified TCP <var>port</var>
2208 (default: 6632). If <var>ip</var>, which must be expressed as an
2209 IP address (not a DNS name), is specified, then connections are
2210 restricted to the specified local IP address.
2213 The <ref table="Open_vSwitch" column="ssl"/> column in the <ref
2214 table="Open_vSwitch"/> table must point to a valid SSL
2215 configuration when this form is used.
2218 SSL support is an optional feature that is not always built as
2219 part of Open vSwitch.
2222 <dt><code>ptcp:</code>[<var>port</var>][<code>:<var>ip</var></code>]</dt>
2224 Listens for connections on the specified TCP <var>port</var>
2225 (default: 6632). If <var>ip</var>, which must be expressed as an
2226 IP address (not a DNS name), is specified, then connections are
2227 restricted to the specified local IP address.
2230 <p>When multiple managers are configured, the <ref column="target"/>
2231 values must be unique. Duplicate <ref column="target"/> values yield
2232 unspecified results.</p>
2235 <column name="connection_mode">
2237 If it is specified, this setting must be one of the following strings
2238 that describes how Open vSwitch contacts this OVSDB client over the
2243 <dt><code>in-band</code></dt>
2245 In this mode, this connection's traffic travels over a bridge
2246 managed by Open vSwitch. With this setting, Open vSwitch allows
2247 traffic to and from the client regardless of the contents of the
2248 OpenFlow flow table. (Otherwise, Open vSwitch would never be able
2249 to connect to the client, because it did not have a flow to enable
2250 it.) This is the most common connection mode because it is not
2251 necessary to maintain two independent networks.
2253 <dt><code>out-of-band</code></dt>
2255 In this mode, the client's traffic uses a control network separate
2256 from that managed by Open vSwitch, that is, Open vSwitch does not
2257 use any of its own network devices to communicate with the client.
2258 The control network must be configured separately, before or after
2259 <code>ovs-vswitchd</code> is started.
2264 If not specified, the default is implementation-specific.
2269 <group title="Client Failure Detection and Handling">
2270 <column name="max_backoff">
2271 Maximum number of milliseconds to wait between connection attempts.
2272 Default is implementation-specific.
2275 <column name="inactivity_probe">
2276 Maximum number of milliseconds of idle time on connection to the client
2277 before sending an inactivity probe message. If Open vSwitch does not
2278 communicate with the client for the specified number of seconds, it
2279 will send a probe. If a response is not received for the same
2280 additional amount of time, Open vSwitch assumes the connection has been
2281 broken and attempts to reconnect. Default is implementation-specific.
2282 A value of 0 disables inactivity probes.
2286 <group title="Other Features">
2287 <column name="external_ids">
2288 Key-value pairs for use by external frameworks that integrate with Open
2289 vSwitch, rather than by Open vSwitch itself. System integrators should
2290 either use the Open vSwitch development mailing list to coordinate on
2291 common key-value definitions, or choose key names that are likely to be
2292 unique. No common key-value pairs are currently defined.
2296 <group title="Status">
2297 <column name="is_connected">
2298 <code>true</code> if currently connected to this manager,
2299 <code>false</code> otherwise.
2302 <column name="status">
2303 <p>Key-value pairs that report manager status.</p>
2305 <dt><code>last_error</code></dt>
2306 <dd>A human-readable description of the last error on the connection
2307 to the manager; i.e. <code>strerror(errno)</code>. This key
2308 will exist only if an error has occurred.</dd>
2311 <dt><code>state</code></dt>
2312 <dd>The state of the connection to the manager. Possible values
2313 are: <code>VOID</code> (connection is disabled),
2314 <code>BACKOFF</code> (attempting to reconnect at an increasing
2315 period), <code>CONNECTING</code> (attempting to connect),
2316 <code>ACTIVE</code> (connected, remote host responsive), and
2317 <code>IDLE</code> (remote host idle, sending keep-alive). These
2318 values may change in the future. They are provided only for human
2322 <dt><code>sec_since_connect</code></dt>
2323 <dd>The amount of time since this manager last successfully connected
2324 to the database (in seconds). Value is empty if manager has never
2325 successfully connected.</dd>
2328 <dt><code>sec_since_disconnect</code></dt>
2329 <dd>The amount of time since this manager last disconnected from the
2330 database (in seconds). Value is empty if manager has never
2334 <dt><code>locks_held</code></dt>
2335 <dt><code>locks_waiting</code></dt>
2336 <dt><code>locks_lost</code></dt>
2338 Space-separated lists of the names of OVSDB locks that the
2339 connection holds, is currently waiting to acquire, or has had
2340 stolen by another OVSDB client, respectively. Key-value pairs for
2341 lists that would be empty are omitted.
2345 <dt><code>n_connections</code></dt>
2348 When <ref column="target"/> specifies a connection method that
2349 listens for inbound connections (e.g. <code>ptcp:</code> or
2350 <code>pssl:</code>) and more than one connection is actually
2351 active, the value is the number of active connections.
2352 Otherwise, this key-value pair is omitted.
2355 When multiple connections are active, status columns and
2356 key-value pairs (other than this one) report the status of one
2357 arbitrarily chosen connection.
2365 <table name="NetFlow">
2366 A NetFlow target. NetFlow is a protocol that exports a number of
2367 details about terminating IP flows, such as the principals involved
2370 <column name="targets">
2371 NetFlow targets in the form
2372 <code><var>ip</var>:<var>port</var></code>. The <var>ip</var>
2373 must be specified numerically, not as a DNS name.
2376 <column name="engine_id">
2377 Engine ID to use in NetFlow messages. Defaults to datapath index
2381 <column name="engine_type">
2382 Engine type to use in NetFlow messages. Defaults to datapath
2383 index if not specified.
2386 <column name="active_timeout">
2387 The interval at which NetFlow records are sent for flows that are
2388 still active, in seconds. A value of <code>0</code> requests the
2389 default timeout (currently 600 seconds); a value of <code>-1</code>
2390 disables active timeouts.
2393 <column name="add_id_to_interface">
2394 <p>If this column's value is <code>false</code>, the ingress and egress
2395 interface fields of NetFlow flow records are derived from OpenFlow port
2396 numbers. When it is <code>true</code>, the 7 most significant bits of
2397 these fields will be replaced by the least significant 7 bits of the
2398 engine id. This is useful because many NetFlow collectors do not
2399 expect multiple switches to be sending messages from the same host, so
2400 they do not store the engine information which could be used to
2401 disambiguate the traffic.</p>
2402 <p>When this option is enabled, a maximum of 508 ports are supported.</p>
2405 <column name="external_ids">
2406 Key-value pairs for use by external frameworks that integrate with Open
2407 vSwitch, rather than by Open vSwitch itself. System integrators should
2408 either use the Open vSwitch development mailing list to coordinate on
2409 common key-value definitions, or choose key names that are likely to be
2410 unique. No common key-value pairs are currently defined.
2415 SSL configuration for an Open_vSwitch.
2417 <column name="private_key">
2418 Name of a PEM file containing the private key used as the switch's
2419 identity for SSL connections to the controller.
2422 <column name="certificate">
2423 Name of a PEM file containing a certificate, signed by the
2424 certificate authority (CA) used by the controller and manager,
2425 that certifies the switch's private key, identifying a trustworthy
2429 <column name="ca_cert">
2430 Name of a PEM file containing the CA certificate used to verify
2431 that the switch is connected to a trustworthy controller.
2434 <column name="bootstrap_ca_cert">
2435 If set to <code>true</code>, then Open vSwitch will attempt to
2436 obtain the CA certificate from the controller on its first SSL
2437 connection and save it to the named PEM file. If it is successful,
2438 it will immediately drop the connection and reconnect, and from then
2439 on all SSL connections must be authenticated by a certificate signed
2440 by the CA certificate thus obtained. <em>This option exposes the
2441 SSL connection to a man-in-the-middle attack obtaining the initial
2442 CA certificate.</em> It may still be useful for bootstrapping.
2445 <column name="external_ids">
2446 Key-value pairs for use by external frameworks that integrate with Open
2447 vSwitch, rather than by Open vSwitch itself. System integrators should
2448 either use the Open vSwitch development mailing list to coordinate on
2449 common key-value definitions, or choose key names that are likely to be
2450 unique. No common key-value pairs are currently defined.
2454 <table name="sFlow">
2455 <p>An sFlow(R) target. sFlow is a protocol for remote monitoring
2458 <column name="agent">
2459 Name of the network device whose IP address should be reported as the
2460 ``agent address'' to collectors. If not specified, the IP address
2461 defaults to the <ref table="Controller" column="local_ip"/> in the
2462 collector's <ref table="Controller"/>. If an agent IP address cannot be
2463 determined either way, sFlow is disabled.
2466 <column name="header">
2467 Number of bytes of a sampled packet to send to the collector.
2468 If not specified, the default is 128 bytes.
2471 <column name="polling">
2472 Polling rate in seconds to send port statistics to the collector.
2473 If not specified, defaults to 30 seconds.
2476 <column name="sampling">
2477 Rate at which packets should be sampled and sent to the collector.
2478 If not specified, defaults to 400, which means one out of 400
2479 packets, on average, will be sent to the collector.
2482 <column name="targets">
2483 sFlow targets in the form
2484 <code><var>ip</var>:<var>port</var></code>.
2487 <column name="external_ids">
2488 Key-value pairs for use by external frameworks that integrate with Open
2489 vSwitch, rather than by Open vSwitch itself. System integrators should
2490 either use the Open vSwitch development mailing list to coordinate on
2491 common key-value definitions, or choose key names that are likely to be
2492 unique. No common key-value pairs are currently defined.
2496 <table name="Capability">
2497 <p>Records in this table describe functionality supported by the hardware
2498 and software platform on which this Open vSwitch is based. Clients
2499 should not modify this table.</p>
2501 <p>A record in this table is meaningful only if it is referenced by the
2502 <ref table="Open_vSwitch" column="capabilities"/> column in the
2503 <ref table="Open_vSwitch"/> table. The key used to reference it, called
2504 the record's ``category,'' determines the meanings of the
2505 <ref column="details"/> column. The following general forms of
2506 categories are currently defined:</p>
2509 <dt><code>qos-<var>type</var></code></dt>
2510 <dd><var>type</var> is supported as the value for
2511 <ref column="type" table="QoS"/> in the <ref table="QoS"/> table.
2515 <column name="details">
2516 <p>Key-value pairs that describe capabilities. The meaning of the pairs
2517 depends on the category key that the <ref table="Open_vSwitch"
2518 column="capabilities"/> column in the <ref table="Open_vSwitch"/> table
2519 uses to reference this record, as described above.</p>
2521 <p>The presence of a record for category <code>qos-<var>type</var></code>
2522 indicates that the switch supports <var>type</var> as the value of
2523 the <ref table="QoS" column="type"/> column in the <ref table="QoS"/>
2524 table. The following key-value pairs are defined to further describe
2525 QoS capabilities:</p>
2528 <dt><code>n-queues</code></dt>
2529 <dd>Number of supported queues, as a positive integer. Keys in the
2530 <ref table="QoS" column="queues"/> column for <ref table="QoS"/>
2531 records whose <ref table="QoS" column="type"/> value
2532 equals <var>type</var> must range between 0 and this value minus one,