vswitchd/ovs-vswitchd.8.in

   1 .\" -*- nroff -*-
   2 .de IQ
   3 .  br
   4 .  ns
   5 .  IP "\\$1"
   6 ..
   7 .TH ovs\-vswitchd 8 "@VERSION@" "Open vSwitch" "Open vSwitch Manual"
   8 .\" This program's name:
   9 .ds PN ovs\-vswitchd
  10 .
  11 .SH NAME
  12 ovs\-vswitchd \- Open vSwitch daemon
  13 .
  14 .SH SYNOPSIS
  15 \fBovs\-vswitchd \fR[\fIdatabase\fR]
  16 .
  17 .SH DESCRIPTION
  18 A daemon that manages and controls any number of Open vSwitch switches
  19 on the local machine.
  20 .PP
  21 The \fIdatabase\fR argument specifies how \fBovs\-vswitchd\fR connects
  22 to \fBovsdb\-server\fR.  \fIdatabase\fR may be an OVSDB active or
  23 passive connection method, as described in \fBovsdb\fR(7).  The
  24 default is \fBunix:@RUNDIR@/db.sock\fR.
  25 .PP
  26 \fBovs\-vswitchd\fR retrieves its configuration from \fIdatabase\fR at
  27 startup.  It sets up Open vSwitch datapaths and then operates
  28 switching across each bridge described in its configuration files.  As
  29 the database changes, \fBovs\-vswitchd\fR automatically updates its
  30 configuration to match.
  31 .PP
  32 \fBovs\-vswitchd\fR switches may be configured with any of the following
  33 features:
  34 .
  35 .IP \(bu
  36 L2 switching with MAC learning.
  37 .
  38 .IP \(bu
  39 NIC bonding with automatic fail-over and source MAC-based TX load
  40 balancing ("SLB").
  41 .
  42 .IP \(bu
  43 802.1Q VLAN support.
  44 .
  45 .IP \(bu
  46 Port mirroring, with optional VLAN tagging.
  47 .
  48 .IP \(bu
  49 NetFlow v5 flow logging.
  50 .
  51 .IP \(bu
  52 sFlow(R) monitoring.
  53 .
  54 .IP \(bu
  55 Connectivity to an external OpenFlow controller, such as NOX.
  56 .
  57 .PP
  58 Only a single instance of \fBovs\-vswitchd\fR is intended to run at a time.
  59 A single \fBovs\-vswitchd\fR can manage any number of switch instances, up
  60 to the maximum number of supported Open vSwitch datapaths.
  61 .PP
  62 \fBovs\-vswitchd\fR does all the necessary management of Open vSwitch datapaths
  63 itself.  Thus, external tools, such \fBovs\-dpctl\fR(8), are not needed for
  64 managing datapaths in conjunction with \fBovs\-vswitchd\fR, and their use
  65 to modify datapaths when \fBovs\-vswitchd\fR is running can interfere with
  66 its operation.  (\fBovs\-dpctl\fR may still be useful for diagnostics.)
  67 .PP
  68 An Open vSwitch datapath kernel module must be loaded for \fBovs\-vswitchd\fR
  69 to be useful.  Refer to the documentation for instructions on how to build and
  70 load the Open vSwitch kernel module.
  71 .PP
  72 .SH OPTIONS
  73 .IP "\fB\-\-mlockall\fR"
  74 Causes \fBovs\-vswitchd\fR to call the \fBmlockall()\fR function, to
  75 attempt to lock all of its process memory into physical RAM,
  76 preventing the kernel from paging any of its memory to disk.  This
  77 helps to avoid networking interruptions due to system memory pressure.
  78 .IP
  79 Some systems do not support \fBmlockall()\fR at all, and other systems
  80 only allow privileged users, such as the superuser, to use it.
  81 \fBovs\-vswitchd\fR emits a log message if \fBmlockall()\fR is
  82 unavailable or unsuccessful.
  83 .
  84 .SS "DPDK Options"
  85 For details on initializing the \fBovs\-vswitchd\fR DPDK datapath,
  86 refer to the documentation or \fBovs\-vswitchd.conf.db\fR(5) for
  87 details.
  88 .SS "Daemon Options"
  89 .ds DD \
  90 \fBovs\-vswitchd\fR detaches only after it has connected to the \
  91 database, retrieved the initial configuration, and set up that \
  92 configuration.
  93 .so lib/daemon.man
  94 .SS "Service Options"
  95 .so lib/service.man
  96 .SS "Public Key Infrastructure Options"
  97 .so lib/ssl.man
  98 .so lib/ssl-bootstrap.man
  99 .SS "Logging Options"
 100 .so lib/vlog.man
 101 .SS "Other Options"
 102 .so lib/unixctl.man
 103 .so lib/common.man
 104 .
 105 .SH "RUNTIME MANAGEMENT COMMANDS"
 106 \fBovs\-appctl\fR(8) can send commands to a running
 107 \fBovs\-vswitchd\fR process.  The currently supported commands are
 108 described below.  The command descriptions assume an understanding of
 109 how to configure Open vSwitch.
 110 .SS "GENERAL COMMANDS"
 111 .IP "\fBexit\fR \fI--cleanup\fR"
 112 Causes \fBovs\-vswitchd\fR to gracefully terminate. If \fI--cleanup\fR
 113 is specified, release datapath resources configured by \fBovs\-vswitchd\fR.
 114 Otherwise, datapath flows and other resources remains undeleted.
 115 .
 116 .IP "\fBqos/show-types\fR \fIinterface\fR"
 117 Queries the interface for a list of Quality of Service types that are
 118 configurable via Open vSwitch for the given \fIinterface\fR.
 119 .IP "\fBqos/show\fR \fIinterface\fR"
 120 Queries the kernel for Quality of Service configuration and statistics
 121 associated with the given \fIinterface\fR.
 122 .IP "\fBbfd/show\fR [\fIinterface\fR]"
 123 Displays detailed information about Bidirectional Forwarding Detection
 124 configured on \fIinterface\fR.  If \fIinterface\fR is not specified,
 125 then displays detailed information about all interfaces with BFD
 126 enabled.
 127 .IP "\fBbfd/set-forwarding\fR [\fIinterface\fR] \fIstatus\fR"
 128 Force the fault status of the BFD module on \fIinterface\fR (or all
 129 interfaces if none is given) to be \fIstatus\fR.  \fIstatus\fR can be
 130 "true", "false", or "normal" which reverts to the standard behavior.
 131 .IP "\fBcfm/show\fR [\fIinterface\fR]"
 132 Displays detailed information about Connectivity Fault Management
 133 configured on \fIinterface\fR.  If \fIinterface\fR is not specified,
 134 then displays detailed information about all interfaces with CFM
 135 enabled.
 136 .IP "\fBcfm/set-fault\fR [\fIinterface\fR] \fIstatus\fR"
 137 Force the fault status of the CFM module on \fIinterface\fR (or all
 138 interfaces if none is given) to be \fIstatus\fR.  \fIstatus\fR can be
 139 "true", "false", or "normal" which reverts to the standard behavior.
 140 .IP "\fBstp/tcn\fR [\fIbridge\fR]"
 141 Forces a topology change event on \fIbridge\fR if it's running STP.  This
 142 may cause it to send Topology Change Notifications to its peers and flush
 143 its MAC table.  If no \fIbridge\fR is given, forces a topology change
 144 event on all bridges.
 145 .IP "\fBstp/show\fR [\fIbridge\fR]"
 146 Displays detailed information about spanning tree on the \fIbridge\fR.  If
 147 \fIbridge\fR is not specified, then displays detailed information about all
 148 bridges with STP enabled.
 149 .IP "\fBrstp/tcn\fR [\fIbridge\fR]"
 150 Forces a topology change event on \fIbridge\fR if it's running RSTP.  This
 151 may cause it to send Topology Change Notifications to its peers and flush
 152 its MAC table.  If no \fIbridge\fR is given, forces a topology change
 153 event on all bridges.
 154 .IP "\fBrstp/show\fR [\fIbridge\fR]"
 155 Displays detailed information about rapid spanning tree on the \fIbridge\fR.
 156 If \fIbridge\fR is not specified, then displays detailed information about all
 157 bridges with RSTP enabled.
 158 .SS "BRIDGE COMMANDS"
 159 These commands manage bridges.
 160 .IP "\fBfdb/flush\fR [\fIbridge\fR]"
 161 Flushes \fIbridge\fR MAC address learning table, or all learning tables
 162 if no \fIbridge\fR is given.
 163 .IP "\fBfdb/show\fR \fIbridge\fR"
 164 Lists each MAC address/VLAN pair learned by the specified \fIbridge\fR,
 165 along with the port on which it was learned and the age of the entry,
 166 in seconds.
 167 .IP "\fBmdb/flush\fR [\fIbridge\fR]"
 168 Flushes \fIbridge\fR multicast snooping table, or all snooping tables
 169 if no \fIbridge\fR is given.
 170 .IP "\fBmdb/show\fR \fIbridge\fR"
 171 Lists each multicast group/VLAN pair learned by the specified \fIbridge\fR,
 172 along with the port on which it was learned and the age of the entry,
 173 in seconds.
 174 .IP "\fBbridge/reconnect\fR [\fIbridge\fR]"
 175 Makes \fIbridge\fR drop all of its OpenFlow controller connections and
 176 reconnect.  If \fIbridge\fR is not specified, then all bridges drop
 177 their controller connections and reconnect.
 178 .IP
 179 This command might be useful for debugging OpenFlow controller issues.
 180 .
 181 .IP "\fBbridge/dump\-flows\fR \fIbridge\fR"
 182 Lists all flows in \fIbridge\fR, including those normally hidden to
 183 commands such as \fBovs\-ofctl dump\-flows\fR.  Flows set up by mechanisms
 184 such as in-band control and fail-open are hidden from the controller
 185 since it is not allowed to modify or override them.
 186 .SS "BOND COMMANDS"
 187 These commands manage bonded ports on an Open vSwitch's bridges.  To
 188 understand some of these commands, it is important to understand a
 189 detail of the bonding implementation called ``source load balancing''
 190 (SLB).  Instead of directly assigning Ethernet source addresses to
 191 slaves, the bonding implementation computes a function that maps an
 192 48-bit Ethernet source addresses into an 8-bit value (a ``MAC hash''
 193 value).  All of the Ethernet addresses that map to a single 8-bit
 194 value are then assigned to a single slave.
 195 .IP "\fBbond/list\fR"
 196 Lists all of the bonds, and their slaves, on each bridge.
 197 .
 198 .IP "\fBbond/show\fR [\fIport\fR]"
 199 Lists all of the bond-specific information (updelay, downdelay, time
 200 until the next rebalance) about the given bonded \fIport\fR, or all
 201 bonded ports if no \fIport\fR is given.  Also lists information about
 202 each slave: whether it is enabled or disabled, the time to completion
 203 of an updelay or downdelay if one is in progress, whether it is the
 204 active slave, the hashes assigned to the slave.  Any LACP information
 205 related to this bond may be found using the \fBlacp/show\fR command.
 206 .
 207 .IP "\fBbond/migrate\fR \fIport\fR \fIhash\fR \fIslave\fR"
 208 Only valid for SLB bonds.  Assigns a given MAC hash to a new slave.
 209 \fIport\fR specifies the bond port, \fIhash\fR the MAC hash to be
 210 migrated (as a decimal number between 0 and 255), and \fIslave\fR the
 211 new slave to be assigned.
 212 .IP
 213 The reassignment is not permanent: rebalancing or fail-over will
 214 cause the MAC hash to be shifted to a new slave in the usual
 215 manner.
 216 .IP
 217 A MAC hash cannot be migrated to a disabled slave.
 218 .IP "\fBbond/set\-active\-slave\fR \fIport\fR \fIslave\fR"
 219 Sets \fIslave\fR as the active slave on \fIport\fR.  \fIslave\fR must
 220 currently be enabled.
 221 .IP
 222 The setting is not permanent: a new active slave will be selected
 223 if \fIslave\fR becomes disabled.
 224 .IP "\fBbond/enable\-slave\fR \fIport\fR \fIslave\fR"
 225 .IQ "\fBbond/disable\-slave\fR \fIport\fR \fIslave\fR"
 226 Enables (or disables) \fIslave\fR on the given bond \fIport\fR, skipping any
 227 updelay (or downdelay).
 228 .IP
 229 This setting is not permanent: it persists only until the carrier
 230 status of \fIslave\fR changes.
 231 .IP "\fBbond/hash\fR \fImac\fR [\fIvlan\fR] [\fIbasis\fR]"
 232 Returns the hash value which would be used for \fImac\fR with \fIvlan\fR
 233 and \fIbasis\fR if specified.
 234 .
 235 .IP "\fBlacp/show\fR [\fIport\fR]"
 236 Lists all of the LACP related information about the given \fIport\fR:
 237 active or passive, aggregation key, system id, and system priority.  Also
 238 lists information about each slave: whether it is enabled or disabled,
 239 whether it is attached or detached, port id and priority, actor
 240 information, and partner information.  If \fIport\fR is not specified,
 241 then displays detailed information about all interfaces with CFM
 242 enabled.
 243 .SS "DPCTL DATAPATH DEBUGGING COMMANDS"
 244 The primary way to configure \fBovs\-vswitchd\fR is through the Open
 245 vSwitch database, e.g. using \fBovs\-vsctl\fR(8).  These commands
 246 provide a debugging interface for managing datapaths.  They implement
 247 the same features (and syntax) as \fBovs\-dpctl\fR(8).  Unlike
 248 \fBovs\-dpctl\fR(8), these commands work with datapaths that are
 249 integrated into \fBovs\-vswitchd\fR (e.g. the \fBnetdev\fR datapath
 250 type).
 251 .PP
 252 .
 253 .ds DX \fBdpctl/\fR
 254 .de DO
 255 \\$2 \\$1 \\$3
 256 ..
 257 .so lib/dpctl.man
 258 .
 259 .SS "DPIF-NETDEV COMMANDS"
 260 These commands are used to expose internal information (mostly statistics)
 261 about the ``dpif-netdev'' userspace datapath. If there is only one datapath
 262 (as is often the case, unless \fBdpctl/\fR commands are used), the \fIdp\fR
 263 argument can be omitted.
 264 .IP "\fBdpif-netdev/pmd-stats-show\fR [\fIdp\fR]"
 265 Shows performance statistics for each pmd thread of the datapath \fIdp\fR.
 266 The special thread ``main'' sums up the statistics of every non pmd thread.
 267 The sum of ``emc hits'', ``masked hits'' and ``miss'' is the number of
 268 packets received by the datapath.  Cycles are counted using the TSC or similar
 269 facilities (when available on the platform).  To reset these counters use
 270 \fBdpif-netdev/pmd-stats-clear\fR. The duration of one cycle depends on the
 271 measuring infrastructure. ``idle cycles'' refers to cycles spent polling
 272 devices but not receiving any packets. ``processing cycles'' refers to cycles
 273 spent polling devices and successfully receiving packets, plus the cycles
 274 spent processing said packets.
 275 .IP "\fBdpif-netdev/pmd-stats-clear\fR [\fIdp\fR]"
 276 Resets to zero the per pmd thread performance numbers shown by the
 277 \fBdpif-netdev/pmd-stats-show\fR command.  It will NOT reset datapath or
 278 bridge statistics, only the values shown by the above command.
 279 .IP "\fBdpif-netdev/pmd-rxq-show\fR [\fIdp\fR]"
 280 For each pmd thread of the datapath \fIdp\fR shows list of queue-ids with
 281 port names, which this thread polls.
 282 .IP "\fBdpif-netdev/pmd-rxq-rebalance\fR [\fIdp\fR]"
 283 Reassigns rxqs to pmds in the datapath \fIdp\fR based on their current usage.
 284 .
 285 .so lib/netdev-dpdk-unixctl.man
 286 .so ofproto/ofproto-dpif-unixctl.man
 287 .so ofproto/ofproto-unixctl.man
 288 .so lib/vlog-unixctl.man
 289 .so lib/memory-unixctl.man
 290 .so lib/coverage-unixctl.man
 291 .so ofproto/ofproto-tnl-unixctl.man
 292 .
 293 .SH "OPENFLOW IMPLEMENTATION"
 294 .
 295 .PP
 296 This section documents aspects of OpenFlow for which the OpenFlow
 297 specification requires documentation.
 298 .
 299 .SS "Packet buffering."
 300 The OpenFlow specification, version 1.2, says:
 301 .
 302 .IP
 303 Switches that implement buffering are expected to expose, through
 304 documentation, both the amount of available buffering, and the length
 305 of time before buffers may be reused.
 306 .
 307 .PP
 308 Open vSwitch does not maintains any packet buffers.
 309 .
 310 .SS "Bundle lifetime"
 311 The OpenFlow specification, version 1.4, says:
 312 .
 313 .IP
 314 If the switch does not receive any OFPT_BUNDLE_CONTROL or
 315 OFPT_BUNDLE_ADD_MESSAGE message for an opened bundle_id for a switch
 316 defined time greater than 1s, it may send an ofp_error_msg with
 317 OFPET_BUNDLE_FAILED type and OFPBFC_TIMEOUT code.  If the switch does
 318 not receive any new message in a bundle apart from echo request and
 319 replies for a switch defined time greater than 1s, it may send an
 320 ofp_error_msg with OFPET_BUNDLE_FAILED type and OFPBFC_TIMEOUT code.
 321 .
 322 .PP
 323 Open vSwitch implements idle bundle lifetime of 10 seconds.
 324 .
 325 .SH "LIMITS"
 326 .
 327 .PP
 328 We believe these limits to be accurate as of this writing.  These
 329 limits assume the use of the Linux kernel datapath.
 330 .
 331 .IP \(bu
 332 \fBovs\-vswitchd\fR started through \fBovs\-ctl\fR(8) provides a limit of 65535
 333 file descriptors.  The limits on the number of bridges and ports is decided by
 334 the availability of file descriptors.  With the Linux kernel datapath, creation
 335 of a single bridge consumes three file descriptors and adding a port consumes
 336 "n-handler-threads" file descriptors per bridge port.  Performance will degrade
 337 beyond 1,024 ports per bridge due to fixed hash table sizing.  Other platforms
 338 may have different limitations.
 339 .
 340 .IP \(bu
 341 2,048 MAC learning entries per bridge, by default.  (This is
 342 configurable via \fBother\-config:mac\-table\-size\fR in the
 343 \fBBridge\fR table.  See \fBovs\-vswitchd.conf.db\fR(5) for details.)
 344 .
 345 .IP \(bu
 346 Kernel flows are limited only by memory available to the kernel.
 347 Performance will degrade beyond 1,048,576 kernel flows per bridge with
 348 a 32-bit kernel, beyond 262,144 with a 64-bit kernel.
 349 (\fBovs\-vswitchd\fR should never install anywhere near that many
 350 flows.)
 351 .
 352 .IP \(bu
 353 OpenFlow flows are limited only by available memory.  Performance is
 354 linear in the number of unique wildcard patterns.  That is, an
 355 OpenFlow table that contains many flows that all match on the same
 356 fields in the same way has a constant-time lookup, but a table that
 357 contains many flows that match on different fields requires lookup
 358 time linear in the number of flows.
 359 .
 360 .IP \(bu
 361 255 ports per bridge participating in 802.1D Spanning Tree Protocol.
 362 .
 363 .IP \(bu
 364 32 mirrors per bridge.
 365 .
 366 .IP \(bu
 367 15 bytes for the name of a port.  (This is a Linux kernel limitation.)
 368 .
 369 .SH "SEE ALSO"
 370 .BR ovs\-appctl (8),
 371 .BR ovsdb\-server (1).