1ovn-controller(8) Open vSwitch Manual ovn-controller(8)
2
6 ovn-controller - Open Virtual Network local controller
7
9 ovn-controller [options] [ovs-database]
10
12 ovn-controller is the local controller daemon for OVN, the Open Virtual
13 Network. It connects up to the OVN Southbound database (see ovn-sb(5))
14 over the OVSDB protocol, and down to the Open vSwitch database (see
15 ovs-vswitchd.conf.db(5)) over the OVSDB protocol and to ovs-vswitchd(8)
16 via OpenFlow. Each hypervisor and software gateway in an OVN deployment
17 runs its own independent copy of ovn-controller; thus, ovn-controller’s
18 downward connections are machine-local and do not run over a physical
19 network.
20
22 ACL log messages are logged through ovn-controller’s logging mechanism.
23 ACL log entries have the module acl_log at log level info. Configuring
24 logging is described below in the Logging Options section.
25
27 Daemon Options
28 --pidfile[=pidfile]
29 Causes a file (by default, program.pid) to be created indicating
30 the PID of the running process. If the pidfile argument is not
31 specified, or if it does not begin with /, then it is created in
32 .
33 If --pidfile is not specified, no pidfile is created.
35 --overwrite-pidfile
37 By default, when --pidfile is specified and the specified pid‐
38 file already exists and is locked by a running process, the dae‐
39 mon refuses to start. Specify --overwrite-pidfile to cause it to
40 instead overwrite the pidfile.
41 When --pidfile is not specified, this option has no effect.
43 --detach
45 Runs this program as a background process. The process forks,
46 and in the child it starts a new session, closes the standard
47 file descriptors (which has the side effect of disabling logging
48 to the console), and changes its current directory to the root
49 (unless --no-chdir is specified). After the child completes its
50 initialization, the parent exits.
51 --monitor
53 Creates an additional process to monitor this program. If it
54 dies due to a signal that indicates a programming error (SIGA‐
55 BRT, SIGALRM, SIGBUS, SIGFPE, SIGILL, SIGPIPE, SIGSEGV, SIGXCPU,
56 or SIGXFSZ) then the monitor process starts a new copy of it. If
57 the daemon dies or exits for another reason, the monitor process
58 exits.
59 This option is normally used with --detach, but it also func‐
61 tions without it.
62 --no-chdir
64 By default, when --detach is specified, the daemon changes its
65 current working directory to the root directory after it
66 detaches. Otherwise, invoking the daemon from a carelessly cho‐
67 sen directory would prevent the administrator from unmounting
68 the file system that holds that directory.
69 Specifying --no-chdir suppresses this behavior, preventing the
71 daemon from changing its current working directory. This may be
72 useful for collecting core files, since it is common behavior to
73 write core dumps into the current working directory and the root
74 directory is not a good directory to use.
75 This option has no effect when --detach is not specified.
77 --no-self-confinement
79 By default this daemon will try to self-confine itself to work
80 with files under well-known directories whitelisted at build
81 time. It is better to stick with this default behavior and not
82 to use this flag unless some other Access Control is used to
83 confine daemon. Note that in contrast to other access control
84 implementations that are typically enforced from kernel-space
85 (e.g. DAC or MAC), self-confinement is imposed from the user-
86 space daemon itself and hence should not be considered as a full
87 confinement strategy, but instead should be viewed as an addi‐
88 tional layer of security.
89 --user=user:group
91 Causes this program to run as a different user specified in
92 user:group, thus dropping most of the root privileges. Short
93 forms user and :group are also allowed, with current user or
94 group assumed, respectively. Only daemons started by the root
95 user accepts this argument.
96 On Linux, daemons will be granted CAP_IPC_LOCK and
98 CAP_NET_BIND_SERVICES before dropping root privileges. Daemons
99 that interact with a datapath, such as ovs-vswitchd, will be
100 granted three additional capabilities, namely CAP_NET_ADMIN,
101 CAP_NET_BROADCAST and CAP_NET_RAW. The capability change will
102 apply even if the new user is root.
103 On Windows, this option is not currently supported. For security
105 reasons, specifying this option will cause the daemon process
106 not to start.
107 Logging Options
109 -v[spec]
110 --verbose=[spec]
111 Sets logging levels. Without any spec, sets the log level for
112 every module and destination to dbg. Otherwise, spec is a list of
113 words separated by spaces or commas or colons, up to one from each
114 category below:
115 · A valid module name, as displayed by the vlog/list command
117 on ovs-appctl(8), limits the log level change to the speci‐
118 fied module.
119 · syslog, console, or file, to limit the log level change to
121 only to the system log, to the console, or to a file,
122 respectively. (If --detach is specified, the daemon closes
123 its standard file descriptors, so logging to the console
124 will have no effect.)
125 On Windows platform, syslog is accepted as a word and is
127 only useful along with the --syslog-target option (the word
128 has no effect otherwise).
129 · off, emer, err, warn, info, or dbg, to control the log
131 level. Messages of the given severity or higher will be
132 logged, and messages of lower severity will be filtered
133 out. off filters out all messages. See ovs-appctl(8) for a
134 definition of each log level.
135 Case is not significant within spec.
137 Regardless of the log levels set for file, logging to a file will
139 not take place unless --log-file is also specified (see below).
140 For compatibility with older versions of OVS, any is accepted as a
142 word but has no effect.
143 -v
145 --verbose
146 Sets the maximum logging verbosity level, equivalent to --ver‐
147 bose=dbg.
148 -vPATTERN:destination:pattern
150 --verbose=PATTERN:destination:pattern
151 Sets the log pattern for destination to pattern. Refer to
152 ovs-appctl(8) for a description of the valid syntax for pattern.
153 -vFACILITY:facility
155 --verbose=FACILITY:facility
156 Sets the RFC5424 facility of the log message. facility can be one
157 of kern, user, mail, daemon, auth, syslog, lpr, news, uucp, clock,
158 ftp, ntp, audit, alert, clock2, local0, local1, local2, local3,
159 local4, local5, local6 or local7. If this option is not specified,
160 daemon is used as the default for the local system syslog and
161 local0 is used while sending a message to the target provided via
162 the --syslog-target option.
163 --log-file[=file]
165 Enables logging to a file. If file is specified, then it is used
166 as the exact name for the log file. The default log file name used
167 if file is omitted is /var/log/ovn/program.log.
168 --syslog-target=host:port
170 Send syslog messages to UDP port on host, in addition to the sys‐
171 tem syslog. The host must be a numerical IP address, not a host‐
172 name.
173 --syslog-method=method
175 Specify method as how syslog messages should be sent to syslog
176 daemon. The following forms are supported:
177 · libc, to use the libc syslog() function. Downside of using
179 this options is that libc adds fixed prefix to every mes‐
180 sage before it is actually sent to the syslog daemon over
181 /dev/log UNIX domain socket.
182 · unix:file, to use a UNIX domain socket directly. It is pos‐
184 sible to specify arbitrary message format with this option.
185 However, rsyslogd 8.9 and older versions use hard coded
186 parser function anyway that limits UNIX domain socket use.
187 If you want to use arbitrary message format with older
188 rsyslogd versions, then use UDP socket to localhost IP
189 address instead.
190 · udp:ip:port, to use a UDP socket. With this method it is
192 possible to use arbitrary message format also with older
193 rsyslogd. When sending syslog messages over UDP socket
194 extra precaution needs to be taken into account, for exam‐
195 ple, syslog daemon needs to be configured to listen on the
196 specified UDP port, accidental iptables rules could be
197 interfering with local syslog traffic and there are some
198 security considerations that apply to UDP sockets, but do
199 not apply to UNIX domain sockets.
200 · null, to discard all messages logged to syslog.
202 The default is taken from the OVS_SYSLOG_METHOD environment vari‐
204 able; if it is unset, the default is libc.
205 PKI Options
207 PKI configuration is required in order to use SSL for the connections
208 to the Northbound and Southbound databases.
209 -p privkey.pem
211 --private-key=privkey.pem
212 Specifies a PEM file containing the private key used as
213 identity for outgoing SSL connections.
214 -c cert.pem
216 --certificate=cert.pem
217 Specifies a PEM file containing a certificate that certi‐
218 fies the private key specified on -p or --private-key to be
219 trustworthy. The certificate must be signed by the certifi‐
220 cate authority (CA) that the peer in SSL connections will
221 use to verify it.
222 -C cacert.pem
224 --ca-cert=cacert.pem
225 Specifies a PEM file containing the CA certificate for ver‐
226 ifying certificates presented to this program by SSL peers.
227 (This may be the same certificate that SSL peers use to
228 verify the certificate specified on -c or --certificate, or
229 it may be a different one, depending on the PKI design in
230 use.)
231 -C none
233 --ca-cert=none
234 Disables verification of certificates presented by SSL
235 peers. This introduces a security risk, because it means
236 that certificates cannot be verified to be those of known
237 trusted hosts.
238 --bootstrap-ca-cert=cacert.pem
240 When cacert.pem exists, this option has the same effect
241 as -C or --ca-cert. If it does not exist, then the exe‐
242 cutable will attempt to obtain the CA certificate from
243 the SSL peer on its first SSL connection and save it to
244 the named PEM file. If it is successful, it will immedi‐
245 ately drop the connection and reconnect, and from then on
246 all SSL connections must be authenticated by a certifi‐
247 cate signed by the CA certificate thus obtained.
248 This option exposes the SSL connection to a man-in-the-
250 middle attack obtaining the initial CA certificate, but
251 it may be useful for bootstrapping.
252 This option is only useful if the SSL peer sends its CA
254 certificate as part of the SSL certificate chain. The SSL
255 protocol does not require the server to send the CA cer‐
256 tificate.
257 This option is mutually exclusive with -C and --ca-cert.
259 --peer-ca-cert=peer-cacert.pem
261 Specifies a PEM file that contains one or more additional
262 certificates to send to SSL peers. peer-cacert.pem should
263 be the CA certificate used to sign the program’s own cer‐
264 tificate, that is, the certificate specified on -c or
265 --certificate. If the program’s certificate is self-
266 signed, then --certificate and --peer-ca-cert should
267 specify the same file.
268 This option is not useful in normal operation, because
270 the SSL peer must already have the CA certificate for the
271 peer to have any confidence in the program’s identity.
272 However, this offers a way for a new installation to
273 bootstrap the CA certificate on its first SSL connection.
274 Other Options
276 -h
277 --help
278 Prints a brief help message to the console.
279 -V
281 --version
282 Prints version information to the console.
283
285 ovn-controller retrieves most of its configuration information from the
286 local Open vSwitch’s ovsdb-server instance. The default location is
287 db.sock in the local Open vSwitch’s "run" directory. It may be overrid‐
288 den by specifying the ovs-database argument as an OVSDB active or pas‐
289 sive connection method, as described in ovsdb(7).
290 ovn-controller assumes it gets configuration information from the fol‐
292 lowing keys in the Open_vSwitch table of the local OVS instance:
293 external_ids:system-id
295 The chassis name to use in the Chassis table.
296 external_ids:hostname
298 The hostname to use in the Chassis table.
299 external_ids:ovn-bridge
301 The integration bridge to which logical ports are
302 attached. The default is br-int. If this bridge does not
303 exist when ovn-controller starts, it will be created
304 automatically with the default configuration suggested in
305 ovn-architecture(7).
306 external_ids:ovn-bridge-datapath-type
308 This configuration is optional. If set, then the datapath
309 type of the integration bridge will be set to the config‐
310 ured value. If this option is not set, then ovn-con‐
311 troller will not modify the existing datapath-type of the
312 integration bridge.
313 external_ids:ovn-remote
315 The OVN database that this system should connect to for
316 its configuration, in one of the same forms documented
317 above for the ovs-database.
318 external_ids:ovn-remote-probe-interval
320 The inactivity probe interval of the connection to the
321 OVN database, in milliseconds. If the value is zero, it
322 disables the connection keepalive feature.
323 If the value is nonzero, then it will be forced to a
325 value of at least 1000 ms.
326 external_ids:ovn-openflow-probe-interval
328 The inactivity probe interval of the OpenFlow connection
329 to the OpenvSwitch integration bridge, in seconds. If the
330 value is zero, it disables the connection keepalive fea‐
331 ture.
332 If the value is nonzero, then it will be forced to a
334 value of at least 5s.
335 external_ids:ovn-encap-type
337 The encapsulation type that a chassis should use to con‐
338 nect to this node. Multiple encapsulation types may be
339 specified with a comma-separated list. Each listed encap‐
340 sulation type will be paired with ovn-encap-ip.
341 Supported tunnel types for connecting hypervisors are
343 geneve and stt. Gateways may use geneve, vxlan, or stt.
344 Due to the limited amount of metadata in vxlan, the capa‐
346 bilities and performance of connected gateways will be
347 reduced versus other tunnel formats.
348 external_ids:ovn-encap-ip
350 The IP address that a chassis should use to connect to
351 this node using encapsulation types specified by exter‐
352 nal_ids:ovn-encap-type.
353 external_ids:ovn-bridge-mappings
355 A list of key-value pairs that map a physical network
356 name to a local ovs bridge that provides connectivity to
357 that network. An example value mapping two physical net‐
358 work names to two ovs bridges would be: phys‐
359 net1:br-eth0,physnet2:br-eth1.
360 external_ids:ovn-encap-csum
362 ovn-encap-csum indicates that encapsulation checksums can
363 be transmitted and received with reasonable performance.
364 It is a hint to senders transmitting data to this chassis
365 that they should use checksums to protect OVN metadata.
366 Set to true to enable or false to disable. Depending on
367 the capabilities of the network interface card, enabling
368 encapsulation checksum may incur performance loss. In
369 such cases, encapsulation checksums can be disabled.
370 external_ids:ovn-cms-options
372 A list of options that will be consumed by the CMS Plugin
373 and which specific to this particular chassis. An example
374 would be: cms_option1,cms_option2:foo.
375 external_ids:ovn-transport-zones
377 The transport zone(s) that this chassis belongs to.
378 Transport zones is a way to group different chassis so
379 that tunnels are only formed between members of the same
380 group(s). Multiple transport zones may be specified with
381 a comma-separated list. For example: tz1,tz2,tz3.
382 If not set, the Chassis will be considered part of a
384 default transport zone.
385 external_ids:ovn-chassis-mac-mappings
387 A list of key-value pairs that map a chassis specific mac
388 to a physical network name. An example value mapping two
389 chassis macs to two physical network names would be:
390 physnet1:aa:bb:cc:dd:ee:ff,physnet2:a1:b2:c3:d4:e5:f6.
391 These are the macs that ovn-controller will replace a
392 router port mac with, if packet is going from a distrib‐
393 uted router port on vlan type logical switch.
394 ovn-controller reads the following values from the Open_vSwitch data‐
396 base of the local OVS instance:
397 datapath-type from Bridge table
399 This value is read from local OVS integration bridge row
400 of Bridge table and populated in external_ids:datapath-
401 type of the Chassis table in the OVN_Southbound database.
402 iface-types from Open_vSwitch table
404 This value is populated in external_ids:iface-types of
405 the Chassis table in the OVN_Southbound database.
406 private_key, certificate, ca_cert, and bootstrap_ca_cert from
408 SSL table
409 These values provide the SSL configuration used for con‐
410 necting to the OVN southbound database server when an SSL
411 connection type is configured via exter‐
412 nal_ids:ovn-remote. Note that this SSL configuration can
413 also be provided via command-line options, the configura‐
414 tion in the database takes precedence if both are
415 present.
416
418 ovn-controller uses a number of external_ids keys in the Open vSwitch
419 database to keep track of ports and interfaces. For proper operation,
420 users should not change or clear these keys:
421 external_ids:ovn-chassis-id in the Port table
423 The presence of this key identifies a tunnel port within
424 the integration bridge as one created by ovn-controller
425 to reach a remote chassis. Its value is the chassis ID of
426 the remote chassis.
427 external_ids:ct-zone-* in the Bridge table
429 Logical ports and gateway routers are assigned a connec‐
430 tion tracking zone by ovn-controller for stateful ser‐
431 vices. To keep state across restarts of ovn-controller,
432 these keys are stored in the integration bridge’s Bridge
433 table. The name contains a prefix of ct-zone- followed by
434 the name of the logical port or gateway router’s zone
435 key. The value for this key identifies the zone used for
436 this port.
437 external_ids:ovn-localnet-port in the Port table
439 The presence of this key identifies a patch port as one
440 created by ovn-controller to connect the integration
441 bridge and another bridge to implement a localnet logical
442 port. Its value is the name of the logical port with type
443 set to localnet that the port implements. See exter‐
444 nal_ids:ovn-bridge-mappings, above, for more information.
445 Each localnet logical port is implemented as a pair of
447 patch ports, one in the integration bridge, one in a dif‐
448 ferent bridge, with the same external_ids:ovn-local‐
449 net-port value.
450 external_ids:ovn-l2gateway-port in the Port table
452 The presence of this key identifies a patch port as one
453 created by ovn-controller to connect the integration
454 bridge and another bridge to implement a l2gateway logi‐
455 cal port. Its value is the name of the logical port with
456 type set to l2gateway that the port implements. See
457 external_ids:ovn-bridge-mappings, above, for more infor‐
458 mation.
459 Each l2gateway logical port is implemented as a pair of
461 patch ports, one in the integration bridge, one in a dif‐
462 ferent bridge, with the same external_ids:ovn-l2gate‐
463 way-port value.
464 external-ids:ovn-l3gateway-port in the Port table
466 This key identifies a patch port as one created by
467 ovn-controller to implement a l3gateway logical port. Its
468 value is the name of the logical port with type set to
469 l3gateway. This patch port is similar to the OVN logical
470 patch port, except that l3gateway port can only be bound
471 to a paticular chassis.
472 external-ids:ovn-logical-patch-port in the Port table
474 This key identifies a patch port as one created by
475 ovn-controller to implement an OVN logical patch port
476 within the integration bridge. Its value is the name of
477 the OVN logical patch port that it implements.
478
480 ovn-controller reads from much of the OVN_Southbound database to guide
481 its operation. ovn-controller also writes to the following tables:
482 Chassis
484 Upon startup, ovn-controller creates a row in this table
485 to represent its own chassis. Upon graceful termination,
486 e.g. with ovs-appctl -t ovn-controller exit (but not
487 SIGTERM), ovn-controller removes its row.
488 Encap Upon startup, ovn-controller creates a row or rows in
490 this table that represent the tunnel encapsulations by
491 which its chassis can be reached, and points its Chassis
492 row to them. Upon graceful termination, ovn-controller
493 removes these rows.
494 Port_Binding
496 At runtime, ovn-controller sets the chassis columns of
497 ports that are resident on its chassis to point to its
498 Chassis row, and, conversely, clears the chassis column
499 of ports that point to its Chassis row but are no longer
500 resident on its chassis. The chassis column has a weak
501 reference type, so when ovn-controller gracefully exits
502 and removes its Chassis row, the database server automat‐
503 ically clears any remaining references to that row.
504 MAC_Binding
506 At runtime, ovn-controller updates the MAC_Binding table
507 as instructed by put_arp and put_nd logical actions.
508 These changes persist beyond the lifetime of ovn-con‐
509 troller.
510
512 ovs-appctl can send commands to a running ovn-controller process. The
513 currently supported commands are described below.
514 exit Causes ovn-controller to gracefully terminate.
516 ct-zone-list
518 Lists each local logical port and its connection tracking
519 zone.
520 meter-table-list
522 Lists each meter table entry and its local meter id.
523 group-table-list
525 Lists each group table entry and its local group id.
526 inject-pkt microflow
528 Injects microflow into the connected Open vSwitch
529 instance. microflow must contain an ingress logical port
530 (inport argument) that is present on the Open vSwitch
531 instance.
532 The microflow argument describes the packet whose for‐
534 warding is to be simulated, in the syntax of an OVN logi‐
535 cal expression, as described in ovn-sb(5), to express
536 constraints. The parser understands prerequisites; for
537 example, if the expression refers to ip4.src, there is no
538 need to explicitly state ip4 or eth.type == 0x800.
539 connection-status
541 Show OVN SBDB connection status for the chassis.
542Open vSwitch 2.12.0 ovn-controller ovn-controller(8)