1TEAMD.CONF(5) Team daemon configuration TEAMD.CONF(5)
2
6 teamd.conf — libteam daemon configuration file
7
9 teamd uses JSON format configuration.
10
12 device (string)
13 Desired name of new team device.
14 debug_level (int)
16 Level of debug messages. The higher it is the more debug mes‐
17 sages will be printed. It is the same as adding "-g" command
18 line options.
19 Default: 0 (disabled)
21 hwaddr (string)
23 Desired hardware address of new team device. Usual MAC address
24 format is accepted.
25 runner.name (string)
27 Name of team device. The following runners are available:
28 broadcast — Simple runner which directs the team device to
30 transmit packets via all ports.
31 roundrobin — Simple runner which directs the team device to
33 transmits packets in a round-robin fashion.
34 random — Simple runner which directs the team device to trans‐
36 mits packets on a randomly selected port.
37 activebackup — Watches for link changes and selects active port
39 to be used for data transfers.
40 loadbalance — To do passive load balancing, runner only sets up
42 BPF hash function which will determine port for packet transmit.
43 To do active load balancing, runner moves hashes among available
44 ports trying to reach perfect balance.
45 lacp — Implements 802.3ad LACP protocol. Can use same Tx port
47 selection possibilities as loadbalance runner.
48 Default: roundrobin
50 notify_peers.count (int)
52 Number of bursts of unsolicited NAs and gratuitous ARP packets
53 sent after port is enabled or disabled.
54 Default: 0 (disabled)
56 Default for activebackup runner: 1
58 notify_peers.interval (int)
60 Value is positive number in milliseconds. Specifies an interval
61 between bursts of notify-peer packets.
62 Default: 0
64 mcast_rejoin.count (int)
66 Number of bursts of multicast group rejoin requests sent after
67 port is enabled or disabled.
68 Default: 0 (disabled)
70 Default for activebackup runner: 1
72 mcast_rejoin.interval (int)
74 Value is positive number in milliseconds. Specifies an interval
75 between bursts of multicast group rejoin requests.
76 Default: 0
78 link_watch.name | ports.PORTIFNAME.link_watch.name (string)
80 Name of link watcher to be used. The following link watchers are
81 available:
82 ethtool — Uses Libteam lib to get port ethtool state changes.
84 arp_ping — ARP requests are sent through a port. If an ARP reply
86 is received, the link is considered to be up.
87 nsna_ping — Similar to the previous, except that it uses IPv6
89 Neighbor Solicitation / Neighbor Advertisement mechanism. This
90 is an alternative to arp_ping and becomes handy in pure-IPv6 en‐
91 vironments.
92 ports (object)
94 List of ports, network devices, to be used in a team device.
95 See examples for more information.
97 ports.PORTIFNAME.queue_id (int)
99 ID of queue which this port should be mapped to.
100 Default: None
102
104 runner.hwaddr_policy (string)
105 This defines the policy of how hardware addresses of team device
106 and port devices should be set during the team lifetime. The
107 following are available:
108 same_all — All ports will always have the same hardware address
110 as the associated team device.
111 by_active — Team device adopts the hardware address of the cur‐
113 rently active port. This is useful when the port device is not
114 able to change its hardware address.
115 only_active — Only the active port adopts the hardware address
117 of the team device. The others have their own.
118 Default: same_all
120 ports.PORTIFNAME.prio (int)
122 Port priority. The higher number means higher priority.
123 Default: 0
125 ports.PORTIFNAME.sticky (bool)
127 Flag which indicates if the port is sticky. If set, it means the
128 port does not get unselected if another port with higher prior‐
129 ity or better parameters becomes available.
130 Default: false
132
134 runner.tx_hash (array)
135 List of fragment types (strings) which should be used for packet
136 Tx hash computation. The following are available:
137 eth — Uses source and destination MAC addresses.
139 vlan — Uses VLAN id.
141 ipv4 — Uses source and destination IPv4 addresses.
143 ipv6 — Uses source and destination IPv6 addresses.
145 ip — Uses source and destination IPv4 and IPv6 addresses.
147 l3 — Uses source and destination IPv4 and IPv6 addresses.
149 tcp — Uses source and destination TCP ports.
151 udp — Uses source and destination UDP ports.
153 sctp — Uses source and destination SCTP ports.
155 l4 — Uses source and destination TCP and UDP and SCTP ports.
157 Default: ["eth", "ipv4", "ipv6"]
159 runner.tx_balancer.name (string)
161 Name of active Tx balancer. Active Tx balancing is disabled by
162 default. The only value available is basic.
163 Default: None
165 runner.tx_balancer.balancing_interval (int)
167 In tenths of a second. Periodic interval between rebalancing.
168 Default: 50
170
172 runner.active (bool)
173 If active is true LACPDU frames are sent along the configured
174 links periodically. If not, it acts as "speak when spoken to".
175 Default: true
177 runner.fast_rate (bool)
179 Option specifies the rate at which our link partner is asked to
180 transmit LACPDU packets. If this is true then packets will be
181 sent once per second. Otherwise they will be sent every 30 sec‐
182 onds.
183 Default: false
185 runner.tx_hash (array)
187 Same as for load balance runner.
188 runner.tx_balancer.name (string)
190 Same as for load balance runner.
191 runner.tx_balancer.balancing_interval (int)
193 Same as for load balance runner.
194 runner.sys_prio (int)
196 System priority, value can be 0 – 65535.
197 Default: 65535
199 runner.min_ports (int)
201 Specifies the minimum number of ports that must be active before
202 asserting carrier in the master interface, value can be 1 –
203 1024.
204 Default: 1
206 runner.agg_select_policy (string)
208 This selects the policy of how the aggregators will be selected.
209 The following are available:
210 lacp_prio — Aggregator with highest priority according to LACP
212 standard will be selected. Aggregator priority is affected by
213 per-port option lacp_prio.
214 lacp_prio_stable — Same as previous one, except do not replace
216 selected aggregator if it is still usable.
217 bandwidth — Select aggregator with highest total bandwidth.
219 count — Select aggregator with highest number of ports.
221 port_config — Aggregator with highest priority according to per-
223 port options prio and sticky will be selected. This means that
224 the aggregator containing the port with the highest priority
225 will be selected unless at least one of the ports in the cur‐
226 rently selected aggregator is sticky.
227 Default: lacp_prio
229 ports.PORTIFNAME.lacp_prio (int)
231 Port priority according to LACP standard. The lower number means
232 higher priority.
233 Default: 255
235 ports.PORTIFNAME.lacp_key (int)
237 Port key according to LACP standard. It is only possible to ag‐
238 gregate ports with the same key.
239 Default: 0
241
243 link_watch.delay_up | ports.PORTIFNAME.link_watch.delay_up (int)
244 Value is a positive number in milliseconds. It is the delay be‐
245 tween the link coming up and the runner being notified about it.
246 Default: 0
248 link_watch.delay_down | ports.PORTIFNAME.link_watch.delay_down (int)
250 Value is a positive number in milliseconds. It is the delay be‐
251 tween the link going down and the runner being notified about
252 it.
253 Default: 0
255
257 link_watch.interval | ports.PORTIFNAME.link_watch.interval (int)
258 Value is a positive number in milliseconds. It is the interval
259 between ARP requests being sent.
260 Default: 1000
262 link_watch.init_wait | ports.PORTIFNAME.link_watch.init_wait (int)
264 Value is a positive number in milliseconds. It is the delay be‐
265 tween link watch initialization and the first ARP request being
266 sent.
267 Default: 0
269 link_watch.missed_max | ports.PORTIFNAME.link_watch.missed_max (int)
271 Maximum number of missed ARP replies. If this number is ex‐
272 ceeded, link is reported as down.
273 Default: 3
275 link_watch.source_host | ports.PORTIFNAME.link_watch.source_host (host‐
277 name)
278 Hostname to be converted to IP address which will be filled into
279 ARP request as source address.
280 Default: 0.0.0.0
282 link_watch.target_host | ports.PORTIFNAME.link_watch.target_host (host‐
284 name)
285 Hostname to be converted to IP address which will be filled into
286 ARP request as destination address.
287 link_watch.validate_active | ports.PORTIFNAME.link_watch.validate_ac‐
289 tive (bool)
290 Validate received ARP packets on active ports. If this is not
291 set, all incoming ARP packets will be considered as a good re‐
292 ply.
293 Default: false
295 link_watch.validate_inactive | ports.PORTIFNAME.link_watch.validate_in‐
297 active (bool)
298 Validate received ARP packets on inactive ports. If this is not
299 set, all incoming ARP packets will be considered as a good re‐
300 ply.
301 Default: false
303 link_watch.vlanid | ports.PORTIFNAME.link_watch.vlanid (int)
305 By default, ARP requests are sent without VLAN tags. This option
306 causes outgoing ARP requests to be sent with the specified VLAN
307 ID number.
308 Default: None
310 link_watch.send_always | ports.PORTIFNAME.link_watch.send_always (bool)
312 By default, ARP requests are sent on active ports only. This op‐
313 tion allows sending even on inactive ports.
314 Default: false
316
318 link_watch.interval | ports.PORTIFNAME.link_watch.interval (int)
319 Value is a positive number in milliseconds. It is the interval
320 between sending NS packets.
321 Default: 1000
323 link_watch.init_wait | ports.PORTIFNAME.link_watch.init_wait (int)
325 Value is a positive number in milliseconds. It is the delay be‐
326 tween link watch initialization and the first NS packet being
327 sent.
328 link_watch.missed_max | ports.PORTIFNAME.link_watch.missed_max (int)
330 Maximum number of missed NA reply packets. If this number is ex‐
331 ceeded, link is reported as down.
332 Default: 3
334 link_watch.target_host | ports.PORTIFNAME.link_watch.target_host (host‐
336 name)
337 Hostname to be converted to IPv6 address which will be filled
338 into NS packet as target address.
339 link_watch_policy (string)
341 Name of link-watchers evaluation policy. Available options are
342 following:
343 any — Link is up if any of the link-watchers reports the link
345 up.
346 all — Link is up if all of the link-watchers reports the link
348 up.
349 Default: any
351
353 {
354 "device": "team0",
355 "runner": {"name": "roundrobin"},
356 "ports": {"eth1": {}, "eth2": {}}
357 }
358 Very basic configuration.
360 {
362 "device": "team0",
363 "runner": {"name": "activebackup"},
364 "link_watch": {"name": "ethtool"},
365 "ports": {
366 "eth1": {
367 "prio": -10,
368 "sticky": true
369 },
370 "eth2": {
371 "prio": 100
372 }
373 }
374 }
375 This configuration uses active-backup runner with ethtool link watcher.
377 Port eth2 has higher priority, but the sticky flag ensures that if eth1
378 becomes active, it stays active while the link remains up.
379 {
381 "device": "team0",
382 "runner": {"name": "activebackup"},
383 "link_watch": {
384 "name": "ethtool",
385 "delay_up": 2500,
386 "delay_down": 1000
387 },
388 "ports": {
389 "eth1": {
390 "prio": -10,
391 "sticky": true
392 },
393 "eth2": {
394 "prio": 100
395 }
396 }
397 }
398 Similar to the previous one. Only difference is that link changes are
400 not propagated to the runner immediately, but delays are applied.
401 {
403 "device": "team0",
404 "runner": {"name": "activebackup"},
405 "link_watch": {
406 "name": "arp_ping",
407 "interval": 100,
408 "missed_max": 30,
409 "target_host": "192.168.23.1"
410 },
411 "ports": {
412 "eth1": {
413 "prio": -10,
414 "sticky": true
415 },
416 "eth2": {
417 "prio": 100
418 }
419 }
420 }
421 This configuration uses ARP ping link watch.
423 {
425 "device": "team0",
426 "runner": {"name": "activebackup"},
427 "link_watch": [
428 {
429 "name": "arp_ping",
430 "interval": 100,
431 "missed_max": 30,
432 "target_host": "192.168.23.1"
433 },
434 {
435 "name": "arp_ping",
436 "interval": 50,
437 "missed_max": 20,
438 "target_host": "192.168.24.1"
439 }
440 ],
441 "ports": {
442 "eth1": {
443 "prio": -10,
444 "sticky": true
445 },
446 "eth2": {
447 "prio": 100
448 }
449 }
450 }
451 Similar to the previous one, only this time two link watchers are used
453 at the same time.
454 {
456 "device": "team0",
457 "runner": {"name": "activebackup"},
458 "link_watch_policy": "all",
459 "link_watch": [
460 {
461 "name": "arp_ping",
462 "interval": 100,
463 "missed_max": 30,
464 "target_host": "192.168.23.1"
465 },
466 {
467 "name": "ethtool"
468 }
469 ],
470 "ports": {
471 "eth1": {
472 "prio": -10,
473 "sticky": true
474 },
475 "eth2": {
476 "prio": 100
477 }
478 }
479 }
480 Two link-watchers are used at the same time. Link is up only if all
482 configured link-watchers report link is up.
483 {
485 "device": "team0",
486 "runner": {
487 "name": "loadbalance",
488 "tx_hash": ["eth", "ipv4", "ipv6"]
489 },
490 "ports": {"eth1": {}, "eth2": {}}
491 }
492 Configuration for hash-based passive Tx load balancing.
494 {
496 "device": "team0",
497 "runner": {
498 "name": "loadbalance",
499 "tx_hash": ["eth", "ipv4", "ipv6"],
500 "tx_balancer": {
501 "name": "basic"
502 }
503 },
504 "ports": {"eth1": {}, "eth2": {}}
505 }
506 Configuration for active Tx load balancing using basic load balancer.
508 {
510 "device": "team0",
511 "runner": {
512 "name": "lacp",
513 "active": true,
514 "fast_rate": true,
515 "tx_hash": ["eth", "ipv4", "ipv6"]
516 },
517 "link_watch": {"name": "ethtool"},
518 "ports": {"eth1": {}, "eth2": {}}
519 }
520 Configuration for connection to LACP capable counterpart.
522
524 teamd(8), teamdctl(8), teamnl(8), bond2team(1)
525
527 Jiri Pirko is the original author and current maintainer of libteam.
528libteam 2013-07-09 TEAMD.CONF(5)