blob: 2cdb8b66caa934b04f42d4290a297664dcd15251 [file] [log] [blame]
Linus Torvalds1da177e2005-04-16 15:20:36 -07001
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002 Linux Ethernet Bonding Driver HOWTO
3
Ben Hutchingsad246c92011-04-26 15:25:52 +00004 Latest update: 27 April 2011
Linus Torvalds1da177e2005-04-16 15:20:36 -07005
6Initial release : Thomas Davis <tadavis at lbl.gov>
7Corrections, HA extensions : 2000/10/03-15 :
8 - Willy Tarreau <willy at meta-x.org>
9 - Constantine Gavrilov <const-g at xpert.com>
10 - Chad N. Tindel <ctindel at ieee dot org>
11 - Janice Girouard <girouard at us dot ibm dot com>
12 - Jay Vosburgh <fubar at us dot ibm dot com>
13
14Reorganized and updated Feb 2005 by Jay Vosburgh
Auke Kok6224e012006-06-08 11:15:35 -070015Added Sysfs information: 2006/04/24
16 - Mitch Williams <mitch.a.williams at intel.com>
Linus Torvalds1da177e2005-04-16 15:20:36 -070017
Jay Vosburgh00354cf2005-07-21 12:18:02 -070018Introduction
19============
Linus Torvalds1da177e2005-04-16 15:20:36 -070020
Jay Vosburgh00354cf2005-07-21 12:18:02 -070021 The Linux bonding driver provides a method for aggregating
22multiple network interfaces into a single logical "bonded" interface.
23The behavior of the bonded interfaces depends upon the mode; generally
24speaking, modes provide either hot standby or load balancing services.
25Additionally, link integrity monitoring may be performed.
26
27 The bonding driver originally came from Donald Becker's
28beowulf patches for kernel 2.0. It has changed quite a bit since, and
29the original tools from extreme-linux and beowulf sites will not work
30with this version of the driver.
31
32 For new versions of the driver, updated userspace tools, and
33who to ask for help, please follow the links at the end of this file.
Linus Torvalds1da177e2005-04-16 15:20:36 -070034
35Table of Contents
36=================
37
381. Bonding Driver Installation
39
402. Bonding Driver Options
41
423. Configuring Bonding Devices
Auke Kok6224e012006-06-08 11:15:35 -0700433.1 Configuration with Sysconfig Support
443.1.1 Using DHCP with Sysconfig
453.1.2 Configuring Multiple Bonds with Sysconfig
463.2 Configuration with Initscripts Support
473.2.1 Using DHCP with Initscripts
483.2.2 Configuring Multiple Bonds with Initscripts
493.3 Configuring Bonding Manually with Ifenslave
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700503.3.1 Configuring Multiple Bonds Manually
Auke Kok6224e012006-06-08 11:15:35 -0700513.4 Configuring Bonding Manually via Sysfs
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +0000523.5 Configuration with Interfaces Support
533.6 Overriding Configuration for Special Cases
Linus Torvalds1da177e2005-04-16 15:20:36 -070054
Auke Kok6224e012006-06-08 11:15:35 -0700554. Querying Bonding Configuration
564.1 Bonding Configuration
574.2 Network Configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -070058
Auke Kok6224e012006-06-08 11:15:35 -0700595. Switch Configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -070060
Auke Kok6224e012006-06-08 11:15:35 -0700616. 802.1q VLAN Support
Linus Torvalds1da177e2005-04-16 15:20:36 -070062
Auke Kok6224e012006-06-08 11:15:35 -0700637. Link Monitoring
647.1 ARP Monitor Operation
657.2 Configuring Multiple ARP Targets
667.3 MII Monitor Operation
Linus Torvalds1da177e2005-04-16 15:20:36 -070067
Auke Kok6224e012006-06-08 11:15:35 -0700688. Potential Trouble Sources
698.1 Adventures in Routing
708.2 Ethernet Device Renaming
718.3 Painfully Slow Or No Failed Link Detection By Miimon
Linus Torvalds1da177e2005-04-16 15:20:36 -070072
Auke Kok6224e012006-06-08 11:15:35 -0700739. SNMP agents
Linus Torvalds1da177e2005-04-16 15:20:36 -070074
Auke Kok6224e012006-06-08 11:15:35 -07007510. Promiscuous mode
Linus Torvalds1da177e2005-04-16 15:20:36 -070076
Auke Kok6224e012006-06-08 11:15:35 -07007711. Configuring Bonding for High Availability
7811.1 High Availability in a Single Switch Topology
7911.2 High Availability in a Multiple Switch Topology
8011.2.1 HA Bonding Mode Selection for Multiple Switch Topology
8111.2.2 HA Link Monitoring for Multiple Switch Topology
Linus Torvalds1da177e2005-04-16 15:20:36 -070082
Auke Kok6224e012006-06-08 11:15:35 -07008312. Configuring Bonding for Maximum Throughput
8412.1 Maximum Throughput in a Single Switch Topology
8512.1.1 MT Bonding Mode Selection for Single Switch Topology
8612.1.2 MT Link Monitoring for Single Switch Topology
8712.2 Maximum Throughput in a Multiple Switch Topology
8812.2.1 MT Bonding Mode Selection for Multiple Switch Topology
8912.2.2 MT Link Monitoring for Multiple Switch Topology
Linus Torvalds1da177e2005-04-16 15:20:36 -070090
Auke Kok6224e012006-06-08 11:15:35 -07009113. Switch Behavior Issues
9213.1 Link Establishment and Failover Delays
9313.2 Duplicated Incoming Packets
Linus Torvalds1da177e2005-04-16 15:20:36 -070094
Auke Kok6224e012006-06-08 11:15:35 -07009514. Hardware Specific Considerations
9614.1 IBM BladeCenter
Jay Vosburgh00354cf2005-07-21 12:18:02 -070097
Auke Kok6224e012006-06-08 11:15:35 -07009815. Frequently Asked Questions
Jay Vosburgh00354cf2005-07-21 12:18:02 -070099
Auke Kok6224e012006-06-08 11:15:35 -070010016. Resources and Links
Linus Torvalds1da177e2005-04-16 15:20:36 -0700101
102
1031. Bonding Driver Installation
104==============================
105
106 Most popular distro kernels ship with the bonding driver
Cong Wangb1098bb2013-05-27 15:49:16 +0000107already available as a module. If your distro does not, or you
Linus Torvalds1da177e2005-04-16 15:20:36 -0700108have need to compile bonding from source (e.g., configuring and
109installing a mainline kernel from kernel.org), you'll need to perform
110the following steps:
111
1121.1 Configure and build the kernel with bonding
113-----------------------------------------------
114
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700115 The current version of the bonding driver is available in the
Linus Torvalds1da177e2005-04-16 15:20:36 -0700116drivers/net/bonding subdirectory of the most recent kernel source
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700117(which is available on http://kernel.org). Most users "rolling their
118own" will want to use the most recent kernel from kernel.org.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700119
120 Configure kernel with "make menuconfig" (or "make xconfig" or
121"make config"), then select "Bonding driver support" in the "Network
122device support" section. It is recommended that you configure the
123driver as module since it is currently the only way to pass parameters
124to the driver or configure more than one bonding device.
125
Cong Wangb1098bb2013-05-27 15:49:16 +0000126 Build and install the new kernel and modules.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700127
Cong Wangb1098bb2013-05-27 15:49:16 +00001281.2 Bonding Control Utility
Linus Torvalds1da177e2005-04-16 15:20:36 -0700129-------------------------------------
130
Cong Wangb1098bb2013-05-27 15:49:16 +0000131 It is recommended to configure bonding via iproute2 (netlink)
132or sysfs, the old ifenslave control utility is obsolete.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700133
1342. Bonding Driver Options
135=========================
136
Jay Vosburgh9a6c6862007-11-13 20:25:48 -0800137 Options for the bonding driver are supplied as parameters to the
138bonding module at load time, or are specified via sysfs.
139
140 Module options may be given as command line arguments to the
141insmod or modprobe command, but are usually specified in either the
Lucas De Marchi970e2482012-03-30 13:37:16 -0700142/etc/modrobe.d/*.conf configuration files, or in a distro-specific
143configuration file (some of which are detailed in the next section).
Jay Vosburgh9a6c6862007-11-13 20:25:48 -0800144
145 Details on bonding support for sysfs is provided in the
146"Configuring Bonding Manually via Sysfs" section, below.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700147
148 The available bonding driver parameters are listed below. If a
149parameter is not specified the default value is used. When initially
150configuring a bond, it is recommended "tail -f /var/log/messages" be
151run in a separate window to watch for bonding driver error messages.
152
153 It is critical that either the miimon or arp_interval and
154arp_ip_target parameters be specified, otherwise serious network
155degradation will occur during link failures. Very few devices do not
156support at least miimon, so there is really no reason not to use it.
157
158 Options with textual values will accept either the text name
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700159or, for backwards compatibility, the option value. E.g.,
160"mode=802.3ad" and "mode=4" set the same mode.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700161
162 The parameters are as follows:
163
Nicolas de Pesloüan1ba9ac72011-12-26 13:35:24 +0000164active_slave
165
166 Specifies the new active slave for modes that support it
167 (active-backup, balance-alb and balance-tlb). Possible values
168 are the name of any currently enslaved interface, or an empty
169 string. If a name is given, the slave and its link must be up in order
170 to be selected as the new active slave. If an empty string is
171 specified, the current active slave is cleared, and a new active
172 slave is selected automatically.
173
174 Note that this is only available through the sysfs interface. No module
175 parameter by this name exists.
176
177 The normal value of this option is the name of the currently
178 active slave, or the empty string if there is no active slave or
179 the current mode does not use an active slave.
180
Jay Vosburghfd989c82008-11-04 17:51:16 -0800181ad_select
182
183 Specifies the 802.3ad aggregation selection logic to use. The
184 possible values and their effects are:
185
186 stable or 0
187
188 The active aggregator is chosen by largest aggregate
189 bandwidth.
190
191 Reselection of the active aggregator occurs only when all
192 slaves of the active aggregator are down or the active
193 aggregator has no slaves.
194
195 This is the default value.
196
197 bandwidth or 1
198
199 The active aggregator is chosen by largest aggregate
200 bandwidth. Reselection occurs if:
201
202 - A slave is added to or removed from the bond
203
204 - Any slave's link state changes
205
206 - Any slave's 802.3ad association state changes
207
Matt LaPlante19f59462009-04-27 15:06:31 +0200208 - The bond's administrative state changes to up
Jay Vosburghfd989c82008-11-04 17:51:16 -0800209
210 count or 2
211
212 The active aggregator is chosen by the largest number of
213 ports (slaves). Reselection occurs as described under the
214 "bandwidth" setting, above.
215
216 The bandwidth and count selection policies permit failover of
217 802.3ad aggregations when partial failure of the active aggregator
218 occurs. This keeps the aggregator with the highest availability
219 (either in bandwidth or in number of ports) active at all times.
220
221 This option was added in bonding version 3.4.0.
222
Nicolas de Pesloüan025890b2011-08-06 07:06:39 +0000223all_slaves_active
224
225 Specifies that duplicate frames (received on inactive ports) should be
226 dropped (0) or delivered (1).
227
228 Normally, bonding will drop duplicate frames (received on inactive
229 ports), which is desirable for most users. But there are some times
230 it is nice to allow duplicate frames to be delivered.
231
232 The default value is 0 (drop duplicate frames received on inactive
233 ports).
234
Linus Torvalds1da177e2005-04-16 15:20:36 -0700235arp_interval
236
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700237 Specifies the ARP link monitoring frequency in milliseconds.
Jay Vosburghf5b2b962006-09-22 21:54:53 -0700238
239 The ARP monitor works by periodically checking the slave
240 devices to determine whether they have sent or received
241 traffic recently (the precise criteria depends upon the
242 bonding mode, and the state of the slave). Regular traffic is
243 generated via ARP probes issued for the addresses specified by
244 the arp_ip_target option.
245
246 This behavior can be modified by the arp_validate option,
247 below.
248
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700249 If ARP monitoring is used in an etherchannel compatible mode
250 (modes 0 and 2), the switch should be configured in a mode
251 that evenly distributes packets across all links. If the
252 switch is configured to distribute the packets in an XOR
Linus Torvalds1da177e2005-04-16 15:20:36 -0700253 fashion, all replies from the ARP targets will be received on
254 the same link which could cause the other team members to
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700255 fail. ARP monitoring should not be used in conjunction with
256 miimon. A value of 0 disables ARP monitoring. The default
Linus Torvalds1da177e2005-04-16 15:20:36 -0700257 value is 0.
258
259arp_ip_target
260
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700261 Specifies the IP addresses to use as ARP monitoring peers when
262 arp_interval is > 0. These are the targets of the ARP request
263 sent to determine the health of the link to the targets.
264 Specify these values in ddd.ddd.ddd.ddd format. Multiple IP
265 addresses must be separated by a comma. At least one IP
266 address must be given for ARP monitoring to function. The
267 maximum number of targets that can be specified is 16. The
268 default value is no IP addresses.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700269
Jay Vosburghf5b2b962006-09-22 21:54:53 -0700270arp_validate
271
272 Specifies whether or not ARP probes and replies should be
273 validated in the active-backup mode. This causes the ARP
274 monitor to examine the incoming ARP requests and replies, and
275 only consider a slave to be up if it is receiving the
276 appropriate ARP traffic.
277
278 Possible values are:
279
280 none or 0
281
282 No validation is performed. This is the default.
283
284 active or 1
285
286 Validation is performed only for the active slave.
287
288 backup or 2
289
290 Validation is performed only for backup slaves.
291
292 all or 3
293
294 Validation is performed for all slaves.
295
296 For the active slave, the validation checks ARP replies to
297 confirm that they were generated by an arp_ip_target. Since
298 backup slaves do not typically receive these replies, the
299 validation performed for backup slaves is on the ARP request
300 sent out via the active slave. It is possible that some
301 switch or network configurations may result in situations
302 wherein the backup slaves do not receive the ARP requests; in
303 such a situation, validation of backup slaves must be
304 disabled.
305
Veaceslav Falicod7d35c62013-06-24 11:49:33 +0200306 The validation of ARP requests on backup slaves is mainly
307 helping bonding to decide which slaves are more likely to
308 work in case of the active slave failure, it doesn't really
309 guarantee that the backup slave will work if it's selected
310 as the next active slave.
311
Jay Vosburghf5b2b962006-09-22 21:54:53 -0700312 This option is useful in network configurations in which
313 multiple bonding hosts are concurrently issuing ARPs to one or
314 more targets beyond a common switch. Should the link between
315 the switch and target fail (but not the switch itself), the
316 probe traffic generated by the multiple bonding instances will
317 fool the standard ARP monitor into considering the links as
318 still up. Use of the arp_validate option can resolve this, as
319 the ARP monitor will only consider ARP requests and replies
320 associated with its own instance of bonding.
321
322 This option was added in bonding version 3.1.0.
323
Veaceslav Falico8599b522013-06-24 11:49:34 +0200324arp_all_targets
325
326 Specifies the quantity of arp_ip_targets that must be reachable
327 in order for the ARP monitor to consider a slave as being up.
328 This option affects only active-backup mode for slaves with
329 arp_validation enabled.
330
331 Possible values are:
332
333 any or 0
334
335 consider the slave up only when any of the arp_ip_targets
336 is reachable
337
338 all or 1
339
340 consider the slave up only when all of the arp_ip_targets
341 are reachable
342
Linus Torvalds1da177e2005-04-16 15:20:36 -0700343downdelay
344
345 Specifies the time, in milliseconds, to wait before disabling
346 a slave after a link failure has been detected. This option
347 is only valid for the miimon link monitor. The downdelay
348 value should be a multiple of the miimon value; if not, it
349 will be rounded down to the nearest multiple. The default
350 value is 0.
351
Jay Vosburghdd957c52007-10-09 19:57:24 -0700352fail_over_mac
353
354 Specifies whether active-backup mode should set all slaves to
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700355 the same MAC address at enslavement (the traditional
356 behavior), or, when enabled, perform special handling of the
357 bond's MAC address in accordance with the selected policy.
Jay Vosburghdd957c52007-10-09 19:57:24 -0700358
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700359 Possible values are:
Jay Vosburghdd957c52007-10-09 19:57:24 -0700360
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700361 none or 0
Jay Vosburghdd957c52007-10-09 19:57:24 -0700362
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700363 This setting disables fail_over_mac, and causes
364 bonding to set all slaves of an active-backup bond to
365 the same MAC address at enslavement time. This is the
366 default.
Jay Vosburghdd957c52007-10-09 19:57:24 -0700367
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700368 active or 1
Jay Vosburghdd957c52007-10-09 19:57:24 -0700369
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700370 The "active" fail_over_mac policy indicates that the
371 MAC address of the bond should always be the MAC
372 address of the currently active slave. The MAC
373 address of the slaves is not changed; instead, the MAC
374 address of the bond changes during a failover.
375
376 This policy is useful for devices that cannot ever
377 alter their MAC address, or for devices that refuse
378 incoming broadcasts with their own source MAC (which
379 interferes with the ARP monitor).
380
381 The down side of this policy is that every device on
382 the network must be updated via gratuitous ARP,
383 vs. just updating a switch or set of switches (which
384 often takes place for any traffic, not just ARP
385 traffic, if the switch snoops incoming traffic to
386 update its tables) for the traditional method. If the
387 gratuitous ARP is lost, communication may be
388 disrupted.
389
Lucas De Marchi25985ed2011-03-30 22:57:33 -0300390 When this policy is used in conjunction with the mii
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700391 monitor, devices which assert link up prior to being
392 able to actually transmit and receive are particularly
Matt LaPlante19f59462009-04-27 15:06:31 +0200393 susceptible to loss of the gratuitous ARP, and an
Jay Vosburgh3915c1e82008-05-17 21:10:14 -0700394 appropriate updelay setting may be required.
395
396 follow or 2
397
398 The "follow" fail_over_mac policy causes the MAC
399 address of the bond to be selected normally (normally
400 the MAC address of the first slave added to the bond).
401 However, the second and subsequent slaves are not set
402 to this MAC address while they are in a backup role; a
403 slave is programmed with the bond's MAC address at
404 failover time (and the formerly active slave receives
405 the newly active slave's MAC address).
406
407 This policy is useful for multiport devices that
408 either become confused or incur a performance penalty
409 when multiple ports are programmed with the same MAC
410 address.
411
412
413 The default policy is none, unless the first slave cannot
414 change its MAC address, in which case the active policy is
415 selected by default.
416
417 This option may be modified via sysfs only when no slaves are
418 present in the bond.
419
420 This option was added in bonding version 3.2.0. The "follow"
421 policy was added in bonding version 3.3.0.
Jay Vosburghdd957c52007-10-09 19:57:24 -0700422
Linus Torvalds1da177e2005-04-16 15:20:36 -0700423lacp_rate
424
425 Option specifying the rate in which we'll ask our link partner
426 to transmit LACPDU packets in 802.3ad mode. Possible values
427 are:
428
429 slow or 0
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700430 Request partner to transmit LACPDUs every 30 seconds
Linus Torvalds1da177e2005-04-16 15:20:36 -0700431
432 fast or 1
433 Request partner to transmit LACPDUs every 1 second
434
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700435 The default is slow.
436
Linus Torvalds1da177e2005-04-16 15:20:36 -0700437max_bonds
438
439 Specifies the number of bonding devices to create for this
440 instance of the bonding driver. E.g., if max_bonds is 3, and
441 the bonding driver is not already loaded, then bond0, bond1
Jay Vosburghb8a97872008-06-13 18:12:04 -0700442 and bond2 will be created. The default value is 1. Specifying
443 a value of 0 will load bonding, but will not create any devices.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700444
445miimon
446
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700447 Specifies the MII link monitoring frequency in milliseconds.
448 This determines how often the link state of each slave is
449 inspected for link failures. A value of zero disables MII
450 link monitoring. A value of 100 is a good starting point.
451 The use_carrier option, below, affects how the link state is
Linus Torvalds1da177e2005-04-16 15:20:36 -0700452 determined. See the High Availability section for additional
453 information. The default value is 0.
454
Nicolas de Pesloüan025890b2011-08-06 07:06:39 +0000455min_links
456
457 Specifies the minimum number of links that must be active before
458 asserting carrier. It is similar to the Cisco EtherChannel min-links
459 feature. This allows setting the minimum number of member ports that
460 must be up (link-up state) before marking the bond device as up
461 (carrier on). This is useful for situations where higher level services
462 such as clustering want to ensure a minimum number of low bandwidth
463 links are active before switchover. This option only affect 802.3ad
464 mode.
465
466 The default value is 0. This will cause carrier to be asserted (for
467 802.3ad mode) whenever there is an active aggregator, regardless of the
468 number of available links in that aggregator. Note that, because an
469 aggregator cannot be active without at least one available link,
470 setting this option to 0 or to 1 has the exact same effect.
471
Linus Torvalds1da177e2005-04-16 15:20:36 -0700472mode
473
474 Specifies one of the bonding policies. The default is
475 balance-rr (round robin). Possible values are:
476
477 balance-rr or 0
478
479 Round-robin policy: Transmit packets in sequential
480 order from the first available slave through the
481 last. This mode provides load balancing and fault
482 tolerance.
483
484 active-backup or 1
485
486 Active-backup policy: Only one slave in the bond is
487 active. A different slave becomes active if, and only
488 if, the active slave fails. The bond's MAC address is
489 externally visible on only one port (network adapter)
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700490 to avoid confusing the switch.
491
492 In bonding version 2.6.2 or later, when a failover
493 occurs in active-backup mode, bonding will issue one
494 or more gratuitous ARPs on the newly active slave.
Auke Kok6224e012006-06-08 11:15:35 -0700495 One gratuitous ARP is issued for the bonding master
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700496 interface and each VLAN interfaces configured above
497 it, provided that the interface has at least one IP
498 address configured. Gratuitous ARPs issued for VLAN
499 interfaces are tagged with the appropriate VLAN id.
500
501 This mode provides fault tolerance. The primary
502 option, documented below, affects the behavior of this
503 mode.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700504
505 balance-xor or 2
506
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700507 XOR policy: Transmit based on the selected transmit
508 hash policy. The default policy is a simple [(source
509 MAC address XOR'd with destination MAC address) modulo
510 slave count]. Alternate transmit policies may be
511 selected via the xmit_hash_policy option, described
512 below.
513
514 This mode provides load balancing and fault tolerance.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700515
516 broadcast or 3
517
518 Broadcast policy: transmits everything on all slave
519 interfaces. This mode provides fault tolerance.
520
521 802.3ad or 4
522
523 IEEE 802.3ad Dynamic link aggregation. Creates
524 aggregation groups that share the same speed and
525 duplex settings. Utilizes all slaves in the active
526 aggregator according to the 802.3ad specification.
527
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700528 Slave selection for outgoing traffic is done according
529 to the transmit hash policy, which may be changed from
530 the default simple XOR policy via the xmit_hash_policy
531 option, documented below. Note that not all transmit
532 policies may be 802.3ad compliant, particularly in
533 regards to the packet mis-ordering requirements of
534 section 43.2.4 of the 802.3ad standard. Differing
535 peer implementations will have varying tolerances for
536 noncompliance.
537
538 Prerequisites:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700539
540 1. Ethtool support in the base drivers for retrieving
541 the speed and duplex of each slave.
542
543 2. A switch that supports IEEE 802.3ad Dynamic link
544 aggregation.
545
546 Most switches will require some type of configuration
547 to enable 802.3ad mode.
548
549 balance-tlb or 5
550
551 Adaptive transmit load balancing: channel bonding that
552 does not require any special switch support. The
553 outgoing traffic is distributed according to the
554 current load (computed relative to the speed) on each
555 slave. Incoming traffic is received by the current
556 slave. If the receiving slave fails, another slave
557 takes over the MAC address of the failed receiving
558 slave.
559
560 Prerequisite:
561
562 Ethtool support in the base drivers for retrieving the
563 speed of each slave.
564
565 balance-alb or 6
566
567 Adaptive load balancing: includes balance-tlb plus
568 receive load balancing (rlb) for IPV4 traffic, and
569 does not require any special switch support. The
570 receive load balancing is achieved by ARP negotiation.
571 The bonding driver intercepts the ARP Replies sent by
572 the local system on their way out and overwrites the
573 source hardware address with the unique hardware
574 address of one of the slaves in the bond such that
575 different peers use different hardware addresses for
576 the server.
577
578 Receive traffic from connections created by the server
579 is also balanced. When the local system sends an ARP
580 Request the bonding driver copies and saves the peer's
581 IP information from the ARP packet. When the ARP
582 Reply arrives from the peer, its hardware address is
583 retrieved and the bonding driver initiates an ARP
584 reply to this peer assigning it to one of the slaves
585 in the bond. A problematic outcome of using ARP
586 negotiation for balancing is that each time that an
587 ARP request is broadcast it uses the hardware address
588 of the bond. Hence, peers learn the hardware address
589 of the bond and the balancing of receive traffic
590 collapses to the current slave. This is handled by
591 sending updates (ARP Replies) to all the peers with
592 their individually assigned hardware address such that
593 the traffic is redistributed. Receive traffic is also
594 redistributed when a new slave is added to the bond
595 and when an inactive slave is re-activated. The
596 receive load is distributed sequentially (round robin)
597 among the group of highest speed slaves in the bond.
598
599 When a link is reconnected or a new slave joins the
600 bond the receive traffic is redistributed among all
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700601 active slaves in the bond by initiating ARP Replies
Auke Kok6224e012006-06-08 11:15:35 -0700602 with the selected MAC address to each of the
Linus Torvalds1da177e2005-04-16 15:20:36 -0700603 clients. The updelay parameter (detailed below) must
604 be set to a value equal or greater than the switch's
605 forwarding delay so that the ARP Replies sent to the
606 peers will not be blocked by the switch.
607
608 Prerequisites:
609
610 1. Ethtool support in the base drivers for retrieving
611 the speed of each slave.
612
613 2. Base driver support for setting the hardware
614 address of a device while it is open. This is
615 required so that there will always be one slave in the
616 team using the bond hardware address (the
617 curr_active_slave) while having a unique hardware
618 address for each slave in the bond. If the
619 curr_active_slave fails its hardware address is
620 swapped with the new curr_active_slave that was
621 chosen.
622
Jay Vosburghb59f9f72008-06-13 18:12:03 -0700623num_grat_arp
Brian Haley305d5522008-11-04 17:51:14 -0800624num_unsol_na
625
Ben Hutchingsad246c92011-04-26 15:25:52 +0000626 Specify the number of peer notifications (gratuitous ARPs and
627 unsolicited IPv6 Neighbor Advertisements) to be issued after a
628 failover event. As soon as the link is up on the new slave
629 (possibly immediately) a peer notification is sent on the
630 bonding device and each VLAN sub-device. This is repeated at
631 each link monitor interval (arp_interval or miimon, whichever
632 is active) if the number is greater than 1.
Brian Haley305d5522008-11-04 17:51:14 -0800633
Ben Hutchingsad246c92011-04-26 15:25:52 +0000634 The valid range is 0 - 255; the default value is 1. These options
635 affect only the active-backup mode. These options were added for
636 bonding versions 3.3.0 and 3.4.0 respectively.
637
Jesper Juhl8fb4e132011-08-01 17:59:44 -0700638 From Linux 3.0 and bonding version 3.7.1, these notifications
Ben Hutchingsad246c92011-04-26 15:25:52 +0000639 are generated by the ipv4 and ipv6 code and the numbers of
640 repetitions cannot be set independently.
Brian Haley305d5522008-11-04 17:51:14 -0800641
Nikolay Aleksandrov12465fb2013-11-05 13:51:42 +0100642packets_per_slave
643
644 Specify the number of packets to transmit through a slave before
645 moving to the next one. When set to 0 then a slave is chosen at
646 random.
647
648 The valid range is 0 - 65535; the default value is 1. This option
649 has effect only in balance-rr mode.
650
Linus Torvalds1da177e2005-04-16 15:20:36 -0700651primary
652
653 A string (eth0, eth2, etc) specifying which slave is the
654 primary device. The specified device will always be the
655 active slave while it is available. Only when the primary is
656 off-line will alternate devices be used. This is useful when
657 one slave is preferred over another, e.g., when one slave has
658 higher throughput than another.
659
660 The primary option is only valid for active-backup mode.
661
Jiri Pirkoa5499522009-09-25 03:28:09 +0000662primary_reselect
663
664 Specifies the reselection policy for the primary slave. This
665 affects how the primary slave is chosen to become the active slave
666 when failure of the active slave or recovery of the primary slave
667 occurs. This option is designed to prevent flip-flopping between
668 the primary slave and other slaves. Possible values are:
669
670 always or 0 (default)
671
672 The primary slave becomes the active slave whenever it
673 comes back up.
674
675 better or 1
676
677 The primary slave becomes the active slave when it comes
678 back up, if the speed and duplex of the primary slave is
679 better than the speed and duplex of the current active
680 slave.
681
682 failure or 2
683
684 The primary slave becomes the active slave only if the
685 current active slave fails and the primary slave is up.
686
687 The primary_reselect setting is ignored in two cases:
688
689 If no slaves are active, the first slave to recover is
690 made the active slave.
691
692 When initially enslaved, the primary slave is always made
693 the active slave.
694
695 Changing the primary_reselect policy via sysfs will cause an
696 immediate selection of the best active slave according to the new
697 policy. This may or may not result in a change of the active
698 slave, depending upon the circumstances.
699
700 This option was added for bonding version 3.6.0.
701
Linus Torvalds1da177e2005-04-16 15:20:36 -0700702updelay
703
704 Specifies the time, in milliseconds, to wait before enabling a
705 slave after a link recovery has been detected. This option is
706 only valid for the miimon link monitor. The updelay value
707 should be a multiple of the miimon value; if not, it will be
708 rounded down to the nearest multiple. The default value is 0.
709
710use_carrier
711
712 Specifies whether or not miimon should use MII or ETHTOOL
713 ioctls vs. netif_carrier_ok() to determine the link
714 status. The MII or ETHTOOL ioctls are less efficient and
715 utilize a deprecated calling sequence within the kernel. The
716 netif_carrier_ok() relies on the device driver to maintain its
717 state with netif_carrier_on/off; at this writing, most, but
718 not all, device drivers support this facility.
719
720 If bonding insists that the link is up when it should not be,
721 it may be that your network device driver does not support
722 netif_carrier_on/off. The default state for netif_carrier is
723 "carrier on," so if a driver does not support netif_carrier,
724 it will appear as if the link is always up. In this case,
725 setting use_carrier to 0 will cause bonding to revert to the
726 MII / ETHTOOL ioctl method to determine the link state.
727
728 A value of 1 enables the use of netif_carrier_ok(), a value of
729 0 will use the deprecated MII / ETHTOOL ioctls. The default
730 value is 1.
731
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700732xmit_hash_policy
733
734 Selects the transmit hash policy to use for slave selection in
735 balance-xor and 802.3ad modes. Possible values are:
736
737 layer2
738
739 Uses XOR of hardware MAC addresses to generate the
740 hash. The formula is
741
742 (source MAC XOR destination MAC) modulo slave count
743
744 This algorithm will place all traffic to a particular
745 network peer on the same slave.
746
747 This algorithm is 802.3ad compliant.
748
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800749 layer2+3
750
751 This policy uses a combination of layer2 and layer3
752 protocol information to generate the hash.
753
754 Uses XOR of hardware MAC addresses and IP addresses to
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200755 generate the hash. The formula is
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800756
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200757 hash = source MAC XOR destination MAC
758 hash = hash XOR source IP XOR destination IP
759 hash = hash XOR (hash RSHIFT 16)
760 hash = hash XOR (hash RSHIFT 8)
761 And then hash is reduced modulo slave count.
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800762
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200763 If the protocol is IPv6 then the source and destination
764 addresses are first hashed using ipv6_addr_hash.
John Eaglesham6b923cb2012-08-21 20:43:35 +0000765
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800766 This algorithm will place all traffic to a particular
767 network peer on the same slave. For non-IP traffic,
768 the formula is the same as for the layer2 transmit
769 hash policy.
770
771 This policy is intended to provide a more balanced
772 distribution of traffic than layer2 alone, especially
773 in environments where a layer3 gateway device is
774 required to reach most destinations.
775
Matt LaPlanted9195882008-07-25 19:45:33 -0700776 This algorithm is 802.3ad compliant.
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800777
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700778 layer3+4
779
780 This policy uses upper layer protocol information,
781 when available, to generate the hash. This allows for
782 traffic to a particular network peer to span multiple
783 slaves, although a single connection will not span
784 multiple slaves.
785
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200786 The formula for unfragmented TCP and UDP packets is
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700787
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200788 hash = source port, destination port (as in the header)
789 hash = hash XOR source IP XOR destination IP
790 hash = hash XOR (hash RSHIFT 16)
791 hash = hash XOR (hash RSHIFT 8)
792 And then hash is reduced modulo slave count.
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700793
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200794 If the protocol is IPv6 then the source and destination
795 addresses are first hashed using ipv6_addr_hash.
John Eaglesham6b923cb2012-08-21 20:43:35 +0000796
797 For fragmented TCP or UDP packets and all other IPv4 and
798 IPv6 protocol traffic, the source and destination port
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700799 information is omitted. For non-IP traffic, the
800 formula is the same as for the layer2 transmit hash
801 policy.
802
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700803 This algorithm is not fully 802.3ad compliant. A
804 single TCP or UDP conversation containing both
805 fragmented and unfragmented packets will see packets
806 striped across two interfaces. This may result in out
807 of order delivery. Most traffic types will not meet
808 this criteria, as TCP rarely fragments traffic, and
809 most UDP traffic is not involved in extended
810 conversations. Other implementations of 802.3ad may
811 or may not tolerate this noncompliance.
812
Nikolay Aleksandrov7a6afab2013-10-02 13:39:26 +0200813 encap2+3
814
815 This policy uses the same formula as layer2+3 but it
816 relies on skb_flow_dissect to obtain the header fields
817 which might result in the use of inner headers if an
818 encapsulation protocol is used. For example this will
819 improve the performance for tunnel users because the
820 packets will be distributed according to the encapsulated
821 flows.
822
823 encap3+4
824
825 This policy uses the same formula as layer3+4 but it
826 relies on skb_flow_dissect to obtain the header fields
827 which might result in the use of inner headers if an
828 encapsulation protocol is used. For example this will
829 improve the performance for tunnel users because the
830 packets will be distributed according to the encapsulated
831 flows.
832
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700833 The default value is layer2. This option was added in bonding
Jay Vosburgh6f6652b2007-12-06 23:40:34 -0800834 version 2.6.3. In earlier versions of bonding, this parameter
835 does not exist, and the layer2 policy is the only policy. The
836 layer2+3 value was added for bonding version 3.2.2.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700837
Flavio Leitnerc2952c32010-10-05 14:23:59 +0000838resend_igmp
839
840 Specifies the number of IGMP membership reports to be issued after
841 a failover event. One membership report is issued immediately after
842 the failover, subsequent packets are sent in each 200ms interval.
843
Flavio Leitner94265cf2011-05-25 08:38:58 +0000844 The valid range is 0 - 255; the default value is 1. A value of 0
845 prevents the IGMP membership report from being issued in response
846 to the failover event.
847
848 This option is useful for bonding modes balance-rr (0), active-backup
849 (1), balance-tlb (5) and balance-alb (6), in which a failover can
850 switch the IGMP traffic from one slave to another. Therefore a fresh
851 IGMP report must be issued to cause the switch to forward the incoming
852 IGMP traffic over the newly selected slave.
853
854 This option was added for bonding version 3.7.0.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700855
8563. Configuring Bonding Devices
857==============================
858
Auke Kok6224e012006-06-08 11:15:35 -0700859 You can configure bonding using either your distro's network
Cong Wangb1098bb2013-05-27 15:49:16 +0000860initialization scripts, or manually using either iproute2 or the
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +0000861sysfs interface. Distros generally use one of three packages for the
862network initialization scripts: initscripts, sysconfig or interfaces.
863Recent versions of these packages have support for bonding, while older
Auke Kok6224e012006-06-08 11:15:35 -0700864versions do not.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700865
866 We will first describe the options for configuring bonding for
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +0000867distros using versions of initscripts, sysconfig and interfaces with full
868or partial support for bonding, then provide information on enabling
Linus Torvalds1da177e2005-04-16 15:20:36 -0700869bonding without support from the network initialization scripts (i.e.,
870older versions of initscripts or sysconfig).
871
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +0000872 If you're unsure whether your distro uses sysconfig,
873initscripts or interfaces, or don't know if it's new enough, have no fear.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700874Determining this is fairly straightforward.
875
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +0000876 First, look for a file called interfaces in /etc/network directory.
877If this file is present in your system, then your system use interfaces. See
878Configuration with Interfaces Support.
879
880 Else, issue the command:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700881
882$ rpm -qf /sbin/ifup
883
884 It will respond with a line of text starting with either
885"initscripts" or "sysconfig," followed by some numbers. This is the
886package that provides your network initialization scripts.
887
888 Next, to determine if your installation supports bonding,
889issue the command:
890
891$ grep ifenslave /sbin/ifup
892
893 If this returns any matches, then your initscripts or
894sysconfig has support for bonding.
895
Auke Kok6224e012006-06-08 11:15:35 -07008963.1 Configuration with Sysconfig Support
Linus Torvalds1da177e2005-04-16 15:20:36 -0700897----------------------------------------
898
899 This section applies to distros using a version of sysconfig
900with bonding support, for example, SuSE Linux Enterprise Server 9.
901
902 SuSE SLES 9's networking configuration system does support
903bonding, however, at this writing, the YaST system configuration
Auke Kok6224e012006-06-08 11:15:35 -0700904front end does not provide any means to work with bonding devices.
Linus Torvalds1da177e2005-04-16 15:20:36 -0700905Bonding devices can be managed by hand, however, as follows.
906
907 First, if they have not already been configured, configure the
908slave devices. On SLES 9, this is most easily done by running the
909yast2 sysconfig configuration utility. The goal is for to create an
910ifcfg-id file for each slave device. The simplest way to accomplish
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700911this is to configure the devices for DHCP (this is only to get the
912file ifcfg-id file created; see below for some issues with DHCP). The
913name of the configuration file for each device will be of the form:
Linus Torvalds1da177e2005-04-16 15:20:36 -0700914
915ifcfg-id-xx:xx:xx:xx:xx:xx
916
917 Where the "xx" portion will be replaced with the digits from
918the device's permanent MAC address.
919
920 Once the set of ifcfg-id-xx:xx:xx:xx:xx:xx files has been
921created, it is necessary to edit the configuration files for the slave
922devices (the MAC addresses correspond to those of the slave devices).
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700923Before editing, the file will contain multiple lines, and will look
Linus Torvalds1da177e2005-04-16 15:20:36 -0700924something like this:
925
926BOOTPROTO='dhcp'
927STARTMODE='on'
928USERCTL='no'
929UNIQUE='XNzu.WeZGOGF+4wE'
930_nm_name='bus-pci-0001:61:01.0'
931
932 Change the BOOTPROTO and STARTMODE lines to the following:
933
934BOOTPROTO='none'
935STARTMODE='off'
936
937 Do not alter the UNIQUE or _nm_name lines. Remove any other
938lines (USERCTL, etc).
939
940 Once the ifcfg-id-xx:xx:xx:xx:xx:xx files have been modified,
941it's time to create the configuration file for the bonding device
942itself. This file is named ifcfg-bondX, where X is the number of the
943bonding device to create, starting at 0. The first such file is
944ifcfg-bond0, the second is ifcfg-bond1, and so on. The sysconfig
945network configuration system will correctly start multiple instances
946of bonding.
947
948 The contents of the ifcfg-bondX file is as follows:
949
950BOOTPROTO="static"
951BROADCAST="10.0.2.255"
952IPADDR="10.0.2.10"
953NETMASK="255.255.0.0"
954NETWORK="10.0.2.0"
955REMOTE_IPADDR=""
956STARTMODE="onboot"
957BONDING_MASTER="yes"
958BONDING_MODULE_OPTS="mode=active-backup miimon=100"
959BONDING_SLAVE0="eth0"
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700960BONDING_SLAVE1="bus-pci-0000:06:08.1"
Linus Torvalds1da177e2005-04-16 15:20:36 -0700961
962 Replace the sample BROADCAST, IPADDR, NETMASK and NETWORK
963values with the appropriate values for your network.
964
Linus Torvalds1da177e2005-04-16 15:20:36 -0700965 The STARTMODE specifies when the device is brought online.
966The possible values are:
967
968 onboot: The device is started at boot time. If you're not
969 sure, this is probably what you want.
970
971 manual: The device is started only when ifup is called
972 manually. Bonding devices may be configured this
973 way if you do not wish them to start automatically
974 at boot for some reason.
975
976 hotplug: The device is started by a hotplug event. This is not
977 a valid choice for a bonding device.
978
979 off or ignore: The device configuration is ignored.
980
981 The line BONDING_MASTER='yes' indicates that the device is a
982bonding master device. The only useful value is "yes."
983
984 The contents of BONDING_MODULE_OPTS are supplied to the
985instance of the bonding module for this device. Specify the options
986for the bonding mode, link monitoring, and so on here. Do not include
987the max_bonds bonding parameter; this will confuse the configuration
988system if you have multiple bonding devices.
989
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700990 Finally, supply one BONDING_SLAVEn="slave device" for each
991slave. where "n" is an increasing value, one for each slave. The
992"slave device" is either an interface name, e.g., "eth0", or a device
993specifier for the network device. The interface name is easier to
994find, but the ethN names are subject to change at boot time if, e.g.,
995a device early in the sequence has failed. The device specifiers
996(bus-pci-0000:06:08.1 in the example above) specify the physical
997network device, and will not change unless the device's bus location
998changes (for example, it is moved from one PCI slot to another). The
999example above uses one of each type for demonstration purposes; most
1000configurations will choose one or the other for all slave devices.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001001
1002 When all configuration files have been modified or created,
1003networking must be restarted for the configuration changes to take
1004effect. This can be accomplished via the following:
1005
1006# /etc/init.d/network restart
1007
1008 Note that the network control script (/sbin/ifdown) will
1009remove the bonding module as part of the network shutdown processing,
1010so it is not necessary to remove the module by hand if, e.g., the
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001011module parameters have changed.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001012
1013 Also, at this writing, YaST/YaST2 will not manage bonding
1014devices (they do not show bonding interfaces on its list of network
1015devices). It is necessary to edit the configuration file by hand to
1016change the bonding configuration.
1017
1018 Additional general options and details of the ifcfg file
1019format can be found in an example ifcfg template file:
1020
1021/etc/sysconfig/network/ifcfg.template
1022
1023 Note that the template does not document the various BONDING_
1024settings described above, but does describe many of the other options.
1025
Auke Kok6224e012006-06-08 11:15:35 -070010263.1.1 Using DHCP with Sysconfig
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001027-------------------------------
1028
1029 Under sysconfig, configuring a device with BOOTPROTO='dhcp'
1030will cause it to query DHCP for its IP address information. At this
1031writing, this does not function for bonding devices; the scripts
1032attempt to obtain the device address from DHCP prior to adding any of
1033the slave devices. Without active slaves, the DHCP requests are not
1034sent to the network.
1035
Auke Kok6224e012006-06-08 11:15:35 -070010363.1.2 Configuring Multiple Bonds with Sysconfig
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001037-----------------------------------------------
1038
1039 The sysconfig network initialization system is capable of
1040handling multiple bonding devices. All that is necessary is for each
1041bonding instance to have an appropriately configured ifcfg-bondX file
1042(as described above). Do not specify the "max_bonds" parameter to any
1043instance of bonding, as this will confuse sysconfig. If you require
1044multiple bonding devices with identical parameters, create multiple
1045ifcfg-bondX files.
1046
1047 Because the sysconfig scripts supply the bonding module
1048options in the ifcfg-bondX file, it is not necessary to add them to
Lucas De Marchi970e2482012-03-30 13:37:16 -07001049the system /etc/modules.d/*.conf configuration files.
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001050
Auke Kok6224e012006-06-08 11:15:35 -070010513.2 Configuration with Initscripts Support
Linus Torvalds1da177e2005-04-16 15:20:36 -07001052------------------------------------------
1053
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001054 This section applies to distros using a recent version of
1055initscripts with bonding support, for example, Red Hat Enterprise Linux
1056version 3 or later, Fedora, etc. On these systems, the network
1057initialization scripts have knowledge of bonding, and can be configured to
1058control bonding devices. Note that older versions of the initscripts
1059package have lower levels of support for bonding; this will be noted where
1060applicable.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001061
1062 These distros will not automatically load the network adapter
1063driver unless the ethX device is configured with an IP address.
1064Because of this constraint, users must manually configure a
1065network-script file for all physical adapters that will be members of
1066a bondX link. Network script files are located in the directory:
1067
1068/etc/sysconfig/network-scripts
1069
1070 The file name must be prefixed with "ifcfg-eth" and suffixed
1071with the adapter's physical adapter number. For example, the script
1072for eth0 would be named /etc/sysconfig/network-scripts/ifcfg-eth0.
1073Place the following text in the file:
1074
1075DEVICE=eth0
1076USERCTL=no
1077ONBOOT=yes
1078MASTER=bond0
1079SLAVE=yes
1080BOOTPROTO=none
1081
1082 The DEVICE= line will be different for every ethX device and
1083must correspond with the name of the file, i.e., ifcfg-eth1 must have
1084a device line of DEVICE=eth1. The setting of the MASTER= line will
1085also depend on the final bonding interface name chosen for your bond.
1086As with other network devices, these typically start at 0, and go up
1087one for each device, i.e., the first bonding instance is bond0, the
1088second is bond1, and so on.
1089
1090 Next, create a bond network script. The file name for this
1091script will be /etc/sysconfig/network-scripts/ifcfg-bondX where X is
1092the number of the bond. For bond0 the file is named "ifcfg-bond0",
1093for bond1 it is named "ifcfg-bond1", and so on. Within that file,
1094place the following text:
1095
1096DEVICE=bond0
1097IPADDR=192.168.1.1
1098NETMASK=255.255.255.0
1099NETWORK=192.168.1.0
1100BROADCAST=192.168.1.255
1101ONBOOT=yes
1102BOOTPROTO=none
1103USERCTL=no
1104
1105 Be sure to change the networking specific lines (IPADDR,
1106NETMASK, NETWORK and BROADCAST) to match your network configuration.
1107
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001108 For later versions of initscripts, such as that found with Fedora
Andy Gospodarek3f8b4b12008-10-22 11:19:48 +000011097 (or later) and Red Hat Enterprise Linux version 5 (or later), it is possible,
1110and, indeed, preferable, to specify the bonding options in the ifcfg-bond0
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001111file, e.g. a line of the format:
1112
Andy Gospodarek3f8b4b12008-10-22 11:19:48 +00001113BONDING_OPTS="mode=active-backup arp_interval=60 arp_ip_target=192.168.1.254"
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001114
1115 will configure the bond with the specified options. The options
1116specified in BONDING_OPTS are identical to the bonding module parameters
Andy Gospodarek3f8b4b12008-10-22 11:19:48 +00001117except for the arp_ip_target field when using versions of initscripts older
1118than and 8.57 (Fedora 8) and 8.45.19 (Red Hat Enterprise Linux 5.2). When
1119using older versions each target should be included as a separate option and
1120should be preceded by a '+' to indicate it should be added to the list of
1121queried targets, e.g.,
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001122
1123 arp_ip_target=+192.168.1.1 arp_ip_target=+192.168.1.2
1124
1125 is the proper syntax to specify multiple targets. When specifying
Lucas De Marchi970e2482012-03-30 13:37:16 -07001126options via BONDING_OPTS, it is not necessary to edit /etc/modprobe.d/*.conf.
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001127
Andy Gospodarek3f8b4b12008-10-22 11:19:48 +00001128 For even older versions of initscripts that do not support
Lucas De Marchi970e2482012-03-30 13:37:16 -07001129BONDING_OPTS, it is necessary to edit /etc/modprobe.d/*.conf, depending upon
1130your distro) to load the bonding module with your desired options when the
1131bond0 interface is brought up. The following lines in /etc/modprobe.d/*.conf
1132will load the bonding module, and select its options:
Linus Torvalds1da177e2005-04-16 15:20:36 -07001133
1134alias bond0 bonding
1135options bond0 mode=balance-alb miimon=100
1136
1137 Replace the sample parameters with the appropriate set of
1138options for your configuration.
1139
1140 Finally run "/etc/rc.d/init.d/network restart" as root. This
1141will restart the networking subsystem and your bond link should be now
1142up and running.
1143
Auke Kok6224e012006-06-08 11:15:35 -070011443.2.1 Using DHCP with Initscripts
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001145---------------------------------
1146
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001147 Recent versions of initscripts (the versions supplied with Fedora
1148Core 3 and Red Hat Enterprise Linux 4, or later versions, are reported to
1149work) have support for assigning IP information to bonding devices via
1150DHCP.
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001151
1152 To configure bonding for DHCP, configure it as described
1153above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp"
1154and add a line consisting of "TYPE=Bonding". Note that the TYPE value
1155is case sensitive.
1156
Auke Kok6224e012006-06-08 11:15:35 -070011573.2.2 Configuring Multiple Bonds with Initscripts
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001158-------------------------------------------------
1159
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001160 Initscripts packages that are included with Fedora 7 and Red Hat
1161Enterprise Linux 5 support multiple bonding interfaces by simply
1162specifying the appropriate BONDING_OPTS= in ifcfg-bondX where X is the
1163number of the bond. This support requires sysfs support in the kernel,
1164and a bonding driver of version 3.0.0 or later. Other configurations may
1165not support this method for specifying multiple bonding interfaces; for
1166those instances, see the "Configuring Multiple Bonds Manually" section,
1167below.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001168
Cong Wangb1098bb2013-05-27 15:49:16 +000011693.3 Configuring Bonding Manually with iproute2
Auke Kok6224e012006-06-08 11:15:35 -07001170-----------------------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07001171
1172 This section applies to distros whose network initialization
1173scripts (the sysconfig or initscripts package) do not have specific
1174knowledge of bonding. One such distro is SuSE Linux Enterprise Server
1175version 8.
1176
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001177 The general method for these systems is to place the bonding
Lucas De Marchi970e2482012-03-30 13:37:16 -07001178module parameters into a config file in /etc/modprobe.d/ (as
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001179appropriate for the installed distro), then add modprobe and/or
Cong Wangb1098bb2013-05-27 15:49:16 +00001180`ip link` commands to the system's global init script. The name of
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001181the global init script differs; for sysconfig, it is
Linus Torvalds1da177e2005-04-16 15:20:36 -07001182/etc/init.d/boot.local and for initscripts it is /etc/rc.d/rc.local.
1183
1184 For example, if you wanted to make a simple bond of two e100
1185devices (presumed to be eth0 and eth1), and have it persist across
1186reboots, edit the appropriate file (/etc/init.d/boot.local or
1187/etc/rc.d/rc.local), and add the following:
1188
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001189modprobe bonding mode=balance-alb miimon=100
Linus Torvalds1da177e2005-04-16 15:20:36 -07001190modprobe e100
1191ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
Cong Wangb1098bb2013-05-27 15:49:16 +00001192ip link set eth0 master bond0
1193ip link set eth1 master bond0
Linus Torvalds1da177e2005-04-16 15:20:36 -07001194
1195 Replace the example bonding module parameters and bond0
1196network configuration (IP address, netmask, etc) with the appropriate
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001197values for your configuration.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001198
1199 Unfortunately, this method will not provide support for the
1200ifup and ifdown scripts on the bond devices. To reload the bonding
1201configuration, it is necessary to run the initialization script, e.g.,
1202
1203# /etc/init.d/boot.local
1204
1205 or
1206
1207# /etc/rc.d/rc.local
1208
1209 It may be desirable in such a case to create a separate script
1210which only initializes the bonding configuration, then call that
1211separate script from within boot.local. This allows for bonding to be
1212enabled without re-running the entire global init script.
1213
1214 To shut down the bonding devices, it is necessary to first
1215mark the bonding device itself as being down, then remove the
1216appropriate device driver modules. For our example above, you can do
1217the following:
1218
1219# ifconfig bond0 down
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001220# rmmod bonding
Linus Torvalds1da177e2005-04-16 15:20:36 -07001221# rmmod e100
1222
1223 Again, for convenience, it may be desirable to create a script
1224with these commands.
1225
1226
Jay Vosburgh00354cf2005-07-21 12:18:02 -070012273.3.1 Configuring Multiple Bonds Manually
1228-----------------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07001229
1230 This section contains information on configuring multiple
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001231bonding devices with differing options for those systems whose network
1232initialization scripts lack support for configuring multiple bonds.
1233
1234 If you require multiple bonding devices, but all with the same
1235options, you may wish to use the "max_bonds" module parameter,
1236documented above.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001237
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001238 To create multiple bonding devices with differing options, it is
Rick Jonesf8b72d32012-07-20 10:51:37 +00001239preferable to use bonding parameters exported by sysfs, documented in the
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001240section below.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001241
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001242 For versions of bonding without sysfs support, the only means to
1243provide multiple instances of bonding with differing options is to load
1244the bonding driver multiple times. Note that current versions of the
1245sysconfig network initialization scripts handle this automatically; if
1246your distro uses these scripts, no special action is needed. See the
1247section Configuring Bonding Devices, above, if you're not sure about your
1248network initialization scripts.
1249
1250 To load multiple instances of the module, it is necessary to
1251specify a different name for each instance (the module loading system
1252requires that every loaded module, even multiple instances of the same
1253module, have a unique name). This is accomplished by supplying multiple
Lucas De Marchi970e2482012-03-30 13:37:16 -07001254sets of bonding options in /etc/modprobe.d/*.conf, for example:
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001255
1256alias bond0 bonding
1257options bond0 -o bond0 mode=balance-rr miimon=100
1258
1259alias bond1 bonding
1260options bond1 -o bond1 mode=balance-alb miimon=50
1261
1262 will load the bonding module two times. The first instance is
1263named "bond0" and creates the bond0 device in balance-rr mode with an
1264miimon of 100. The second instance is named "bond1" and creates the
1265bond1 device in balance-alb mode with an miimon of 50.
1266
1267 In some circumstances (typically with older distributions),
1268the above does not work, and the second bonding instance never sees
1269its options. In that case, the second options line can be substituted
1270as follows:
1271
1272install bond1 /sbin/modprobe --ignore-install bonding -o bond1 \
1273 mode=balance-alb miimon=50
1274
1275 This may be repeated any number of times, specifying a new and
1276unique name in place of bond1 for each subsequent instance.
1277
1278 It has been observed that some Red Hat supplied kernels are unable
1279to rename modules at load time (the "-o bond1" part). Attempts to pass
1280that option to modprobe will produce an "Operation not permitted" error.
1281This has been reported on some Fedora Core kernels, and has been seen on
1282RHEL 4 as well. On kernels exhibiting this problem, it will be impossible
1283to configure multiple bonds with differing parameters (as they are older
1284kernels, and also lack sysfs support).
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001285
Auke Kok6224e012006-06-08 11:15:35 -070012863.4 Configuring Bonding Manually via Sysfs
1287------------------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07001288
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08001289 Starting with version 3.0.0, Channel Bonding may be configured
Auke Kok6224e012006-06-08 11:15:35 -07001290via the sysfs interface. This interface allows dynamic configuration
1291of all bonds in the system without unloading the module. It also
1292allows for adding and removing bonds at runtime. Ifenslave is no
1293longer required, though it is still supported.
1294
1295 Use of the sysfs interface allows you to use multiple bonds
1296with different configurations without having to reload the module.
1297It also allows you to use multiple, differently configured bonds when
1298bonding is compiled into the kernel.
1299
1300 You must have the sysfs filesystem mounted to configure
1301bonding this way. The examples in this document assume that you
1302are using the standard mount point for sysfs, e.g. /sys. If your
1303sysfs filesystem is mounted elsewhere, you will need to adjust the
1304example paths accordingly.
1305
1306Creating and Destroying Bonds
1307-----------------------------
1308To add a new bond foo:
1309# echo +foo > /sys/class/net/bonding_masters
1310
1311To remove an existing bond bar:
1312# echo -bar > /sys/class/net/bonding_masters
1313
1314To show all existing bonds:
1315# cat /sys/class/net/bonding_masters
1316
1317NOTE: due to 4K size limitation of sysfs files, this list may be
1318truncated if you have more than a few hundred bonds. This is unlikely
1319to occur under normal operating conditions.
1320
1321Adding and Removing Slaves
1322--------------------------
1323 Interfaces may be enslaved to a bond using the file
1324/sys/class/net/<bond>/bonding/slaves. The semantics for this file
1325are the same as for the bonding_masters file.
1326
1327To enslave interface eth0 to bond bond0:
1328# ifconfig bond0 up
1329# echo +eth0 > /sys/class/net/bond0/bonding/slaves
1330
1331To free slave eth0 from bond bond0:
1332# echo -eth0 > /sys/class/net/bond0/bonding/slaves
1333
Auke Kok6224e012006-06-08 11:15:35 -07001334 When an interface is enslaved to a bond, symlinks between the
1335two are created in the sysfs filesystem. In this case, you would get
1336/sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and
1337/sys/class/net/eth0/master pointing to /sys/class/net/bond0.
1338
1339 This means that you can tell quickly whether or not an
1340interface is enslaved by looking for the master symlink. Thus:
1341# echo -eth0 > /sys/class/net/eth0/master/bonding/slaves
1342will free eth0 from whatever bond it is enslaved to, regardless of
1343the name of the bond interface.
1344
1345Changing a Bond's Configuration
1346-------------------------------
1347 Each bond may be configured individually by manipulating the
1348files located in /sys/class/net/<bond name>/bonding
1349
1350 The names of these files correspond directly with the command-
Paolo Ornati670e9f32006-10-03 22:57:56 +02001351line parameters described elsewhere in this file, and, with the
Auke Kok6224e012006-06-08 11:15:35 -07001352exception of arp_ip_target, they accept the same values. To see the
1353current setting, simply cat the appropriate file.
1354
1355 A few examples will be given here; for specific usage
1356guidelines for each parameter, see the appropriate section in this
1357document.
1358
1359To configure bond0 for balance-alb mode:
1360# ifconfig bond0 down
1361# echo 6 > /sys/class/net/bond0/bonding/mode
1362 - or -
1363# echo balance-alb > /sys/class/net/bond0/bonding/mode
1364 NOTE: The bond interface must be down before the mode can be
1365changed.
1366
1367To enable MII monitoring on bond0 with a 1 second interval:
1368# echo 1000 > /sys/class/net/bond0/bonding/miimon
1369 NOTE: If ARP monitoring is enabled, it will disabled when MII
1370monitoring is enabled, and vice-versa.
1371
1372To add ARP targets:
1373# echo +192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1374# echo +192.168.0.101 > /sys/class/net/bond0/bonding/arp_ip_target
Brian Haley5a31bec2009-04-13 00:11:30 -07001375 NOTE: up to 16 target addresses may be specified.
Auke Kok6224e012006-06-08 11:15:35 -07001376
1377To remove an ARP target:
1378# echo -192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1379
Neil Horman7eacd032013-09-13 11:05:33 -04001380To configure the interval between learning packet transmits:
1381# echo 12 > /sys/class/net/bond0/bonding/lp_interval
1382 NOTE: the lp_inteval is the number of seconds between instances where
1383the bonding driver sends learning packets to each slaves peer switch. The
1384default interval is 1 second.
1385
Auke Kok6224e012006-06-08 11:15:35 -07001386Example Configuration
1387---------------------
1388 We begin with the same example that is shown in section 3.3,
1389executed with sysfs, and without using ifenslave.
1390
1391 To make a simple bond of two e100 devices (presumed to be eth0
1392and eth1), and have it persist across reboots, edit the appropriate
1393file (/etc/init.d/boot.local or /etc/rc.d/rc.local), and add the
1394following:
1395
1396modprobe bonding
1397modprobe e100
1398echo balance-alb > /sys/class/net/bond0/bonding/mode
1399ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
1400echo 100 > /sys/class/net/bond0/bonding/miimon
1401echo +eth0 > /sys/class/net/bond0/bonding/slaves
1402echo +eth1 > /sys/class/net/bond0/bonding/slaves
1403
1404 To add a second bond, with two e1000 interfaces in
1405active-backup mode, using ARP monitoring, add the following lines to
1406your init script:
1407
1408modprobe e1000
1409echo +bond1 > /sys/class/net/bonding_masters
1410echo active-backup > /sys/class/net/bond1/bonding/mode
1411ifconfig bond1 192.168.2.1 netmask 255.255.255.0 up
1412echo +192.168.2.100 /sys/class/net/bond1/bonding/arp_ip_target
1413echo 2000 > /sys/class/net/bond1/bonding/arp_interval
1414echo +eth2 > /sys/class/net/bond1/bonding/slaves
1415echo +eth3 > /sys/class/net/bond1/bonding/slaves
1416
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +000014173.5 Configuration with Interfaces Support
1418-----------------------------------------
1419
1420 This section applies to distros which use /etc/network/interfaces file
1421to describe network interface configuration, most notably Debian and it's
1422derivatives.
1423
1424 The ifup and ifdown commands on Debian don't support bonding out of
1425the box. The ifenslave-2.6 package should be installed to provide bonding
1426support. Once installed, this package will provide bond-* options to be used
1427into /etc/network/interfaces.
1428
1429 Note that ifenslave-2.6 package will load the bonding module and use
1430the ifenslave command when appropriate.
1431
1432Example Configurations
1433----------------------
1434
1435In /etc/network/interfaces, the following stanza will configure bond0, in
1436active-backup mode, with eth0 and eth1 as slaves.
1437
1438auto bond0
1439iface bond0 inet dhcp
1440 bond-slaves eth0 eth1
1441 bond-mode active-backup
1442 bond-miimon 100
1443 bond-primary eth0 eth1
1444
1445If the above configuration doesn't work, you might have a system using
1446upstart for system startup. This is most notably true for recent
1447Ubuntu versions. The following stanza in /etc/network/interfaces will
1448produce the same result on those systems.
1449
1450auto bond0
1451iface bond0 inet dhcp
1452 bond-slaves none
1453 bond-mode active-backup
1454 bond-miimon 100
1455
1456auto eth0
1457iface eth0 inet manual
1458 bond-master bond0
1459 bond-primary eth0 eth1
1460
1461auto eth1
1462iface eth1 inet manual
1463 bond-master bond0
1464 bond-primary eth0 eth1
1465
1466For a full list of bond-* supported options in /etc/network/interfaces and some
1467more advanced examples tailored to you particular distros, see the files in
1468/usr/share/doc/ifenslave-2.6.
1469
14703.6 Overriding Configuration for Special Cases
Andy Gospodarekbb1d9122010-06-02 08:40:18 +00001471----------------------------------------------
Nicolas de Pesloüande221bd2011-01-24 13:21:37 +00001472
Andy Gospodarekbb1d9122010-06-02 08:40:18 +00001473When using the bonding driver, the physical port which transmits a frame is
1474typically selected by the bonding driver, and is not relevant to the user or
1475system administrator. The output port is simply selected using the policies of
1476the selected bonding mode. On occasion however, it is helpful to direct certain
1477classes of traffic to certain physical interfaces on output to implement
1478slightly more complex policies. For example, to reach a web server over a
1479bonded interface in which eth0 connects to a private network, while eth1
1480connects via a public network, it may be desirous to bias the bond to send said
1481traffic over eth0 first, using eth1 only as a fall back, while all other traffic
1482can safely be sent over either interface. Such configurations may be achieved
1483using the traffic control utilities inherent in linux.
Auke Kok6224e012006-06-08 11:15:35 -07001484
Andy Gospodarekbb1d9122010-06-02 08:40:18 +00001485By default the bonding driver is multiqueue aware and 16 queues are created
1486when the driver initializes (see Documentation/networking/multiqueue.txt
1487for details). If more or less queues are desired the module parameter
1488tx_queues can be used to change this value. There is no sysfs parameter
1489available as the allocation is done at module init time.
1490
1491The output of the file /proc/net/bonding/bondX has changed so the output Queue
1492ID is now printed for each slave:
1493
1494Bonding Mode: fault-tolerance (active-backup)
1495Primary Slave: None
1496Currently Active Slave: eth0
1497MII Status: up
1498MII Polling Interval (ms): 0
1499Up Delay (ms): 0
1500Down Delay (ms): 0
1501
1502Slave Interface: eth0
1503MII Status: up
1504Link Failure Count: 0
1505Permanent HW addr: 00:1a:a0:12:8f:cb
1506Slave queue ID: 0
1507
1508Slave Interface: eth1
1509MII Status: up
1510Link Failure Count: 0
1511Permanent HW addr: 00:1a:a0:12:8f:cc
1512Slave queue ID: 2
1513
1514The queue_id for a slave can be set using the command:
1515
1516# echo "eth1:2" > /sys/class/net/bond0/bonding/queue_id
1517
1518Any interface that needs a queue_id set should set it with multiple calls
1519like the one above until proper priorities are set for all interfaces. On
1520distributions that allow configuration via initscripts, multiple 'queue_id'
1521arguments can be added to BONDING_OPTS to set all needed slave queues.
1522
1523These queue id's can be used in conjunction with the tc utility to configure
1524a multiqueue qdisc and filters to bias certain traffic to transmit on certain
1525slave devices. For instance, say we wanted, in the above configuration to
1526force all traffic bound to 192.168.1.100 to use eth1 in the bond as its output
1527device. The following commands would accomplish this:
1528
1529# tc qdisc add dev bond0 handle 1 root multiq
1530
1531# tc filter add dev bond0 protocol ip parent 1: prio 1 u32 match ip dst \
1532 192.168.1.100 action skbedit queue_mapping 2
1533
1534These commands tell the kernel to attach a multiqueue queue discipline to the
1535bond0 interface and filter traffic enqueued to it, such that packets with a dst
1536ip of 192.168.1.100 have their output queue mapping value overwritten to 2.
1537This value is then passed into the driver, causing the normal output path
1538selection policy to be overridden, selecting instead qid 2, which maps to eth1.
1539
1540Note that qid values begin at 1. Qid 0 is reserved to initiate to the driver
1541that normal output policy selection should take place. One benefit to simply
1542leaving the qid for a slave to 0 is the multiqueue awareness in the bonding
1543driver that is now present. This awareness allows tc filters to be placed on
1544slave devices as well as bond devices and the bonding driver will simply act as
1545a pass-through for selecting output queues on the slave device rather than
1546output port selection.
1547
1548This feature first appeared in bonding driver version 3.7.0 and support for
1549output slave selection was limited to round-robin and active-backup modes.
1550
15514 Querying Bonding Configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -07001552=================================
1553
Auke Kok6224e012006-06-08 11:15:35 -070015544.1 Bonding Configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -07001555-------------------------
1556
1557 Each bonding device has a read-only file residing in the
1558/proc/net/bonding directory. The file contents include information
1559about the bonding configuration, options and state of each slave.
1560
1561 For example, the contents of /proc/net/bonding/bond0 after the
1562driver is loaded with parameters of mode=0 and miimon=1000 is
1563generally as follows:
1564
1565 Ethernet Channel Bonding Driver: 2.6.1 (October 29, 2004)
1566 Bonding Mode: load balancing (round-robin)
1567 Currently Active Slave: eth0
1568 MII Status: up
1569 MII Polling Interval (ms): 1000
1570 Up Delay (ms): 0
1571 Down Delay (ms): 0
1572
1573 Slave Interface: eth1
1574 MII Status: up
1575 Link Failure Count: 1
1576
1577 Slave Interface: eth0
1578 MII Status: up
1579 Link Failure Count: 1
1580
1581 The precise format and contents will change depending upon the
1582bonding configuration, state, and version of the bonding driver.
1583
Auke Kok6224e012006-06-08 11:15:35 -070015844.2 Network configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -07001585-------------------------
1586
1587 The network configuration can be inspected using the ifconfig
1588command. Bonding devices will have the MASTER flag set; Bonding slave
1589devices will have the SLAVE flag set. The ifconfig output does not
1590contain information on which slaves are associated with which masters.
1591
1592 In the example below, the bond0 interface is the master
1593(MASTER) while eth0 and eth1 are slaves (SLAVE). Notice all slaves of
1594bond0 have the same MAC address (HWaddr) as bond0 for all modes except
1595TLB and ALB that require a unique MAC address for each slave.
1596
1597# /sbin/ifconfig
1598bond0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1599 inet addr:XXX.XXX.XXX.YYY Bcast:XXX.XXX.XXX.255 Mask:255.255.252.0
1600 UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
1601 RX packets:7224794 errors:0 dropped:0 overruns:0 frame:0
1602 TX packets:3286647 errors:1 dropped:0 overruns:1 carrier:0
1603 collisions:0 txqueuelen:0
1604
1605eth0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
Linus Torvalds1da177e2005-04-16 15:20:36 -07001606 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1607 RX packets:3573025 errors:0 dropped:0 overruns:0 frame:0
1608 TX packets:1643167 errors:1 dropped:0 overruns:1 carrier:0
1609 collisions:0 txqueuelen:100
1610 Interrupt:10 Base address:0x1080
1611
1612eth1 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
Linus Torvalds1da177e2005-04-16 15:20:36 -07001613 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1614 RX packets:3651769 errors:0 dropped:0 overruns:0 frame:0
1615 TX packets:1643480 errors:0 dropped:0 overruns:0 carrier:0
1616 collisions:0 txqueuelen:100
1617 Interrupt:9 Base address:0x1400
1618
Auke Kok6224e012006-06-08 11:15:35 -070016195. Switch Configuration
Linus Torvalds1da177e2005-04-16 15:20:36 -07001620=======================
1621
1622 For this section, "switch" refers to whatever system the
1623bonded devices are directly connected to (i.e., where the other end of
1624the cable plugs into). This may be an actual dedicated switch device,
1625or it may be another regular system (e.g., another computer running
1626Linux),
1627
1628 The active-backup, balance-tlb and balance-alb modes do not
1629require any specific configuration of the switch.
1630
1631 The 802.3ad mode requires that the switch have the appropriate
1632ports configured as an 802.3ad aggregation. The precise method used
1633to configure this varies from switch to switch, but, for example, a
1634Cisco 3550 series switch requires that the appropriate ports first be
1635grouped together in a single etherchannel instance, then that
1636etherchannel is set to mode "lacp" to enable 802.3ad (instead of
1637standard EtherChannel).
1638
1639 The balance-rr, balance-xor and broadcast modes generally
1640require that the switch have the appropriate ports grouped together.
1641The nomenclature for such a group differs between switches, it may be
1642called an "etherchannel" (as in the Cisco example, above), a "trunk
1643group" or some other similar variation. For these modes, each switch
1644will also have its own configuration options for the switch's transmit
1645policy to the bond. Typical choices include XOR of either the MAC or
1646IP addresses. The transmit policy of the two peers does not need to
1647match. For these three modes, the bonding mode really selects a
1648transmit policy for an EtherChannel group; all three will interoperate
1649with another EtherChannel group.
1650
1651
Auke Kok6224e012006-06-08 11:15:35 -070016526. 802.1q VLAN Support
Linus Torvalds1da177e2005-04-16 15:20:36 -07001653======================
1654
1655 It is possible to configure VLAN devices over a bond interface
1656using the 8021q driver. However, only packets coming from the 8021q
1657driver and passing through bonding will be tagged by default. Self
1658generated packets, for example, bonding's learning packets or ARP
1659packets generated by either ALB mode or the ARP monitor mechanism, are
1660tagged internally by bonding itself. As a result, bonding must
1661"learn" the VLAN IDs configured above it, and use those IDs to tag
1662self generated packets.
1663
1664 For reasons of simplicity, and to support the use of adapters
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001665that can do VLAN hardware acceleration offloading, the bonding
1666interface declares itself as fully hardware offloading capable, it gets
Linus Torvalds1da177e2005-04-16 15:20:36 -07001667the add_vid/kill_vid notifications to gather the necessary
1668information, and it propagates those actions to the slaves. In case
1669of mixed adapter types, hardware accelerated tagged packets that
1670should go through an adapter that is not offloading capable are
1671"un-accelerated" by the bonding driver so the VLAN tag sits in the
1672regular location.
1673
1674 VLAN interfaces *must* be added on top of a bonding interface
1675only after enslaving at least one slave. The bonding interface has a
1676hardware address of 00:00:00:00:00:00 until the first slave is added.
1677If the VLAN interface is created prior to the first enslavement, it
1678would pick up the all-zeroes hardware address. Once the first slave
1679is attached to the bond, the bond device itself will pick up the
1680slave's hardware address, which is then available for the VLAN device.
1681
1682 Also, be aware that a similar problem can occur if all slaves
1683are released from a bond that still has one or more VLAN interfaces on
1684top of it. When a new slave is added, the bonding interface will
1685obtain its hardware address from the first slave, which might not
1686match the hardware address of the VLAN interfaces (which was
1687ultimately copied from an earlier slave).
1688
1689 There are two methods to insure that the VLAN device operates
1690with the correct hardware address if all slaves are removed from a
1691bond interface:
1692
1693 1. Remove all VLAN interfaces then recreate them
1694
1695 2. Set the bonding interface's hardware address so that it
1696matches the hardware address of the VLAN interfaces.
1697
1698 Note that changing a VLAN interface's HW address would set the
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001699underlying device -- i.e. the bonding interface -- to promiscuous
Linus Torvalds1da177e2005-04-16 15:20:36 -07001700mode, which might not be what you want.
1701
1702
Auke Kok6224e012006-06-08 11:15:35 -070017037. Link Monitoring
Linus Torvalds1da177e2005-04-16 15:20:36 -07001704==================
1705
1706 The bonding driver at present supports two schemes for
1707monitoring a slave device's link state: the ARP monitor and the MII
1708monitor.
1709
1710 At the present time, due to implementation restrictions in the
1711bonding driver itself, it is not possible to enable both ARP and MII
1712monitoring simultaneously.
1713
Auke Kok6224e012006-06-08 11:15:35 -070017147.1 ARP Monitor Operation
Linus Torvalds1da177e2005-04-16 15:20:36 -07001715-------------------------
1716
1717 The ARP monitor operates as its name suggests: it sends ARP
1718queries to one or more designated peer systems on the network, and
1719uses the response as an indication that the link is operating. This
1720gives some assurance that traffic is actually flowing to and from one
1721or more peers on the local network.
1722
1723 The ARP monitor relies on the device driver itself to verify
1724that traffic is flowing. In particular, the driver must keep up to
1725date the last receive time, dev->last_rx, and transmit start time,
1726dev->trans_start. If these are not updated by the driver, then the
1727ARP monitor will immediately fail any slaves using that driver, and
1728those slaves will stay down. If networking monitoring (tcpdump, etc)
1729shows the ARP requests and replies on the network, then it may be that
1730your device driver is not updating last_rx and trans_start.
1731
Auke Kok6224e012006-06-08 11:15:35 -070017327.2 Configuring Multiple ARP Targets
Linus Torvalds1da177e2005-04-16 15:20:36 -07001733------------------------------------
1734
1735 While ARP monitoring can be done with just one target, it can
1736be useful in a High Availability setup to have several targets to
1737monitor. In the case of just one target, the target itself may go
1738down or have a problem making it unresponsive to ARP requests. Having
1739an additional target (or several) increases the reliability of the ARP
1740monitoring.
1741
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001742 Multiple ARP targets must be separated by commas as follows:
Linus Torvalds1da177e2005-04-16 15:20:36 -07001743
1744# example options for ARP monitoring with three targets
1745alias bond0 bonding
1746options bond0 arp_interval=60 arp_ip_target=192.168.0.1,192.168.0.3,192.168.0.9
1747
1748 For just a single target the options would resemble:
1749
1750# example options for ARP monitoring with one target
1751alias bond0 bonding
1752options bond0 arp_interval=60 arp_ip_target=192.168.0.100
1753
1754
Auke Kok6224e012006-06-08 11:15:35 -070017557.3 MII Monitor Operation
Linus Torvalds1da177e2005-04-16 15:20:36 -07001756-------------------------
1757
1758 The MII monitor monitors only the carrier state of the local
1759network interface. It accomplishes this in one of three ways: by
1760depending upon the device driver to maintain its carrier state, by
1761querying the device's MII registers, or by making an ethtool query to
1762the device.
1763
1764 If the use_carrier module parameter is 1 (the default value),
1765then the MII monitor will rely on the driver for carrier state
1766information (via the netif_carrier subsystem). As explained in the
1767use_carrier parameter information, above, if the MII monitor fails to
1768detect carrier loss on the device (e.g., when the cable is physically
1769disconnected), it may be that the driver does not support
1770netif_carrier.
1771
1772 If use_carrier is 0, then the MII monitor will first query the
1773device's (via ioctl) MII registers and check the link state. If that
1774request fails (not just that it returns carrier down), then the MII
1775monitor will make an ethtool ETHOOL_GLINK request to attempt to obtain
1776the same information. If both methods fail (i.e., the driver either
1777does not support or had some error in processing both the MII register
1778and ethtool requests), then the MII monitor will assume the link is
1779up.
1780
Auke Kok6224e012006-06-08 11:15:35 -070017818. Potential Sources of Trouble
Linus Torvalds1da177e2005-04-16 15:20:36 -07001782===============================
1783
Auke Kok6224e012006-06-08 11:15:35 -070017848.1 Adventures in Routing
Linus Torvalds1da177e2005-04-16 15:20:36 -07001785-------------------------
1786
1787 When bonding is configured, it is important that the slave
Auke Kok6224e012006-06-08 11:15:35 -07001788devices not have routes that supersede routes of the master (or,
Linus Torvalds1da177e2005-04-16 15:20:36 -07001789generally, not have routes at all). For example, suppose the bonding
1790device bond0 has two slaves, eth0 and eth1, and the routing table is
1791as follows:
1792
1793Kernel IP routing table
1794Destination Gateway Genmask Flags MSS Window irtt Iface
179510.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth0
179610.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth1
179710.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 bond0
1798127.0.0.0 0.0.0.0 255.0.0.0 U 40 0 0 lo
1799
1800 This routing configuration will likely still update the
1801receive/transmit times in the driver (needed by the ARP monitor), but
1802may bypass the bonding driver (because outgoing traffic to, in this
1803case, another host on network 10 would use eth0 or eth1 before bond0).
1804
1805 The ARP monitor (and ARP itself) may become confused by this
1806configuration, because ARP requests (generated by the ARP monitor)
1807will be sent on one interface (bond0), but the corresponding reply
1808will arrive on a different interface (eth0). This reply looks to ARP
1809as an unsolicited ARP reply (because ARP matches replies on an
1810interface basis), and is discarded. The MII monitor is not affected
1811by the state of the routing table.
1812
1813 The solution here is simply to insure that slaves do not have
1814routes of their own, and if for some reason they must, those routes do
Auke Kok6224e012006-06-08 11:15:35 -07001815not supersede routes of their master. This should generally be the
Linus Torvalds1da177e2005-04-16 15:20:36 -07001816case, but unusual configurations or errant manual or automatic static
1817route additions may cause trouble.
1818
Auke Kok6224e012006-06-08 11:15:35 -070018198.2 Ethernet Device Renaming
Linus Torvalds1da177e2005-04-16 15:20:36 -07001820----------------------------
1821
1822 On systems with network configuration scripts that do not
1823associate physical devices directly with network interface names (so
1824that the same physical device always has the same "ethX" name), it may
Lucas De Marchi970e2482012-03-30 13:37:16 -07001825be necessary to add some special logic to config files in
1826/etc/modprobe.d/.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001827
1828 For example, given a modules.conf containing the following:
1829
1830alias bond0 bonding
1831options bond0 mode=some-mode miimon=50
1832alias eth0 tg3
1833alias eth1 tg3
1834alias eth2 e1000
1835alias eth3 e1000
1836
1837 If neither eth0 and eth1 are slaves to bond0, then when the
1838bond0 interface comes up, the devices may end up reordered. This
1839happens because bonding is loaded first, then its slave device's
1840drivers are loaded next. Since no other drivers have been loaded,
1841when the e1000 driver loads, it will receive eth0 and eth1 for its
1842devices, but the bonding configuration tries to enslave eth2 and eth3
1843(which may later be assigned to the tg3 devices).
1844
1845 Adding the following:
1846
1847add above bonding e1000 tg3
1848
1849 causes modprobe to load e1000 then tg3, in that order, when
1850bonding is loaded. This command is fully documented in the
1851modules.conf manual page.
1852
Lucas De Marchi970e2482012-03-30 13:37:16 -07001853 On systems utilizing modprobe an equivalent problem can occur.
1854In this case, the following can be added to config files in
1855/etc/modprobe.d/ as:
Linus Torvalds1da177e2005-04-16 15:20:36 -07001856
Lucas De Marchi78286cd2012-03-30 13:37:20 -07001857softdep bonding pre: tg3 e1000
Linus Torvalds1da177e2005-04-16 15:20:36 -07001858
Lucas De Marchi970e2482012-03-30 13:37:16 -07001859 This will load tg3 and e1000 modules before loading the bonding one.
1860Full documentation on this can be found in the modprobe.d and modprobe
1861manual pages.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001862
Auke Kok6224e012006-06-08 11:15:35 -070018638.3. Painfully Slow Or No Failed Link Detection By Miimon
Linus Torvalds1da177e2005-04-16 15:20:36 -07001864---------------------------------------------------------
1865
1866 By default, bonding enables the use_carrier option, which
1867instructs bonding to trust the driver to maintain carrier state.
1868
1869 As discussed in the options section, above, some drivers do
1870not support the netif_carrier_on/_off link state tracking system.
1871With use_carrier enabled, bonding will always see these links as up,
1872regardless of their actual state.
1873
1874 Additionally, other drivers do support netif_carrier, but do
1875not maintain it in real time, e.g., only polling the link state at
1876some fixed interval. In this case, miimon will detect failures, but
1877only after some long period of time has expired. If it appears that
1878miimon is very slow in detecting link failures, try specifying
1879use_carrier=0 to see if that improves the failure detection time. If
1880it does, then it may be that the driver checks the carrier state at a
1881fixed interval, but does not cache the MII register values (so the
1882use_carrier=0 method of querying the registers directly works). If
1883use_carrier=0 does not improve the failover, then the driver may cache
1884the registers, or the problem may be elsewhere.
1885
1886 Also, remember that miimon only checks for the device's
1887carrier state. It has no way to determine the state of devices on or
1888beyond other ports of a switch, or if a switch is refusing to pass
1889traffic while still maintaining carrier on.
1890
Auke Kok6224e012006-06-08 11:15:35 -070018919. SNMP agents
Linus Torvalds1da177e2005-04-16 15:20:36 -07001892===============
1893
1894 If running SNMP agents, the bonding driver should be loaded
1895before any network drivers participating in a bond. This requirement
Tobias Klauserd533f672005-09-10 00:26:46 -07001896is due to the interface index (ipAdEntIfIndex) being associated to
Linus Torvalds1da177e2005-04-16 15:20:36 -07001897the first interface found with a given IP address. That is, there is
1898only one ipAdEntIfIndex for each IP address. For example, if eth0 and
1899eth1 are slaves of bond0 and the driver for eth0 is loaded before the
1900bonding driver, the interface for the IP address will be associated
1901with the eth0 interface. This configuration is shown below, the IP
1902address 192.168.1.1 has an interface index of 2 which indexes to eth0
1903in the ifDescr table (ifDescr.2).
1904
1905 interfaces.ifTable.ifEntry.ifDescr.1 = lo
1906 interfaces.ifTable.ifEntry.ifDescr.2 = eth0
1907 interfaces.ifTable.ifEntry.ifDescr.3 = eth1
1908 interfaces.ifTable.ifEntry.ifDescr.4 = eth2
1909 interfaces.ifTable.ifEntry.ifDescr.5 = eth3
1910 interfaces.ifTable.ifEntry.ifDescr.6 = bond0
1911 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 5
1912 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
1913 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 4
1914 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
1915
1916 This problem is avoided by loading the bonding driver before
1917any network drivers participating in a bond. Below is an example of
1918loading the bonding driver first, the IP address 192.168.1.1 is
1919correctly associated with ifDescr.2.
1920
1921 interfaces.ifTable.ifEntry.ifDescr.1 = lo
1922 interfaces.ifTable.ifEntry.ifDescr.2 = bond0
1923 interfaces.ifTable.ifEntry.ifDescr.3 = eth0
1924 interfaces.ifTable.ifEntry.ifDescr.4 = eth1
1925 interfaces.ifTable.ifEntry.ifDescr.5 = eth2
1926 interfaces.ifTable.ifEntry.ifDescr.6 = eth3
1927 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 6
1928 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
1929 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 5
1930 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
1931
1932 While some distributions may not report the interface name in
1933ifDescr, the association between the IP address and IfIndex remains
1934and SNMP functions such as Interface_Scan_Next will report that
1935association.
1936
Auke Kok6224e012006-06-08 11:15:35 -0700193710. Promiscuous mode
Linus Torvalds1da177e2005-04-16 15:20:36 -07001938====================
1939
1940 When running network monitoring tools, e.g., tcpdump, it is
1941common to enable promiscuous mode on the device, so that all traffic
1942is seen (instead of seeing only traffic destined for the local host).
1943The bonding driver handles promiscuous mode changes to the bonding
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001944master device (e.g., bond0), and propagates the setting to the slave
Linus Torvalds1da177e2005-04-16 15:20:36 -07001945devices.
1946
1947 For the balance-rr, balance-xor, broadcast, and 802.3ad modes,
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001948the promiscuous mode setting is propagated to all slaves.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001949
1950 For the active-backup, balance-tlb and balance-alb modes, the
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001951promiscuous mode setting is propagated only to the active slave.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001952
1953 For balance-tlb mode, the active slave is the slave currently
1954receiving inbound traffic.
1955
1956 For balance-alb mode, the active slave is the slave used as a
1957"primary." This slave is used for mode-specific control traffic, for
1958sending to peers that are unassigned or if the load is unbalanced.
1959
1960 For the active-backup, balance-tlb and balance-alb modes, when
1961the active slave changes (e.g., due to a link failure), the
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001962promiscuous setting will be propagated to the new active slave.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001963
Auke Kok6224e012006-06-08 11:15:35 -0700196411. Configuring Bonding for High Availability
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001965=============================================
Linus Torvalds1da177e2005-04-16 15:20:36 -07001966
1967 High Availability refers to configurations that provide
1968maximum network availability by having redundant or backup devices,
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001969links or switches between the host and the rest of the world. The
1970goal is to provide the maximum availability of network connectivity
1971(i.e., the network always works), even though other configurations
1972could provide higher throughput.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001973
Auke Kok6224e012006-06-08 11:15:35 -0700197411.1 High Availability in a Single Switch Topology
Linus Torvalds1da177e2005-04-16 15:20:36 -07001975--------------------------------------------------
1976
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001977 If two hosts (or a host and a single switch) are directly
1978connected via multiple physical links, then there is no availability
1979penalty to optimizing for maximum bandwidth. In this case, there is
1980only one switch (or peer), so if it fails, there is no alternative
1981access to fail over to. Additionally, the bonding load balance modes
1982support link monitoring of their members, so if individual links fail,
1983the load will be rebalanced across the remaining devices.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001984
Rick Jonesf8b72d32012-07-20 10:51:37 +00001985 See Section 12, "Configuring Bonding for Maximum Throughput"
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001986for information on configuring bonding with one peer device.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001987
Auke Kok6224e012006-06-08 11:15:35 -0700198811.2 High Availability in a Multiple Switch Topology
Linus Torvalds1da177e2005-04-16 15:20:36 -07001989----------------------------------------------------
1990
1991 With multiple switches, the configuration of bonding and the
1992network changes dramatically. In multiple switch topologies, there is
Jay Vosburgh00354cf2005-07-21 12:18:02 -07001993a trade off between network availability and usable bandwidth.
Linus Torvalds1da177e2005-04-16 15:20:36 -07001994
1995 Below is a sample network, configured to maximize the
1996availability of the network:
1997
1998 | |
1999 |port3 port3|
2000 +-----+----+ +-----+----+
2001 | |port2 ISL port2| |
2002 | switch A +--------------------------+ switch B |
2003 | | | |
2004 +-----+----+ +-----++---+
2005 |port1 port1|
2006 | +-------+ |
2007 +-------------+ host1 +---------------+
2008 eth0 +-------+ eth1
2009
2010 In this configuration, there is a link between the two
2011switches (ISL, or inter switch link), and multiple ports connecting to
2012the outside world ("port3" on each switch). There is no technical
2013reason that this could not be extended to a third switch.
2014
Auke Kok6224e012006-06-08 11:15:35 -0700201511.2.1 HA Bonding Mode Selection for Multiple Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002016-------------------------------------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07002017
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002018 In a topology such as the example above, the active-backup and
2019broadcast modes are the only useful bonding modes when optimizing for
2020availability; the other modes require all links to terminate on the
2021same peer for them to behave rationally.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002022
2023active-backup: This is generally the preferred mode, particularly if
2024 the switches have an ISL and play together well. If the
2025 network configuration is such that one switch is specifically
2026 a backup switch (e.g., has lower capacity, higher cost, etc),
2027 then the primary option can be used to insure that the
2028 preferred link is always used when it is available.
2029
2030broadcast: This mode is really a special purpose mode, and is suitable
2031 only for very specific needs. For example, if the two
2032 switches are not connected (no ISL), and the networks beyond
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002033 them are totally independent. In this case, if it is
Linus Torvalds1da177e2005-04-16 15:20:36 -07002034 necessary for some specific one-way traffic to reach both
2035 independent networks, then the broadcast mode may be suitable.
2036
Auke Kok6224e012006-06-08 11:15:35 -0700203711.2.2 HA Link Monitoring Selection for Multiple Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002038----------------------------------------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07002039
2040 The choice of link monitoring ultimately depends upon your
2041switch. If the switch can reliably fail ports in response to other
2042failures, then either the MII or ARP monitors should work. For
2043example, in the above example, if the "port3" link fails at the remote
2044end, the MII monitor has no direct means to detect this. The ARP
2045monitor could be configured with a target at the remote end of port3,
2046thus detecting that failure without switch support.
2047
2048 In general, however, in a multiple switch topology, the ARP
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002049monitor can provide a higher level of reliability in detecting end to
2050end connectivity failures (which may be caused by the failure of any
2051individual component to pass traffic for any reason). Additionally,
2052the ARP monitor should be configured with multiple targets (at least
2053one for each switch in the network). This will insure that,
Linus Torvalds1da177e2005-04-16 15:20:36 -07002054regardless of which switch is active, the ARP monitor has a suitable
2055target to query.
2056
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002057 Note, also, that of late many switches now support a functionality
2058generally referred to as "trunk failover." This is a feature of the
2059switch that causes the link state of a particular switch port to be set
2060down (or up) when the state of another switch port goes down (or up).
Matt LaPlante19f59462009-04-27 15:06:31 +02002061Its purpose is to propagate link failures from logically "exterior" ports
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002062to the logically "interior" ports that bonding is able to monitor via
2063miimon. Availability and configuration for trunk failover varies by
2064switch, but this can be a viable alternative to the ARP monitor when using
2065suitable switches.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002066
Auke Kok6224e012006-06-08 11:15:35 -0700206712. Configuring Bonding for Maximum Throughput
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002068==============================================
Linus Torvalds1da177e2005-04-16 15:20:36 -07002069
Auke Kok6224e012006-06-08 11:15:35 -0700207012.1 Maximizing Throughput in a Single Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002071------------------------------------------------------
2072
2073 In a single switch configuration, the best method to maximize
2074throughput depends upon the application and network environment. The
2075various load balancing modes each have strengths and weaknesses in
2076different environments, as detailed below.
2077
2078 For this discussion, we will break down the topologies into
2079two categories. Depending upon the destination of most traffic, we
2080categorize them into either "gatewayed" or "local" configurations.
2081
2082 In a gatewayed configuration, the "switch" is acting primarily
2083as a router, and the majority of traffic passes through this router to
2084other networks. An example would be the following:
2085
2086
2087 +----------+ +----------+
2088 | |eth0 port1| | to other networks
2089 | Host A +---------------------+ router +------------------->
2090 | +---------------------+ | Hosts B and C are out
2091 | |eth1 port2| | here somewhere
2092 +----------+ +----------+
2093
2094 The router may be a dedicated router device, or another host
2095acting as a gateway. For our discussion, the important point is that
2096the majority of traffic from Host A will pass through the router to
2097some other network before reaching its final destination.
2098
2099 In a gatewayed network configuration, although Host A may
2100communicate with many other systems, all of its traffic will be sent
2101and received via one other peer on the local network, the router.
2102
2103 Note that the case of two systems connected directly via
2104multiple physical links is, for purposes of configuring bonding, the
2105same as a gatewayed configuration. In that case, it happens that all
2106traffic is destined for the "gateway" itself, not some other network
2107beyond the gateway.
2108
2109 In a local configuration, the "switch" is acting primarily as
2110a switch, and the majority of traffic passes through this switch to
2111reach other stations on the same network. An example would be the
2112following:
2113
2114 +----------+ +----------+ +--------+
2115 | |eth0 port1| +-------+ Host B |
2116 | Host A +------------+ switch |port3 +--------+
2117 | +------------+ | +--------+
2118 | |eth1 port2| +------------------+ Host C |
2119 +----------+ +----------+port4 +--------+
2120
2121
2122 Again, the switch may be a dedicated switch device, or another
2123host acting as a gateway. For our discussion, the important point is
2124that the majority of traffic from Host A is destined for other hosts
2125on the same local network (Hosts B and C in the above example).
2126
2127 In summary, in a gatewayed configuration, traffic to and from
2128the bonded device will be to the same MAC level peer on the network
2129(the gateway itself, i.e., the router), regardless of its final
2130destination. In a local configuration, traffic flows directly to and
2131from the final destinations, thus, each destination (Host B, Host C)
2132will be addressed directly by their individual MAC addresses.
2133
2134 This distinction between a gatewayed and a local network
2135configuration is important because many of the load balancing modes
2136available use the MAC addresses of the local network source and
2137destination to make load balancing decisions. The behavior of each
2138mode is described below.
2139
2140
Auke Kok6224e012006-06-08 11:15:35 -0700214112.1.1 MT Bonding Mode Selection for Single Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002142-----------------------------------------------------------
2143
2144 This configuration is the easiest to set up and to understand,
2145although you will have to decide which bonding mode best suits your
2146needs. The trade offs for each mode are detailed below:
2147
2148balance-rr: This mode is the only mode that will permit a single
2149 TCP/IP connection to stripe traffic across multiple
2150 interfaces. It is therefore the only mode that will allow a
2151 single TCP/IP stream to utilize more than one interface's
2152 worth of throughput. This comes at a cost, however: the
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002153 striping generally results in peer systems receiving packets out
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002154 of order, causing TCP/IP's congestion control system to kick
2155 in, often by retransmitting segments.
2156
2157 It is possible to adjust TCP/IP's congestion limits by
2158 altering the net.ipv4.tcp_reordering sysctl parameter. The
2159 usual default value is 3, and the maximum useful value is 127.
2160 For a four interface balance-rr bond, expect that a single
2161 TCP/IP stream will utilize no more than approximately 2.3
2162 interface's worth of throughput, even after adjusting
2163 tcp_reordering.
2164
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002165 Note that the fraction of packets that will be delivered out of
2166 order is highly variable, and is unlikely to be zero. The level
2167 of reordering depends upon a variety of factors, including the
2168 networking interfaces, the switch, and the topology of the
2169 configuration. Speaking in general terms, higher speed network
2170 cards produce more reordering (due to factors such as packet
2171 coalescing), and a "many to many" topology will reorder at a
2172 higher rate than a "many slow to one fast" configuration.
2173
2174 Many switches do not support any modes that stripe traffic
2175 (instead choosing a port based upon IP or MAC level addresses);
2176 for those devices, traffic for a particular connection flowing
2177 through the switch to a balance-rr bond will not utilize greater
2178 than one interface's worth of bandwidth.
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002179
2180 If you are utilizing protocols other than TCP/IP, UDP for
2181 example, and your application can tolerate out of order
2182 delivery, then this mode can allow for single stream datagram
2183 performance that scales near linearly as interfaces are added
2184 to the bond.
2185
2186 This mode requires the switch to have the appropriate ports
2187 configured for "etherchannel" or "trunking."
2188
2189active-backup: There is not much advantage in this network topology to
2190 the active-backup mode, as the inactive backup devices are all
2191 connected to the same peer as the primary. In this case, a
2192 load balancing mode (with link monitoring) will provide the
2193 same level of network availability, but with increased
2194 available bandwidth. On the plus side, active-backup mode
2195 does not require any configuration of the switch, so it may
2196 have value if the hardware available does not support any of
2197 the load balance modes.
2198
2199balance-xor: This mode will limit traffic such that packets destined
2200 for specific peers will always be sent over the same
2201 interface. Since the destination is determined by the MAC
2202 addresses involved, this mode works best in a "local" network
2203 configuration (as described above), with destinations all on
2204 the same local network. This mode is likely to be suboptimal
2205 if all your traffic is passed through a single router (i.e., a
2206 "gatewayed" network configuration, as described above).
2207
2208 As with balance-rr, the switch ports need to be configured for
2209 "etherchannel" or "trunking."
2210
2211broadcast: Like active-backup, there is not much advantage to this
2212 mode in this type of network topology.
2213
2214802.3ad: This mode can be a good choice for this type of network
2215 topology. The 802.3ad mode is an IEEE standard, so all peers
2216 that implement 802.3ad should interoperate well. The 802.3ad
2217 protocol includes automatic configuration of the aggregates,
2218 so minimal manual configuration of the switch is needed
2219 (typically only to designate that some set of devices is
2220 available for 802.3ad). The 802.3ad standard also mandates
2221 that frames be delivered in order (within certain limits), so
2222 in general single connections will not see misordering of
2223 packets. The 802.3ad mode does have some drawbacks: the
2224 standard mandates that all devices in the aggregate operate at
2225 the same speed and duplex. Also, as with all bonding load
2226 balance modes other than balance-rr, no single connection will
2227 be able to utilize more than a single interface's worth of
2228 bandwidth.
2229
2230 Additionally, the linux bonding 802.3ad implementation
2231 distributes traffic by peer (using an XOR of MAC addresses),
2232 so in a "gatewayed" configuration, all outgoing traffic will
2233 generally use the same device. Incoming traffic may also end
2234 up on a single device, but that is dependent upon the
2235 balancing policy of the peer's 8023.ad implementation. In a
2236 "local" configuration, traffic will be distributed across the
2237 devices in the bond.
2238
2239 Finally, the 802.3ad mode mandates the use of the MII monitor,
2240 therefore, the ARP monitor is not available in this mode.
2241
2242balance-tlb: The balance-tlb mode balances outgoing traffic by peer.
2243 Since the balancing is done according to MAC address, in a
2244 "gatewayed" configuration (as described above), this mode will
2245 send all traffic across a single device. However, in a
2246 "local" network configuration, this mode balances multiple
2247 local network peers across devices in a vaguely intelligent
2248 manner (not a simple XOR as in balance-xor or 802.3ad mode),
2249 so that mathematically unlucky MAC addresses (i.e., ones that
2250 XOR to the same value) will not all "bunch up" on a single
2251 interface.
2252
2253 Unlike 802.3ad, interfaces may be of differing speeds, and no
2254 special switch configuration is required. On the down side,
2255 in this mode all incoming traffic arrives over a single
2256 interface, this mode requires certain ethtool support in the
2257 network device driver of the slave interfaces, and the ARP
2258 monitor is not available.
2259
2260balance-alb: This mode is everything that balance-tlb is, and more.
2261 It has all of the features (and restrictions) of balance-tlb,
2262 and will also balance incoming traffic from local network
2263 peers (as described in the Bonding Module Options section,
2264 above).
2265
2266 The only additional down side to this mode is that the network
2267 device driver must support changing the hardware address while
2268 the device is open.
2269
Auke Kok6224e012006-06-08 11:15:35 -0700227012.1.2 MT Link Monitoring for Single Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002271----------------------------------------------------
2272
2273 The choice of link monitoring may largely depend upon which
2274mode you choose to use. The more advanced load balancing modes do not
2275support the use of the ARP monitor, and are thus restricted to using
2276the MII monitor (which does not provide as high a level of end to end
2277assurance as the ARP monitor).
2278
Auke Kok6224e012006-06-08 11:15:35 -0700227912.2 Maximum Throughput in a Multiple Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002280-----------------------------------------------------
2281
2282 Multiple switches may be utilized to optimize for throughput
2283when they are configured in parallel as part of an isolated network
2284between two or more systems, for example:
2285
2286 +-----------+
2287 | Host A |
2288 +-+---+---+-+
2289 | | |
2290 +--------+ | +---------+
2291 | | |
2292 +------+---+ +-----+----+ +-----+----+
2293 | Switch A | | Switch B | | Switch C |
2294 +------+---+ +-----+----+ +-----+----+
2295 | | |
2296 +--------+ | +---------+
2297 | | |
2298 +-+---+---+-+
2299 | Host B |
2300 +-----------+
2301
2302 In this configuration, the switches are isolated from one
2303another. One reason to employ a topology such as this is for an
2304isolated network with many hosts (a cluster configured for high
2305performance, for example), using multiple smaller switches can be more
2306cost effective than a single larger switch, e.g., on a network with 24
2307hosts, three 24 port switches can be significantly less expensive than
2308a single 72 port switch.
2309
2310 If access beyond the network is required, an individual host
2311can be equipped with an additional network device connected to an
2312external network; this host then additionally acts as a gateway.
2313
Auke Kok6224e012006-06-08 11:15:35 -0700231412.2.1 MT Bonding Mode Selection for Multiple Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002315-------------------------------------------------------------
2316
2317 In actual practice, the bonding mode typically employed in
2318configurations of this type is balance-rr. Historically, in this
2319network configuration, the usual caveats about out of order packet
2320delivery are mitigated by the use of network adapters that do not do
2321any kind of packet coalescing (via the use of NAPI, or because the
2322device itself does not generate interrupts until some number of
2323packets has arrived). When employed in this fashion, the balance-rr
2324mode allows individual connections between two hosts to effectively
2325utilize greater than one interface's bandwidth.
2326
Auke Kok6224e012006-06-08 11:15:35 -0700232712.2.2 MT Link Monitoring for Multiple Switch Topology
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002328------------------------------------------------------
2329
2330 Again, in actual practice, the MII monitor is most often used
2331in this configuration, as performance is given preference over
2332availability. The ARP monitor will function in this topology, but its
2333advantages over the MII monitor are mitigated by the volume of probes
2334needed as the number of systems involved grows (remember that each
2335host in the network is configured with bonding).
2336
Auke Kok6224e012006-06-08 11:15:35 -0700233713. Switch Behavior Issues
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002338==========================
2339
Auke Kok6224e012006-06-08 11:15:35 -0700234013.1 Link Establishment and Failover Delays
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002341-------------------------------------------
2342
2343 Some switches exhibit undesirable behavior with regard to the
2344timing of link up and down reporting by the switch.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002345
2346 First, when a link comes up, some switches may indicate that
2347the link is up (carrier available), but not pass traffic over the
2348interface for some period of time. This delay is typically due to
2349some type of autonegotiation or routing protocol, but may also occur
2350during switch initialization (e.g., during recovery after a switch
2351failure). If you find this to be a problem, specify an appropriate
2352value to the updelay bonding module option to delay the use of the
2353relevant interface(s).
2354
2355 Second, some switches may "bounce" the link state one or more
2356times while a link is changing state. This occurs most commonly while
2357the switch is initializing. Again, an appropriate updelay value may
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002358help.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002359
2360 Note that when a bonding interface has no active links, the
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002361driver will immediately reuse the first link that goes up, even if the
2362updelay parameter has been specified (the updelay is ignored in this
2363case). If there are slave interfaces waiting for the updelay timeout
2364to expire, the interface that first went into that state will be
2365immediately reused. This reduces down time of the network if the
2366value of updelay has been overestimated, and since this occurs only in
2367cases with no connectivity, there is no additional penalty for
2368ignoring the updelay.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002369
2370 In addition to the concerns about switch timings, if your
2371switches take a long time to go into backup mode, it may be desirable
2372to not activate a backup interface immediately after a link goes down.
2373Failover may be delayed via the downdelay bonding module option.
2374
Auke Kok6224e012006-06-08 11:15:35 -0700237513.2 Duplicated Incoming Packets
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002376--------------------------------
2377
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002378 NOTE: Starting with version 3.0.2, the bonding driver has logic to
2379suppress duplicate packets, which should largely eliminate this problem.
2380The following description is kept for reference.
2381
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002382 It is not uncommon to observe a short burst of duplicated
2383traffic when the bonding device is first used, or after it has been
2384idle for some period of time. This is most easily observed by issuing
2385a "ping" to some other host on the network, and noticing that the
2386output from ping flags duplicates (typically one per slave).
2387
2388 For example, on a bond in active-backup mode with five slaves
2389all connected to one switch, the output may appear as follows:
2390
2391# ping -n 10.0.4.2
2392PING 10.0.4.2 (10.0.4.2) from 10.0.3.10 : 56(84) bytes of data.
239364 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.7 ms
239464 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
239564 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
239664 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
239764 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
239864 bytes from 10.0.4.2: icmp_seq=2 ttl=64 time=0.216 ms
239964 bytes from 10.0.4.2: icmp_seq=3 ttl=64 time=0.267 ms
240064 bytes from 10.0.4.2: icmp_seq=4 ttl=64 time=0.222 ms
2401
2402 This is not due to an error in the bonding driver, rather, it
2403is a side effect of how many switches update their MAC forwarding
2404tables. Initially, the switch does not associate the MAC address in
2405the packet with a particular switch port, and so it may send the
2406traffic to all ports until its MAC forwarding table is updated. Since
2407the interfaces attached to the bond may occupy multiple ports on a
2408single switch, when the switch (temporarily) floods the traffic to all
2409ports, the bond device receives multiple copies of the same packet
2410(one per slave device).
2411
2412 The duplicated packet behavior is switch dependent, some
2413switches exhibit this, and some do not. On switches that display this
2414behavior, it can be induced by clearing the MAC forwarding table (on
2415most Cisco switches, the privileged command "clear mac address-table
2416dynamic" will accomplish this).
2417
Auke Kok6224e012006-06-08 11:15:35 -0700241814. Hardware Specific Considerations
Linus Torvalds1da177e2005-04-16 15:20:36 -07002419====================================
2420
2421 This section contains additional information for configuring
2422bonding on specific hardware platforms, or for interfacing bonding
2423with particular switches or other devices.
2424
Auke Kok6224e012006-06-08 11:15:35 -0700242514.1 IBM BladeCenter
Linus Torvalds1da177e2005-04-16 15:20:36 -07002426--------------------
2427
2428 This applies to the JS20 and similar systems.
2429
2430 On the JS20 blades, the bonding driver supports only
2431balance-rr, active-backup, balance-tlb and balance-alb modes. This is
2432largely due to the network topology inside the BladeCenter, detailed
2433below.
2434
2435JS20 network adapter information
2436--------------------------------
2437
2438 All JS20s come with two Broadcom Gigabit Ethernet ports
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002439integrated on the planar (that's "motherboard" in IBM-speak). In the
2440BladeCenter chassis, the eth0 port of all JS20 blades is hard wired to
2441I/O Module #1; similarly, all eth1 ports are wired to I/O Module #2.
2442An add-on Broadcom daughter card can be installed on a JS20 to provide
2443two more Gigabit Ethernet ports. These ports, eth2 and eth3, are
2444wired to I/O Modules 3 and 4, respectively.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002445
2446 Each I/O Module may contain either a switch or a passthrough
2447module (which allows ports to be directly connected to an external
2448switch). Some bonding modes require a specific BladeCenter internal
2449network topology in order to function; these are detailed below.
2450
2451 Additional BladeCenter-specific networking information can be
2452found in two IBM Redbooks (www.ibm.com/redbooks):
2453
2454"IBM eServer BladeCenter Networking Options"
2455"IBM eServer BladeCenter Layer 2-7 Network Switching"
2456
2457BladeCenter networking configuration
2458------------------------------------
2459
2460 Because a BladeCenter can be configured in a very large number
2461of ways, this discussion will be confined to describing basic
2462configurations.
2463
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002464 Normally, Ethernet Switch Modules (ESMs) are used in I/O
Linus Torvalds1da177e2005-04-16 15:20:36 -07002465modules 1 and 2. In this configuration, the eth0 and eth1 ports of a
2466JS20 will be connected to different internal switches (in the
2467respective I/O modules).
2468
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002469 A passthrough module (OPM or CPM, optical or copper,
2470passthrough module) connects the I/O module directly to an external
2471switch. By using PMs in I/O module #1 and #2, the eth0 and eth1
2472interfaces of a JS20 can be redirected to the outside world and
2473connected to a common external switch.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002474
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002475 Depending upon the mix of ESMs and PMs, the network will
2476appear to bonding as either a single switch topology (all PMs) or as a
2477multiple switch topology (one or more ESMs, zero or more PMs). It is
2478also possible to connect ESMs together, resulting in a configuration
2479much like the example in "High Availability in a Multiple Switch
2480Topology," above.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002481
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002482Requirements for specific modes
2483-------------------------------
Linus Torvalds1da177e2005-04-16 15:20:36 -07002484
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002485 The balance-rr mode requires the use of passthrough modules
2486for devices in the bond, all connected to an common external switch.
2487That switch must be configured for "etherchannel" or "trunking" on the
Linus Torvalds1da177e2005-04-16 15:20:36 -07002488appropriate ports, as is usual for balance-rr.
2489
2490 The balance-alb and balance-tlb modes will function with
2491either switch modules or passthrough modules (or a mix). The only
2492specific requirement for these modes is that all network interfaces
2493must be able to reach all destinations for traffic sent over the
2494bonding device (i.e., the network must converge at some point outside
2495the BladeCenter).
2496
2497 The active-backup mode has no additional requirements.
2498
2499Link monitoring issues
2500----------------------
2501
2502 When an Ethernet Switch Module is in place, only the ARP
2503monitor will reliably detect link loss to an external switch. This is
2504nothing unusual, but examination of the BladeCenter cabinet would
2505suggest that the "external" network ports are the ethernet ports for
2506the system, when it fact there is a switch between these "external"
2507ports and the devices on the JS20 system itself. The MII monitor is
2508only able to detect link failures between the ESM and the JS20 system.
2509
2510 When a passthrough module is in place, the MII monitor does
2511detect failures to the "external" port, which is then directly
2512connected to the JS20 system.
2513
2514Other concerns
2515--------------
2516
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002517 The Serial Over LAN (SoL) link is established over the primary
Linus Torvalds1da177e2005-04-16 15:20:36 -07002518ethernet (eth0) only, therefore, any loss of link to eth0 will result
2519in losing your SoL connection. It will not fail over with other
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002520network traffic, as the SoL system is beyond the control of the
2521bonding driver.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002522
2523 It may be desirable to disable spanning tree on the switch
2524(either the internal Ethernet Switch Module, or an external switch) to
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002525avoid fail-over delay issues when using bonding.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002526
2527
Auke Kok6224e012006-06-08 11:15:35 -0700252815. Frequently Asked Questions
Linus Torvalds1da177e2005-04-16 15:20:36 -07002529==============================
2530
25311. Is it SMP safe?
2532
2533 Yes. The old 2.0.xx channel bonding patch was not SMP safe.
2534The new driver was designed to be SMP safe from the start.
2535
25362. What type of cards will work with it?
2537
2538 Any Ethernet type cards (you can even mix cards - a Intel
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002539EtherExpress PRO/100 and a 3com 3c905b, for example). For most modes,
2540devices need not be of the same speed.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002541
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002542 Starting with version 3.2.1, bonding also supports Infiniband
2543slaves in active-backup mode.
2544
Linus Torvalds1da177e2005-04-16 15:20:36 -070025453. How many bonding devices can I have?
2546
2547 There is no limit.
2548
25494. How many slaves can a bonding device have?
2550
2551 This is limited only by the number of network interfaces Linux
2552supports and/or the number of network cards you can place in your
2553system.
2554
25555. What happens when a slave link dies?
2556
2557 If link monitoring is enabled, then the failing device will be
2558disabled. The active-backup mode will fail over to a backup link, and
2559other modes will ignore the failed link. The link will continue to be
2560monitored, and should it recover, it will rejoin the bond (in whatever
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002561manner is appropriate for the mode). See the sections on High
2562Availability and the documentation for each mode for additional
2563information.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002564
2565 Link monitoring can be enabled via either the miimon or
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002566arp_interval parameters (described in the module parameters section,
Linus Torvalds1da177e2005-04-16 15:20:36 -07002567above). In general, miimon monitors the carrier state as sensed by
2568the underlying network device, and the arp monitor (arp_interval)
2569monitors connectivity to another host on the local network.
2570
2571 If no link monitoring is configured, the bonding driver will
2572be unable to detect link failures, and will assume that all links are
2573always available. This will likely result in lost packets, and a
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002574resulting degradation of performance. The precise performance loss
Linus Torvalds1da177e2005-04-16 15:20:36 -07002575depends upon the bonding mode and network configuration.
2576
25776. Can bonding be used for High Availability?
2578
2579 Yes. See the section on High Availability for details.
2580
25817. Which switches/systems does it work with?
2582
2583 The full answer to this depends upon the desired mode.
2584
2585 In the basic balance modes (balance-rr and balance-xor), it
2586works with any system that supports etherchannel (also called
2587trunking). Most managed switches currently available have such
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002588support, and many unmanaged switches as well.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002589
2590 The advanced balance modes (balance-tlb and balance-alb) do
2591not have special switch requirements, but do need device drivers that
2592support specific features (described in the appropriate section under
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002593module parameters, above).
Linus Torvalds1da177e2005-04-16 15:20:36 -07002594
Auke Kok6224e012006-06-08 11:15:35 -07002595 In 802.3ad mode, it works with systems that support IEEE
Linus Torvalds1da177e2005-04-16 15:20:36 -07002596802.3ad Dynamic Link Aggregation. Most managed and many unmanaged
2597switches currently available support 802.3ad.
2598
2599 The active-backup mode should work with any Layer-II switch.
2600
26018. Where does a bonding device get its MAC address from?
2602
Jay Vosburgh9a6c6862007-11-13 20:25:48 -08002603 When using slave devices that have fixed MAC addresses, or when
2604the fail_over_mac option is enabled, the bonding device's MAC address is
2605the MAC address of the active slave.
2606
2607 For other configurations, if not explicitly configured (with
2608ifconfig or ip link), the MAC address of the bonding device is taken from
2609its first slave device. This MAC address is then passed to all following
2610slaves and remains persistent (even if the first slave is removed) until
2611the bonding device is brought down or reconfigured.
Linus Torvalds1da177e2005-04-16 15:20:36 -07002612
2613 If you wish to change the MAC address, you can set it with
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002614ifconfig or ip link:
Linus Torvalds1da177e2005-04-16 15:20:36 -07002615
2616# ifconfig bond0 hw ether 00:11:22:33:44:55
2617
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002618# ip link set bond0 address 66:77:88:99:aa:bb
2619
Linus Torvalds1da177e2005-04-16 15:20:36 -07002620 The MAC address can be also changed by bringing down/up the
2621device and then changing its slaves (or their order):
2622
2623# ifconfig bond0 down ; modprobe -r bonding
2624# ifconfig bond0 .... up
2625# ifenslave bond0 eth...
2626
2627 This method will automatically take the address from the next
2628slave that is added.
2629
2630 To restore your slaves' MAC addresses, you need to detach them
2631from the bond (`ifenslave -d bond0 eth0'). The bonding driver will
2632then restore the MAC addresses that the slaves had before they were
2633enslaved.
2634
Jay Vosburgh00354cf2005-07-21 12:18:02 -0700263516. Resources and Links
Linus Torvalds1da177e2005-04-16 15:20:36 -07002636=======================
2637
Nicolas de Pesloüana23c37f12011-03-13 10:34:22 +00002638 The latest version of the bonding driver can be found in the latest
Linus Torvalds1da177e2005-04-16 15:20:36 -07002639version of the linux kernel, found on http://kernel.org
2640
Nicolas de Pesloüana23c37f12011-03-13 10:34:22 +00002641 The latest version of this document can be found in the latest kernel
2642source (named Documentation/networking/bonding.txt).
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002643
Nicolas de Pesloüana23c37f12011-03-13 10:34:22 +00002644 Discussions regarding the usage of the bonding driver take place on the
2645bonding-devel mailing list, hosted at sourceforge.net. If you have questions or
2646problems, post them to the list. The list address is:
Linus Torvalds1da177e2005-04-16 15:20:36 -07002647
2648bonding-devel@lists.sourceforge.net
2649
Jay Vosburgh00354cf2005-07-21 12:18:02 -07002650 The administrative interface (to subscribe or unsubscribe) can
2651be found at:
2652
Linus Torvalds1da177e2005-04-16 15:20:36 -07002653https://lists.sourceforge.net/lists/listinfo/bonding-devel
2654
Rick Jonesf8b72d32012-07-20 10:51:37 +00002655 Discussions regarding the development of the bonding driver take place
Nicolas de Pesloüana23c37f12011-03-13 10:34:22 +00002656on the main Linux network mailing list, hosted at vger.kernel.org. The list
2657address is:
2658
2659netdev@vger.kernel.org
2660
2661 The administrative interface (to subscribe or unsubscribe) can
2662be found at:
2663
2664http://vger.kernel.org/vger-lists.html#netdev
2665
Linus Torvalds1da177e2005-04-16 15:20:36 -07002666Donald Becker's Ethernet Drivers and diag programs may be found at :
Justin P. Mattock0ea6e612010-07-23 20:51:24 -07002667 - http://web.archive.org/web/*/http://www.scyld.com/network/
Linus Torvalds1da177e2005-04-16 15:20:36 -07002668
2669You will also find a lot of information regarding Ethernet, NWay, MII,
2670etc. at www.scyld.com.
2671
2672-- END --