1 .. SPDX-License-Identifier: GPL-2.0
3 ===================================
4 Linux Ethernet Bonding Driver HOWTO
5 ===================================
7 Latest update: 27 April 2011
9 Initial release: Thomas Davis <tadavis at lbl.gov>
11 Corrections, HA extensions: 2000/10/03-15:
13 - Willy Tarreau <willy at meta-x.org>
14 - Constantine Gavrilov <const-g at xpert.com>
15 - Chad N. Tindel <ctindel at ieee dot org>
16 - Janice Girouard <girouard at us dot ibm dot com>
17 - Jay Vosburgh <fubar at us dot ibm dot com>
19 Reorganized and updated Feb 2005 by Jay Vosburgh
20 Added Sysfs information: 2006/04/24
22 - Mitch Williams <mitch.a.williams at intel.com>
27 The Linux bonding driver provides a method for aggregating
28 multiple network interfaces into a single logical "bonded" interface.
29 The behavior of the bonded interfaces depends upon the mode; generally
30 speaking, modes provide either hot standby or load balancing services.
31 Additionally, link integrity monitoring may be performed.
33 The bonding driver originally came from Donald Becker's
34 beowulf patches for kernel 2.0. It has changed quite a bit since, and
35 the original tools from extreme-linux and beowulf sites will not work
36 with this version of the driver.
38 For new versions of the driver, updated userspace tools, and
39 who to ask for help, please follow the links at the end of this file.
43 1. Bonding Driver Installation
45 2. Bonding Driver Options
47 3. Configuring Bonding Devices
48 3.1 Configuration with Sysconfig Support
49 3.1.1 Using DHCP with Sysconfig
50 3.1.2 Configuring Multiple Bonds with Sysconfig
51 3.2 Configuration with Initscripts Support
52 3.2.1 Using DHCP with Initscripts
53 3.2.2 Configuring Multiple Bonds with Initscripts
54 3.3 Configuring Bonding Manually with Ifenslave
55 3.3.1 Configuring Multiple Bonds Manually
56 3.4 Configuring Bonding Manually via Sysfs
57 3.5 Configuration with Interfaces Support
58 3.6 Overriding Configuration for Special Cases
59 3.7 Configuring LACP for 802.3ad mode in a more secure way
61 4. Querying Bonding Configuration
62 4.1 Bonding Configuration
63 4.2 Network Configuration
65 5. Switch Configuration
67 6. 802.1q VLAN Support
70 7.1 ARP Monitor Operation
71 7.2 Configuring Multiple ARP Targets
72 7.3 MII Monitor Operation
74 8. Potential Trouble Sources
75 8.1 Adventures in Routing
76 8.2 Ethernet Device Renaming
77 8.3 Painfully Slow Or No Failed Link Detection By Miimon
83 11. Configuring Bonding for High Availability
84 11.1 High Availability in a Single Switch Topology
85 11.2 High Availability in a Multiple Switch Topology
86 11.2.1 HA Bonding Mode Selection for Multiple Switch Topology
87 11.2.2 HA Link Monitoring for Multiple Switch Topology
89 12. Configuring Bonding for Maximum Throughput
90 12.1 Maximum Throughput in a Single Switch Topology
91 12.1.1 MT Bonding Mode Selection for Single Switch Topology
92 12.1.2 MT Link Monitoring for Single Switch Topology
93 12.2 Maximum Throughput in a Multiple Switch Topology
94 12.2.1 MT Bonding Mode Selection for Multiple Switch Topology
95 12.2.2 MT Link Monitoring for Multiple Switch Topology
97 13. Switch Behavior Issues
98 13.1 Link Establishment and Failover Delays
99 13.2 Duplicated Incoming Packets
101 14. Hardware Specific Considerations
104 15. Frequently Asked Questions
106 16. Resources and Links
109 1. Bonding Driver Installation
110 ==============================
112 Most popular distro kernels ship with the bonding driver
113 already available as a module. If your distro does not, or you
114 have need to compile bonding from source (e.g., configuring and
115 installing a mainline kernel from kernel.org), you'll need to perform
118 1.1 Configure and build the kernel with bonding
119 -----------------------------------------------
121 The current version of the bonding driver is available in the
122 drivers/net/bonding subdirectory of the most recent kernel source
123 (which is available on http://kernel.org). Most users "rolling their
124 own" will want to use the most recent kernel from kernel.org.
126 Configure kernel with "make menuconfig" (or "make xconfig" or
127 "make config"), then select "Bonding driver support" in the "Network
128 device support" section. It is recommended that you configure the
129 driver as module since it is currently the only way to pass parameters
130 to the driver or configure more than one bonding device.
132 Build and install the new kernel and modules.
134 1.2 Bonding Control Utility
135 ---------------------------
137 It is recommended to configure bonding via iproute2 (netlink)
138 or sysfs, the old ifenslave control utility is obsolete.
140 2. Bonding Driver Options
141 =========================
143 Options for the bonding driver are supplied as parameters to the
144 bonding module at load time, or are specified via sysfs.
146 Module options may be given as command line arguments to the
147 insmod or modprobe command, but are usually specified in either the
148 ``/etc/modprobe.d/*.conf`` configuration files, or in a distro-specific
149 configuration file (some of which are detailed in the next section).
151 Details on bonding support for sysfs is provided in the
152 "Configuring Bonding Manually via Sysfs" section, below.
154 The available bonding driver parameters are listed below. If a
155 parameter is not specified the default value is used. When initially
156 configuring a bond, it is recommended "tail -f /var/log/messages" be
157 run in a separate window to watch for bonding driver error messages.
159 It is critical that either the miimon or arp_interval and
160 arp_ip_target parameters be specified, otherwise serious network
161 degradation will occur during link failures. Very few devices do not
162 support at least miimon, so there is really no reason not to use it.
164 Options with textual values will accept either the text name
165 or, for backwards compatibility, the option value. E.g.,
166 "mode=802.3ad" and "mode=4" set the same mode.
168 The parameters are as follows:
172 Specifies the new active slave for modes that support it
173 (active-backup, balance-alb and balance-tlb). Possible values
174 are the name of any currently enslaved interface, or an empty
175 string. If a name is given, the slave and its link must be up in order
176 to be selected as the new active slave. If an empty string is
177 specified, the current active slave is cleared, and a new active
178 slave is selected automatically.
180 Note that this is only available through the sysfs interface. No module
181 parameter by this name exists.
183 The normal value of this option is the name of the currently
184 active slave, or the empty string if there is no active slave or
185 the current mode does not use an active slave.
189 In an AD system, this specifies the system priority. The allowed range
190 is 1 - 65535. If the value is not specified, it takes 65535 as the
193 This parameter has effect only in 802.3ad mode and is available through
198 In an AD system, this specifies the mac-address for the actor in
199 protocol packet exchanges (LACPDUs). The value cannot be a multicast
200 address. If the all-zeroes MAC is specified, bonding will internally
201 use the MAC of the bond itself. It is preferred to have the
202 local-admin bit set for this mac but driver does not enforce it. If
203 the value is not given then system defaults to using the masters'
204 mac address as actors' system address.
206 This parameter has effect only in 802.3ad mode and is available through
211 Specifies the 802.3ad aggregation selection logic to use. The
212 possible values and their effects are:
216 The active aggregator is chosen by largest aggregate
219 Reselection of the active aggregator occurs only when all
220 slaves of the active aggregator are down or the active
221 aggregator has no slaves.
223 This is the default value.
227 The active aggregator is chosen by largest aggregate
228 bandwidth. Reselection occurs if:
230 - A slave is added to or removed from the bond
232 - Any slave's link state changes
234 - Any slave's 802.3ad association state changes
236 - The bond's administrative state changes to up
240 The active aggregator is chosen by the largest number of
241 ports (slaves). Reselection occurs as described under the
242 "bandwidth" setting, above.
244 The bandwidth and count selection policies permit failover of
245 802.3ad aggregations when partial failure of the active aggregator
246 occurs. This keeps the aggregator with the highest availability
247 (either in bandwidth or in number of ports) active at all times.
249 This option was added in bonding version 3.4.0.
253 In an AD system, the port-key has three parts as shown below -
263 This defines the upper 10 bits of the port key. The values can be
264 from 0 - 1023. If not given, the system defaults to 0.
266 This parameter has effect only in 802.3ad mode and is available through
271 Specifies that duplicate frames (received on inactive ports) should be
272 dropped (0) or delivered (1).
274 Normally, bonding will drop duplicate frames (received on inactive
275 ports), which is desirable for most users. But there are some times
276 it is nice to allow duplicate frames to be delivered.
278 The default value is 0 (drop duplicate frames received on inactive
283 Specifies the ARP link monitoring frequency in milliseconds.
285 The ARP monitor works by periodically checking the slave
286 devices to determine whether they have sent or received
287 traffic recently (the precise criteria depends upon the
288 bonding mode, and the state of the slave). Regular traffic is
289 generated via ARP probes issued for the addresses specified by
290 the arp_ip_target option.
292 This behavior can be modified by the arp_validate option,
295 If ARP monitoring is used in an etherchannel compatible mode
296 (modes 0 and 2), the switch should be configured in a mode
297 that evenly distributes packets across all links. If the
298 switch is configured to distribute the packets in an XOR
299 fashion, all replies from the ARP targets will be received on
300 the same link which could cause the other team members to
301 fail. ARP monitoring should not be used in conjunction with
302 miimon. A value of 0 disables ARP monitoring. The default
307 Specifies the IP addresses to use as ARP monitoring peers when
308 arp_interval is > 0. These are the targets of the ARP request
309 sent to determine the health of the link to the targets.
310 Specify these values in ddd.ddd.ddd.ddd format. Multiple IP
311 addresses must be separated by a comma. At least one IP
312 address must be given for ARP monitoring to function. The
313 maximum number of targets that can be specified is 16. The
314 default value is no IP addresses.
318 Specifies the IPv6 addresses to use as IPv6 monitoring peers when
319 arp_interval is > 0. These are the targets of the NS request
320 sent to determine the health of the link to the targets.
321 Specify these values in ffff:ffff::ffff:ffff format. Multiple IPv6
322 addresses must be separated by a comma. At least one IPv6
323 address must be given for NS/NA monitoring to function. The
324 maximum number of targets that can be specified is 16. The
325 default value is no IPv6 addresses.
329 Specifies whether or not ARP probes and replies should be
330 validated in any mode that supports arp monitoring, or whether
331 non-ARP traffic should be filtered (disregarded) for link
338 No validation or filtering is performed.
342 Validation is performed only for the active slave.
346 Validation is performed only for backup slaves.
350 Validation is performed for all slaves.
354 Filtering is applied to all slaves. No validation is
359 Filtering is applied to all slaves, validation is performed
360 only for the active slave.
364 Filtering is applied to all slaves, validation is performed
365 only for backup slaves.
369 Enabling validation causes the ARP monitor to examine the incoming
370 ARP requests and replies, and only consider a slave to be up if it
371 is receiving the appropriate ARP traffic.
373 For an active slave, the validation checks ARP replies to confirm
374 that they were generated by an arp_ip_target. Since backup slaves
375 do not typically receive these replies, the validation performed
376 for backup slaves is on the broadcast ARP request sent out via the
377 active slave. It is possible that some switch or network
378 configurations may result in situations wherein the backup slaves
379 do not receive the ARP requests; in such a situation, validation
380 of backup slaves must be disabled.
382 The validation of ARP requests on backup slaves is mainly helping
383 bonding to decide which slaves are more likely to work in case of
384 the active slave failure, it doesn't really guarantee that the
385 backup slave will work if it's selected as the next active slave.
387 Validation is useful in network configurations in which multiple
388 bonding hosts are concurrently issuing ARPs to one or more targets
389 beyond a common switch. Should the link between the switch and
390 target fail (but not the switch itself), the probe traffic
391 generated by the multiple bonding instances will fool the standard
392 ARP monitor into considering the links as still up. Use of
393 validation can resolve this, as the ARP monitor will only consider
394 ARP requests and replies associated with its own instance of
399 Enabling filtering causes the ARP monitor to only use incoming ARP
400 packets for link availability purposes. Arriving packets that are
401 not ARPs are delivered normally, but do not count when determining
402 if a slave is available.
404 Filtering operates by only considering the reception of ARP
405 packets (any ARP packet, regardless of source or destination) when
406 determining if a slave has received traffic for link availability
409 Filtering is useful in network configurations in which significant
410 levels of third party broadcast traffic would fool the standard
411 ARP monitor into considering the links as still up. Use of
412 filtering can resolve this, as only ARP traffic is considered for
413 link availability purposes.
415 This option was added in bonding version 3.1.0.
419 Specifies the quantity of arp_ip_targets that must be reachable
420 in order for the ARP monitor to consider a slave as being up.
421 This option affects only active-backup mode for slaves with
422 arp_validation enabled.
428 consider the slave up only when any of the arp_ip_targets
433 consider the slave up only when all of the arp_ip_targets
438 Specifies the number of arp_interval monitor checks that must
439 fail in order for an interface to be marked down by the ARP monitor.
441 In order to provide orderly failover semantics, backup interfaces
442 are permitted an extra monitor check (i.e., they must fail
443 arp_missed_max + 1 times before being marked down).
445 The default value is 2, and the allowable range is 1 - 255.
449 Specifies the time, in milliseconds, to wait before disabling
450 a slave after a link failure has been detected. This option
451 is only valid for the miimon link monitor. The downdelay
452 value should be a multiple of the miimon value; if not, it
453 will be rounded down to the nearest multiple. The default
458 Specifies whether active-backup mode should set all slaves to
459 the same MAC address at enslavement (the traditional
460 behavior), or, when enabled, perform special handling of the
461 bond's MAC address in accordance with the selected policy.
467 This setting disables fail_over_mac, and causes
468 bonding to set all slaves of an active-backup bond to
469 the same MAC address at enslavement time. This is the
474 The "active" fail_over_mac policy indicates that the
475 MAC address of the bond should always be the MAC
476 address of the currently active slave. The MAC
477 address of the slaves is not changed; instead, the MAC
478 address of the bond changes during a failover.
480 This policy is useful for devices that cannot ever
481 alter their MAC address, or for devices that refuse
482 incoming broadcasts with their own source MAC (which
483 interferes with the ARP monitor).
485 The down side of this policy is that every device on
486 the network must be updated via gratuitous ARP,
487 vs. just updating a switch or set of switches (which
488 often takes place for any traffic, not just ARP
489 traffic, if the switch snoops incoming traffic to
490 update its tables) for the traditional method. If the
491 gratuitous ARP is lost, communication may be
494 When this policy is used in conjunction with the mii
495 monitor, devices which assert link up prior to being
496 able to actually transmit and receive are particularly
497 susceptible to loss of the gratuitous ARP, and an
498 appropriate updelay setting may be required.
502 The "follow" fail_over_mac policy causes the MAC
503 address of the bond to be selected normally (normally
504 the MAC address of the first slave added to the bond).
505 However, the second and subsequent slaves are not set
506 to this MAC address while they are in a backup role; a
507 slave is programmed with the bond's MAC address at
508 failover time (and the formerly active slave receives
509 the newly active slave's MAC address).
511 This policy is useful for multiport devices that
512 either become confused or incur a performance penalty
513 when multiple ports are programmed with the same MAC
517 The default policy is none, unless the first slave cannot
518 change its MAC address, in which case the active policy is
521 This option may be modified via sysfs only when no slaves are
524 This option was added in bonding version 3.2.0. The "follow"
525 policy was added in bonding version 3.3.0.
528 Option specifying whether to send LACPDU frames periodically.
531 LACPDU frames acts as "speak when spoken to".
534 LACPDU frames are sent along the configured links
535 periodically. See lacp_rate for more details.
541 Option specifying the rate in which we'll ask our link partner
542 to transmit LACPDU packets in 802.3ad mode. Possible values
546 Request partner to transmit LACPDUs every 30 seconds
549 Request partner to transmit LACPDUs every 1 second
555 Specifies the number of bonding devices to create for this
556 instance of the bonding driver. E.g., if max_bonds is 3, and
557 the bonding driver is not already loaded, then bond0, bond1
558 and bond2 will be created. The default value is 1. Specifying
559 a value of 0 will load bonding, but will not create any devices.
563 Specifies the MII link monitoring frequency in milliseconds.
564 This determines how often the link state of each slave is
565 inspected for link failures. A value of zero disables MII
566 link monitoring. A value of 100 is a good starting point.
567 The use_carrier option, below, affects how the link state is
568 determined. See the High Availability section for additional
569 information. The default value is 0.
573 Specifies the minimum number of links that must be active before
574 asserting carrier. It is similar to the Cisco EtherChannel min-links
575 feature. This allows setting the minimum number of member ports that
576 must be up (link-up state) before marking the bond device as up
577 (carrier on). This is useful for situations where higher level services
578 such as clustering want to ensure a minimum number of low bandwidth
579 links are active before switchover. This option only affect 802.3ad
582 The default value is 0. This will cause carrier to be asserted (for
583 802.3ad mode) whenever there is an active aggregator, regardless of the
584 number of available links in that aggregator. Note that, because an
585 aggregator cannot be active without at least one available link,
586 setting this option to 0 or to 1 has the exact same effect.
590 Specifies one of the bonding policies. The default is
591 balance-rr (round robin). Possible values are:
595 Round-robin policy: Transmit packets in sequential
596 order from the first available slave through the
597 last. This mode provides load balancing and fault
602 Active-backup policy: Only one slave in the bond is
603 active. A different slave becomes active if, and only
604 if, the active slave fails. The bond's MAC address is
605 externally visible on only one port (network adapter)
606 to avoid confusing the switch.
608 In bonding version 2.6.2 or later, when a failover
609 occurs in active-backup mode, bonding will issue one
610 or more gratuitous ARPs on the newly active slave.
611 One gratuitous ARP is issued for the bonding master
612 interface and each VLAN interfaces configured above
613 it, provided that the interface has at least one IP
614 address configured. Gratuitous ARPs issued for VLAN
615 interfaces are tagged with the appropriate VLAN id.
617 This mode provides fault tolerance. The primary
618 option, documented below, affects the behavior of this
623 XOR policy: Transmit based on the selected transmit
624 hash policy. The default policy is a simple [(source
625 MAC address XOR'd with destination MAC address XOR
626 packet type ID) modulo slave count]. Alternate transmit
627 policies may be selected via the xmit_hash_policy option,
630 This mode provides load balancing and fault tolerance.
634 Broadcast policy: transmits everything on all slave
635 interfaces. This mode provides fault tolerance.
639 IEEE 802.3ad Dynamic link aggregation. Creates
640 aggregation groups that share the same speed and
641 duplex settings. Utilizes all slaves in the active
642 aggregator according to the 802.3ad specification.
644 Slave selection for outgoing traffic is done according
645 to the transmit hash policy, which may be changed from
646 the default simple XOR policy via the xmit_hash_policy
647 option, documented below. Note that not all transmit
648 policies may be 802.3ad compliant, particularly in
649 regards to the packet mis-ordering requirements of
650 section 43.2.4 of the 802.3ad standard. Differing
651 peer implementations will have varying tolerances for
656 1. Ethtool support in the base drivers for retrieving
657 the speed and duplex of each slave.
659 2. A switch that supports IEEE 802.3ad Dynamic link
662 Most switches will require some type of configuration
663 to enable 802.3ad mode.
667 Adaptive transmit load balancing: channel bonding that
668 does not require any special switch support.
670 In tlb_dynamic_lb=1 mode; the outgoing traffic is
671 distributed according to the current load (computed
672 relative to the speed) on each slave.
674 In tlb_dynamic_lb=0 mode; the load balancing based on
675 current load is disabled and the load is distributed
676 only using the hash distribution.
678 Incoming traffic is received by the current slave.
679 If the receiving slave fails, another slave takes over
680 the MAC address of the failed receiving slave.
684 Ethtool support in the base drivers for retrieving the
689 Adaptive load balancing: includes balance-tlb plus
690 receive load balancing (rlb) for IPV4 traffic, and
691 does not require any special switch support. The
692 receive load balancing is achieved by ARP negotiation.
693 The bonding driver intercepts the ARP Replies sent by
694 the local system on their way out and overwrites the
695 source hardware address with the unique hardware
696 address of one of the slaves in the bond such that
697 different peers use different hardware addresses for
700 Receive traffic from connections created by the server
701 is also balanced. When the local system sends an ARP
702 Request the bonding driver copies and saves the peer's
703 IP information from the ARP packet. When the ARP
704 Reply arrives from the peer, its hardware address is
705 retrieved and the bonding driver initiates an ARP
706 reply to this peer assigning it to one of the slaves
707 in the bond. A problematic outcome of using ARP
708 negotiation for balancing is that each time that an
709 ARP request is broadcast it uses the hardware address
710 of the bond. Hence, peers learn the hardware address
711 of the bond and the balancing of receive traffic
712 collapses to the current slave. This is handled by
713 sending updates (ARP Replies) to all the peers with
714 their individually assigned hardware address such that
715 the traffic is redistributed. Receive traffic is also
716 redistributed when a new slave is added to the bond
717 and when an inactive slave is re-activated. The
718 receive load is distributed sequentially (round robin)
719 among the group of highest speed slaves in the bond.
721 When a link is reconnected or a new slave joins the
722 bond the receive traffic is redistributed among all
723 active slaves in the bond by initiating ARP Replies
724 with the selected MAC address to each of the
725 clients. The updelay parameter (detailed below) must
726 be set to a value equal or greater than the switch's
727 forwarding delay so that the ARP Replies sent to the
728 peers will not be blocked by the switch.
732 1. Ethtool support in the base drivers for retrieving
733 the speed of each slave.
735 2. Base driver support for setting the hardware
736 address of a device while it is open. This is
737 required so that there will always be one slave in the
738 team using the bond hardware address (the
739 curr_active_slave) while having a unique hardware
740 address for each slave in the bond. If the
741 curr_active_slave fails its hardware address is
742 swapped with the new curr_active_slave that was
748 Specify the number of peer notifications (gratuitous ARPs and
749 unsolicited IPv6 Neighbor Advertisements) to be issued after a
750 failover event. As soon as the link is up on the new slave
751 (possibly immediately) a peer notification is sent on the
752 bonding device and each VLAN sub-device. This is repeated at
753 the rate specified by peer_notif_delay if the number is
756 The valid range is 0 - 255; the default value is 1. These options
757 affect only the active-backup mode. These options were added for
758 bonding versions 3.3.0 and 3.4.0 respectively.
760 From Linux 3.0 and bonding version 3.7.1, these notifications
761 are generated by the ipv4 and ipv6 code and the numbers of
762 repetitions cannot be set independently.
766 Specify the number of packets to transmit through a slave before
767 moving to the next one. When set to 0 then a slave is chosen at
770 The valid range is 0 - 65535; the default value is 1. This option
771 has effect only in balance-rr mode.
775 Specify the delay, in milliseconds, between each peer
776 notification (gratuitous ARP and unsolicited IPv6 Neighbor
777 Advertisement) when they are issued after a failover event.
778 This delay should be a multiple of the link monitor interval
779 (arp_interval or miimon, whichever is active). The default
780 value is 0 which means to match the value of the link monitor
784 Slave priority. A higher number means higher priority.
785 The primary slave has the highest priority. This option also
786 follows the primary_reselect rules.
788 This option could only be configured via netlink, and is only valid
789 for active-backup(1), balance-tlb (5) and balance-alb (6) mode.
790 The valid value range is a signed 32 bit integer.
792 The default value is 0.
796 A string (eth0, eth2, etc) specifying which slave is the
797 primary device. The specified device will always be the
798 active slave while it is available. Only when the primary is
799 off-line will alternate devices be used. This is useful when
800 one slave is preferred over another, e.g., when one slave has
801 higher throughput than another.
803 The primary option is only valid for active-backup(1),
804 balance-tlb (5) and balance-alb (6) mode.
808 Specifies the reselection policy for the primary slave. This
809 affects how the primary slave is chosen to become the active slave
810 when failure of the active slave or recovery of the primary slave
811 occurs. This option is designed to prevent flip-flopping between
812 the primary slave and other slaves. Possible values are:
814 always or 0 (default)
816 The primary slave becomes the active slave whenever it
821 The primary slave becomes the active slave when it comes
822 back up, if the speed and duplex of the primary slave is
823 better than the speed and duplex of the current active
828 The primary slave becomes the active slave only if the
829 current active slave fails and the primary slave is up.
831 The primary_reselect setting is ignored in two cases:
833 If no slaves are active, the first slave to recover is
834 made the active slave.
836 When initially enslaved, the primary slave is always made
839 Changing the primary_reselect policy via sysfs will cause an
840 immediate selection of the best active slave according to the new
841 policy. This may or may not result in a change of the active
842 slave, depending upon the circumstances.
844 This option was added for bonding version 3.6.0.
848 Specifies if dynamic shuffling of flows is enabled in tlb
849 mode. The value has no effect on any other modes.
851 The default behavior of tlb mode is to shuffle active flows across
852 slaves based on the load in that interval. This gives nice lb
853 characteristics but can cause packet reordering. If re-ordering is
854 a concern use this variable to disable flow shuffling and rely on
855 load balancing provided solely by the hash distribution.
856 xmit-hash-policy can be used to select the appropriate hashing for
859 The sysfs entry can be used to change the setting per bond device
860 and the initial value is derived from the module parameter. The
861 sysfs entry is allowed to be changed only if the bond device is
864 The default value is "1" that enables flow shuffling while value "0"
865 disables it. This option was added in bonding driver 3.7.1
870 Specifies the time, in milliseconds, to wait before enabling a
871 slave after a link recovery has been detected. This option is
872 only valid for the miimon link monitor. The updelay value
873 should be a multiple of the miimon value; if not, it will be
874 rounded down to the nearest multiple. The default value is 0.
878 Specifies whether or not miimon should use MII or ETHTOOL
879 ioctls vs. netif_carrier_ok() to determine the link
880 status. The MII or ETHTOOL ioctls are less efficient and
881 utilize a deprecated calling sequence within the kernel. The
882 netif_carrier_ok() relies on the device driver to maintain its
883 state with netif_carrier_on/off; at this writing, most, but
884 not all, device drivers support this facility.
886 If bonding insists that the link is up when it should not be,
887 it may be that your network device driver does not support
888 netif_carrier_on/off. The default state for netif_carrier is
889 "carrier on," so if a driver does not support netif_carrier,
890 it will appear as if the link is always up. In this case,
891 setting use_carrier to 0 will cause bonding to revert to the
892 MII / ETHTOOL ioctl method to determine the link state.
894 A value of 1 enables the use of netif_carrier_ok(), a value of
895 0 will use the deprecated MII / ETHTOOL ioctls. The default
900 Selects the transmit hash policy to use for slave selection in
901 balance-xor, 802.3ad, and tlb modes. Possible values are:
905 Uses XOR of hardware MAC addresses and packet type ID
906 field to generate the hash. The formula is
908 hash = source MAC[5] XOR destination MAC[5] XOR packet type ID
909 slave number = hash modulo slave count
911 This algorithm will place all traffic to a particular
912 network peer on the same slave.
914 This algorithm is 802.3ad compliant.
918 This policy uses a combination of layer2 and layer3
919 protocol information to generate the hash.
921 Uses XOR of hardware MAC addresses and IP addresses to
922 generate the hash. The formula is
924 hash = source MAC[5] XOR destination MAC[5] XOR packet type ID
925 hash = hash XOR source IP XOR destination IP
926 hash = hash XOR (hash RSHIFT 16)
927 hash = hash XOR (hash RSHIFT 8)
928 And then hash is reduced modulo slave count.
930 If the protocol is IPv6 then the source and destination
931 addresses are first hashed using ipv6_addr_hash.
933 This algorithm will place all traffic to a particular
934 network peer on the same slave. For non-IP traffic,
935 the formula is the same as for the layer2 transmit
938 This policy is intended to provide a more balanced
939 distribution of traffic than layer2 alone, especially
940 in environments where a layer3 gateway device is
941 required to reach most destinations.
943 This algorithm is 802.3ad compliant.
947 This policy uses upper layer protocol information,
948 when available, to generate the hash. This allows for
949 traffic to a particular network peer to span multiple
950 slaves, although a single connection will not span
953 The formula for unfragmented TCP and UDP packets is
955 hash = source port, destination port (as in the header)
956 hash = hash XOR source IP XOR destination IP
957 hash = hash XOR (hash RSHIFT 16)
958 hash = hash XOR (hash RSHIFT 8)
959 And then hash is reduced modulo slave count.
961 If the protocol is IPv6 then the source and destination
962 addresses are first hashed using ipv6_addr_hash.
964 For fragmented TCP or UDP packets and all other IPv4 and
965 IPv6 protocol traffic, the source and destination port
966 information is omitted. For non-IP traffic, the
967 formula is the same as for the layer2 transmit hash
970 This algorithm is not fully 802.3ad compliant. A
971 single TCP or UDP conversation containing both
972 fragmented and unfragmented packets will see packets
973 striped across two interfaces. This may result in out
974 of order delivery. Most traffic types will not meet
975 this criteria, as TCP rarely fragments traffic, and
976 most UDP traffic is not involved in extended
977 conversations. Other implementations of 802.3ad may
978 or may not tolerate this noncompliance.
982 This policy uses the same formula as layer2+3 but it
983 relies on skb_flow_dissect to obtain the header fields
984 which might result in the use of inner headers if an
985 encapsulation protocol is used. For example this will
986 improve the performance for tunnel users because the
987 packets will be distributed according to the encapsulated
992 This policy uses the same formula as layer3+4 but it
993 relies on skb_flow_dissect to obtain the header fields
994 which might result in the use of inner headers if an
995 encapsulation protocol is used. For example this will
996 improve the performance for tunnel users because the
997 packets will be distributed according to the encapsulated
1002 This policy uses a very rudimentary vlan ID and source mac
1003 hash to load-balance traffic per-vlan, with failover
1004 should one leg fail. The intended use case is for a bond
1005 shared by multiple virtual machines, all configured to
1006 use their own vlan, to give lacp-like functionality
1007 without requiring lacp-capable switching hardware.
1009 The formula for the hash is simply
1011 hash = (vlan ID) XOR (source MAC vendor) XOR (source MAC dev)
1013 The default value is layer2. This option was added in bonding
1014 version 2.6.3. In earlier versions of bonding, this parameter
1015 does not exist, and the layer2 policy is the only policy. The
1016 layer2+3 value was added for bonding version 3.2.2.
1020 Specifies the number of IGMP membership reports to be issued after
1021 a failover event. One membership report is issued immediately after
1022 the failover, subsequent packets are sent in each 200ms interval.
1024 The valid range is 0 - 255; the default value is 1. A value of 0
1025 prevents the IGMP membership report from being issued in response
1026 to the failover event.
1028 This option is useful for bonding modes balance-rr (0), active-backup
1029 (1), balance-tlb (5) and balance-alb (6), in which a failover can
1030 switch the IGMP traffic from one slave to another. Therefore a fresh
1031 IGMP report must be issued to cause the switch to forward the incoming
1032 IGMP traffic over the newly selected slave.
1034 This option was added for bonding version 3.7.0.
1038 Specifies the number of seconds between instances where the bonding
1039 driver sends learning packets to each slaves peer switch.
1041 The valid range is 1 - 0x7fffffff; the default value is 1. This Option
1042 has effect only in balance-tlb and balance-alb modes.
1044 3. Configuring Bonding Devices
1045 ==============================
1047 You can configure bonding using either your distro's network
1048 initialization scripts, or manually using either iproute2 or the
1049 sysfs interface. Distros generally use one of three packages for the
1050 network initialization scripts: initscripts, sysconfig or interfaces.
1051 Recent versions of these packages have support for bonding, while older
1054 We will first describe the options for configuring bonding for
1055 distros using versions of initscripts, sysconfig and interfaces with full
1056 or partial support for bonding, then provide information on enabling
1057 bonding without support from the network initialization scripts (i.e.,
1058 older versions of initscripts or sysconfig).
1060 If you're unsure whether your distro uses sysconfig,
1061 initscripts or interfaces, or don't know if it's new enough, have no fear.
1062 Determining this is fairly straightforward.
1064 First, look for a file called interfaces in /etc/network directory.
1065 If this file is present in your system, then your system use interfaces. See
1066 Configuration with Interfaces Support.
1068 Else, issue the command::
1070 $ rpm -qf /sbin/ifup
1072 It will respond with a line of text starting with either
1073 "initscripts" or "sysconfig," followed by some numbers. This is the
1074 package that provides your network initialization scripts.
1076 Next, to determine if your installation supports bonding,
1079 $ grep ifenslave /sbin/ifup
1081 If this returns any matches, then your initscripts or
1082 sysconfig has support for bonding.
1084 3.1 Configuration with Sysconfig Support
1085 ----------------------------------------
1087 This section applies to distros using a version of sysconfig
1088 with bonding support, for example, SuSE Linux Enterprise Server 9.
1090 SuSE SLES 9's networking configuration system does support
1091 bonding, however, at this writing, the YaST system configuration
1092 front end does not provide any means to work with bonding devices.
1093 Bonding devices can be managed by hand, however, as follows.
1095 First, if they have not already been configured, configure the
1096 slave devices. On SLES 9, this is most easily done by running the
1097 yast2 sysconfig configuration utility. The goal is for to create an
1098 ifcfg-id file for each slave device. The simplest way to accomplish
1099 this is to configure the devices for DHCP (this is only to get the
1100 file ifcfg-id file created; see below for some issues with DHCP). The
1101 name of the configuration file for each device will be of the form::
1103 ifcfg-id-xx:xx:xx:xx:xx:xx
1105 Where the "xx" portion will be replaced with the digits from
1106 the device's permanent MAC address.
1108 Once the set of ifcfg-id-xx:xx:xx:xx:xx:xx files has been
1109 created, it is necessary to edit the configuration files for the slave
1110 devices (the MAC addresses correspond to those of the slave devices).
1111 Before editing, the file will contain multiple lines, and will look
1112 something like this::
1117 UNIQUE='XNzu.WeZGOGF+4wE'
1118 _nm_name='bus-pci-0001:61:01.0'
1120 Change the BOOTPROTO and STARTMODE lines to the following::
1125 Do not alter the UNIQUE or _nm_name lines. Remove any other
1126 lines (USERCTL, etc).
1128 Once the ifcfg-id-xx:xx:xx:xx:xx:xx files have been modified,
1129 it's time to create the configuration file for the bonding device
1130 itself. This file is named ifcfg-bondX, where X is the number of the
1131 bonding device to create, starting at 0. The first such file is
1132 ifcfg-bond0, the second is ifcfg-bond1, and so on. The sysconfig
1133 network configuration system will correctly start multiple instances
1136 The contents of the ifcfg-bondX file is as follows::
1139 BROADCAST="10.0.2.255"
1141 NETMASK="255.255.0.0"
1145 BONDING_MASTER="yes"
1146 BONDING_MODULE_OPTS="mode=active-backup miimon=100"
1147 BONDING_SLAVE0="eth0"
1148 BONDING_SLAVE1="bus-pci-0000:06:08.1"
1150 Replace the sample BROADCAST, IPADDR, NETMASK and NETWORK
1151 values with the appropriate values for your network.
1153 The STARTMODE specifies when the device is brought online.
1154 The possible values are:
1156 ======== ======================================================
1157 onboot The device is started at boot time. If you're not
1158 sure, this is probably what you want.
1160 manual The device is started only when ifup is called
1161 manually. Bonding devices may be configured this
1162 way if you do not wish them to start automatically
1163 at boot for some reason.
1165 hotplug The device is started by a hotplug event. This is not
1166 a valid choice for a bonding device.
1168 off or The device configuration is ignored.
1170 ======== ======================================================
1172 The line BONDING_MASTER='yes' indicates that the device is a
1173 bonding master device. The only useful value is "yes."
1175 The contents of BONDING_MODULE_OPTS are supplied to the
1176 instance of the bonding module for this device. Specify the options
1177 for the bonding mode, link monitoring, and so on here. Do not include
1178 the max_bonds bonding parameter; this will confuse the configuration
1179 system if you have multiple bonding devices.
1181 Finally, supply one BONDING_SLAVEn="slave device" for each
1182 slave. where "n" is an increasing value, one for each slave. The
1183 "slave device" is either an interface name, e.g., "eth0", or a device
1184 specifier for the network device. The interface name is easier to
1185 find, but the ethN names are subject to change at boot time if, e.g.,
1186 a device early in the sequence has failed. The device specifiers
1187 (bus-pci-0000:06:08.1 in the example above) specify the physical
1188 network device, and will not change unless the device's bus location
1189 changes (for example, it is moved from one PCI slot to another). The
1190 example above uses one of each type for demonstration purposes; most
1191 configurations will choose one or the other for all slave devices.
1193 When all configuration files have been modified or created,
1194 networking must be restarted for the configuration changes to take
1195 effect. This can be accomplished via the following::
1197 # /etc/init.d/network restart
1199 Note that the network control script (/sbin/ifdown) will
1200 remove the bonding module as part of the network shutdown processing,
1201 so it is not necessary to remove the module by hand if, e.g., the
1202 module parameters have changed.
1204 Also, at this writing, YaST/YaST2 will not manage bonding
1205 devices (they do not show bonding interfaces on its list of network
1206 devices). It is necessary to edit the configuration file by hand to
1207 change the bonding configuration.
1209 Additional general options and details of the ifcfg file
1210 format can be found in an example ifcfg template file::
1212 /etc/sysconfig/network/ifcfg.template
1214 Note that the template does not document the various ``BONDING_*``
1215 settings described above, but does describe many of the other options.
1217 3.1.1 Using DHCP with Sysconfig
1218 -------------------------------
1220 Under sysconfig, configuring a device with BOOTPROTO='dhcp'
1221 will cause it to query DHCP for its IP address information. At this
1222 writing, this does not function for bonding devices; the scripts
1223 attempt to obtain the device address from DHCP prior to adding any of
1224 the slave devices. Without active slaves, the DHCP requests are not
1225 sent to the network.
1227 3.1.2 Configuring Multiple Bonds with Sysconfig
1228 -----------------------------------------------
1230 The sysconfig network initialization system is capable of
1231 handling multiple bonding devices. All that is necessary is for each
1232 bonding instance to have an appropriately configured ifcfg-bondX file
1233 (as described above). Do not specify the "max_bonds" parameter to any
1234 instance of bonding, as this will confuse sysconfig. If you require
1235 multiple bonding devices with identical parameters, create multiple
1238 Because the sysconfig scripts supply the bonding module
1239 options in the ifcfg-bondX file, it is not necessary to add them to
1240 the system ``/etc/modules.d/*.conf`` configuration files.
1242 3.2 Configuration with Initscripts Support
1243 ------------------------------------------
1245 This section applies to distros using a recent version of
1246 initscripts with bonding support, for example, Red Hat Enterprise Linux
1247 version 3 or later, Fedora, etc. On these systems, the network
1248 initialization scripts have knowledge of bonding, and can be configured to
1249 control bonding devices. Note that older versions of the initscripts
1250 package have lower levels of support for bonding; this will be noted where
1253 These distros will not automatically load the network adapter
1254 driver unless the ethX device is configured with an IP address.
1255 Because of this constraint, users must manually configure a
1256 network-script file for all physical adapters that will be members of
1257 a bondX link. Network script files are located in the directory:
1259 /etc/sysconfig/network-scripts
1261 The file name must be prefixed with "ifcfg-eth" and suffixed
1262 with the adapter's physical adapter number. For example, the script
1263 for eth0 would be named /etc/sysconfig/network-scripts/ifcfg-eth0.
1264 Place the following text in the file::
1273 The DEVICE= line will be different for every ethX device and
1274 must correspond with the name of the file, i.e., ifcfg-eth1 must have
1275 a device line of DEVICE=eth1. The setting of the MASTER= line will
1276 also depend on the final bonding interface name chosen for your bond.
1277 As with other network devices, these typically start at 0, and go up
1278 one for each device, i.e., the first bonding instance is bond0, the
1279 second is bond1, and so on.
1281 Next, create a bond network script. The file name for this
1282 script will be /etc/sysconfig/network-scripts/ifcfg-bondX where X is
1283 the number of the bond. For bond0 the file is named "ifcfg-bond0",
1284 for bond1 it is named "ifcfg-bond1", and so on. Within that file,
1285 place the following text::
1289 NETMASK=255.255.255.0
1291 BROADCAST=192.168.1.255
1296 Be sure to change the networking specific lines (IPADDR,
1297 NETMASK, NETWORK and BROADCAST) to match your network configuration.
1299 For later versions of initscripts, such as that found with Fedora
1300 7 (or later) and Red Hat Enterprise Linux version 5 (or later), it is possible,
1301 and, indeed, preferable, to specify the bonding options in the ifcfg-bond0
1302 file, e.g. a line of the format::
1304 BONDING_OPTS="mode=active-backup arp_interval=60 arp_ip_target=192.168.1.254"
1306 will configure the bond with the specified options. The options
1307 specified in BONDING_OPTS are identical to the bonding module parameters
1308 except for the arp_ip_target field when using versions of initscripts older
1309 than and 8.57 (Fedora 8) and 8.45.19 (Red Hat Enterprise Linux 5.2). When
1310 using older versions each target should be included as a separate option and
1311 should be preceded by a '+' to indicate it should be added to the list of
1312 queried targets, e.g.,::
1314 arp_ip_target=+192.168.1.1 arp_ip_target=+192.168.1.2
1316 is the proper syntax to specify multiple targets. When specifying
1317 options via BONDING_OPTS, it is not necessary to edit
1318 ``/etc/modprobe.d/*.conf``.
1320 For even older versions of initscripts that do not support
1321 BONDING_OPTS, it is necessary to edit /etc/modprobe.d/*.conf, depending upon
1322 your distro) to load the bonding module with your desired options when the
1323 bond0 interface is brought up. The following lines in /etc/modprobe.d/*.conf
1324 will load the bonding module, and select its options:
1327 options bond0 mode=balance-alb miimon=100
1329 Replace the sample parameters with the appropriate set of
1330 options for your configuration.
1332 Finally run "/etc/rc.d/init.d/network restart" as root. This
1333 will restart the networking subsystem and your bond link should be now
1336 3.2.1 Using DHCP with Initscripts
1337 ---------------------------------
1339 Recent versions of initscripts (the versions supplied with Fedora
1340 Core 3 and Red Hat Enterprise Linux 4, or later versions, are reported to
1341 work) have support for assigning IP information to bonding devices via
1344 To configure bonding for DHCP, configure it as described
1345 above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp"
1346 and add a line consisting of "TYPE=Bonding". Note that the TYPE value
1349 3.2.2 Configuring Multiple Bonds with Initscripts
1350 -------------------------------------------------
1352 Initscripts packages that are included with Fedora 7 and Red Hat
1353 Enterprise Linux 5 support multiple bonding interfaces by simply
1354 specifying the appropriate BONDING_OPTS= in ifcfg-bondX where X is the
1355 number of the bond. This support requires sysfs support in the kernel,
1356 and a bonding driver of version 3.0.0 or later. Other configurations may
1357 not support this method for specifying multiple bonding interfaces; for
1358 those instances, see the "Configuring Multiple Bonds Manually" section,
1361 3.3 Configuring Bonding Manually with iproute2
1362 -----------------------------------------------
1364 This section applies to distros whose network initialization
1365 scripts (the sysconfig or initscripts package) do not have specific
1366 knowledge of bonding. One such distro is SuSE Linux Enterprise Server
1369 The general method for these systems is to place the bonding
1370 module parameters into a config file in /etc/modprobe.d/ (as
1371 appropriate for the installed distro), then add modprobe and/or
1372 `ip link` commands to the system's global init script. The name of
1373 the global init script differs; for sysconfig, it is
1374 /etc/init.d/boot.local and for initscripts it is /etc/rc.d/rc.local.
1376 For example, if you wanted to make a simple bond of two e100
1377 devices (presumed to be eth0 and eth1), and have it persist across
1378 reboots, edit the appropriate file (/etc/init.d/boot.local or
1379 /etc/rc.d/rc.local), and add the following::
1381 modprobe bonding mode=balance-alb miimon=100
1383 ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
1384 ip link set eth0 master bond0
1385 ip link set eth1 master bond0
1387 Replace the example bonding module parameters and bond0
1388 network configuration (IP address, netmask, etc) with the appropriate
1389 values for your configuration.
1391 Unfortunately, this method will not provide support for the
1392 ifup and ifdown scripts on the bond devices. To reload the bonding
1393 configuration, it is necessary to run the initialization script, e.g.,::
1395 # /etc/init.d/boot.local
1399 # /etc/rc.d/rc.local
1401 It may be desirable in such a case to create a separate script
1402 which only initializes the bonding configuration, then call that
1403 separate script from within boot.local. This allows for bonding to be
1404 enabled without re-running the entire global init script.
1406 To shut down the bonding devices, it is necessary to first
1407 mark the bonding device itself as being down, then remove the
1408 appropriate device driver modules. For our example above, you can do
1411 # ifconfig bond0 down
1415 Again, for convenience, it may be desirable to create a script
1416 with these commands.
1419 3.3.1 Configuring Multiple Bonds Manually
1420 -----------------------------------------
1422 This section contains information on configuring multiple
1423 bonding devices with differing options for those systems whose network
1424 initialization scripts lack support for configuring multiple bonds.
1426 If you require multiple bonding devices, but all with the same
1427 options, you may wish to use the "max_bonds" module parameter,
1430 To create multiple bonding devices with differing options, it is
1431 preferable to use bonding parameters exported by sysfs, documented in the
1434 For versions of bonding without sysfs support, the only means to
1435 provide multiple instances of bonding with differing options is to load
1436 the bonding driver multiple times. Note that current versions of the
1437 sysconfig network initialization scripts handle this automatically; if
1438 your distro uses these scripts, no special action is needed. See the
1439 section Configuring Bonding Devices, above, if you're not sure about your
1440 network initialization scripts.
1442 To load multiple instances of the module, it is necessary to
1443 specify a different name for each instance (the module loading system
1444 requires that every loaded module, even multiple instances of the same
1445 module, have a unique name). This is accomplished by supplying multiple
1446 sets of bonding options in ``/etc/modprobe.d/*.conf``, for example::
1449 options bond0 -o bond0 mode=balance-rr miimon=100
1452 options bond1 -o bond1 mode=balance-alb miimon=50
1454 will load the bonding module two times. The first instance is
1455 named "bond0" and creates the bond0 device in balance-rr mode with an
1456 miimon of 100. The second instance is named "bond1" and creates the
1457 bond1 device in balance-alb mode with an miimon of 50.
1459 In some circumstances (typically with older distributions),
1460 the above does not work, and the second bonding instance never sees
1461 its options. In that case, the second options line can be substituted
1464 install bond1 /sbin/modprobe --ignore-install bonding -o bond1 \
1465 mode=balance-alb miimon=50
1467 This may be repeated any number of times, specifying a new and
1468 unique name in place of bond1 for each subsequent instance.
1470 It has been observed that some Red Hat supplied kernels are unable
1471 to rename modules at load time (the "-o bond1" part). Attempts to pass
1472 that option to modprobe will produce an "Operation not permitted" error.
1473 This has been reported on some Fedora Core kernels, and has been seen on
1474 RHEL 4 as well. On kernels exhibiting this problem, it will be impossible
1475 to configure multiple bonds with differing parameters (as they are older
1476 kernels, and also lack sysfs support).
1478 3.4 Configuring Bonding Manually via Sysfs
1479 ------------------------------------------
1481 Starting with version 3.0.0, Channel Bonding may be configured
1482 via the sysfs interface. This interface allows dynamic configuration
1483 of all bonds in the system without unloading the module. It also
1484 allows for adding and removing bonds at runtime. Ifenslave is no
1485 longer required, though it is still supported.
1487 Use of the sysfs interface allows you to use multiple bonds
1488 with different configurations without having to reload the module.
1489 It also allows you to use multiple, differently configured bonds when
1490 bonding is compiled into the kernel.
1492 You must have the sysfs filesystem mounted to configure
1493 bonding this way. The examples in this document assume that you
1494 are using the standard mount point for sysfs, e.g. /sys. If your
1495 sysfs filesystem is mounted elsewhere, you will need to adjust the
1496 example paths accordingly.
1498 Creating and Destroying Bonds
1499 -----------------------------
1500 To add a new bond foo::
1502 # echo +foo > /sys/class/net/bonding_masters
1504 To remove an existing bond bar::
1506 # echo -bar > /sys/class/net/bonding_masters
1508 To show all existing bonds::
1510 # cat /sys/class/net/bonding_masters
1514 due to 4K size limitation of sysfs files, this list may be
1515 truncated if you have more than a few hundred bonds. This is unlikely
1516 to occur under normal operating conditions.
1518 Adding and Removing Slaves
1519 --------------------------
1520 Interfaces may be enslaved to a bond using the file
1521 /sys/class/net/<bond>/bonding/slaves. The semantics for this file
1522 are the same as for the bonding_masters file.
1524 To enslave interface eth0 to bond bond0::
1527 # echo +eth0 > /sys/class/net/bond0/bonding/slaves
1529 To free slave eth0 from bond bond0::
1531 # echo -eth0 > /sys/class/net/bond0/bonding/slaves
1533 When an interface is enslaved to a bond, symlinks between the
1534 two are created in the sysfs filesystem. In this case, you would get
1535 /sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and
1536 /sys/class/net/eth0/master pointing to /sys/class/net/bond0.
1538 This means that you can tell quickly whether or not an
1539 interface is enslaved by looking for the master symlink. Thus:
1540 # echo -eth0 > /sys/class/net/eth0/master/bonding/slaves
1541 will free eth0 from whatever bond it is enslaved to, regardless of
1542 the name of the bond interface.
1544 Changing a Bond's Configuration
1545 -------------------------------
1546 Each bond may be configured individually by manipulating the
1547 files located in /sys/class/net/<bond name>/bonding
1549 The names of these files correspond directly with the command-
1550 line parameters described elsewhere in this file, and, with the
1551 exception of arp_ip_target, they accept the same values. To see the
1552 current setting, simply cat the appropriate file.
1554 A few examples will be given here; for specific usage
1555 guidelines for each parameter, see the appropriate section in this
1558 To configure bond0 for balance-alb mode::
1560 # ifconfig bond0 down
1561 # echo 6 > /sys/class/net/bond0/bonding/mode
1563 # echo balance-alb > /sys/class/net/bond0/bonding/mode
1567 The bond interface must be down before the mode can be changed.
1569 To enable MII monitoring on bond0 with a 1 second interval::
1571 # echo 1000 > /sys/class/net/bond0/bonding/miimon
1575 If ARP monitoring is enabled, it will disabled when MII
1576 monitoring is enabled, and vice-versa.
1578 To add ARP targets::
1580 # echo +192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1581 # echo +192.168.0.101 > /sys/class/net/bond0/bonding/arp_ip_target
1585 up to 16 target addresses may be specified.
1587 To remove an ARP target::
1589 # echo -192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1591 To configure the interval between learning packet transmits::
1593 # echo 12 > /sys/class/net/bond0/bonding/lp_interval
1597 the lp_interval is the number of seconds between instances where
1598 the bonding driver sends learning packets to each slaves peer switch. The
1599 default interval is 1 second.
1601 Example Configuration
1602 ---------------------
1603 We begin with the same example that is shown in section 3.3,
1604 executed with sysfs, and without using ifenslave.
1606 To make a simple bond of two e100 devices (presumed to be eth0
1607 and eth1), and have it persist across reboots, edit the appropriate
1608 file (/etc/init.d/boot.local or /etc/rc.d/rc.local), and add the
1613 echo balance-alb > /sys/class/net/bond0/bonding/mode
1614 ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
1615 echo 100 > /sys/class/net/bond0/bonding/miimon
1616 echo +eth0 > /sys/class/net/bond0/bonding/slaves
1617 echo +eth1 > /sys/class/net/bond0/bonding/slaves
1619 To add a second bond, with two e1000 interfaces in
1620 active-backup mode, using ARP monitoring, add the following lines to
1624 echo +bond1 > /sys/class/net/bonding_masters
1625 echo active-backup > /sys/class/net/bond1/bonding/mode
1626 ifconfig bond1 192.168.2.1 netmask 255.255.255.0 up
1627 echo +192.168.2.100 /sys/class/net/bond1/bonding/arp_ip_target
1628 echo 2000 > /sys/class/net/bond1/bonding/arp_interval
1629 echo +eth2 > /sys/class/net/bond1/bonding/slaves
1630 echo +eth3 > /sys/class/net/bond1/bonding/slaves
1632 3.5 Configuration with Interfaces Support
1633 -----------------------------------------
1635 This section applies to distros which use /etc/network/interfaces file
1636 to describe network interface configuration, most notably Debian and it's
1639 The ifup and ifdown commands on Debian don't support bonding out of
1640 the box. The ifenslave-2.6 package should be installed to provide bonding
1641 support. Once installed, this package will provide ``bond-*`` options
1642 to be used into /etc/network/interfaces.
1644 Note that ifenslave-2.6 package will load the bonding module and use
1645 the ifenslave command when appropriate.
1647 Example Configurations
1648 ----------------------
1650 In /etc/network/interfaces, the following stanza will configure bond0, in
1651 active-backup mode, with eth0 and eth1 as slaves::
1654 iface bond0 inet dhcp
1655 bond-slaves eth0 eth1
1656 bond-mode active-backup
1658 bond-primary eth0 eth1
1660 If the above configuration doesn't work, you might have a system using
1661 upstart for system startup. This is most notably true for recent
1662 Ubuntu versions. The following stanza in /etc/network/interfaces will
1663 produce the same result on those systems::
1666 iface bond0 inet dhcp
1668 bond-mode active-backup
1672 iface eth0 inet manual
1674 bond-primary eth0 eth1
1677 iface eth1 inet manual
1679 bond-primary eth0 eth1
1681 For a full list of ``bond-*`` supported options in /etc/network/interfaces and
1682 some more advanced examples tailored to you particular distros, see the files in
1683 /usr/share/doc/ifenslave-2.6.
1685 3.6 Overriding Configuration for Special Cases
1686 ----------------------------------------------
1688 When using the bonding driver, the physical port which transmits a frame is
1689 typically selected by the bonding driver, and is not relevant to the user or
1690 system administrator. The output port is simply selected using the policies of
1691 the selected bonding mode. On occasion however, it is helpful to direct certain
1692 classes of traffic to certain physical interfaces on output to implement
1693 slightly more complex policies. For example, to reach a web server over a
1694 bonded interface in which eth0 connects to a private network, while eth1
1695 connects via a public network, it may be desirous to bias the bond to send said
1696 traffic over eth0 first, using eth1 only as a fall back, while all other traffic
1697 can safely be sent over either interface. Such configurations may be achieved
1698 using the traffic control utilities inherent in linux.
1700 By default the bonding driver is multiqueue aware and 16 queues are created
1701 when the driver initializes (see Documentation/networking/multiqueue.rst
1702 for details). If more or less queues are desired the module parameter
1703 tx_queues can be used to change this value. There is no sysfs parameter
1704 available as the allocation is done at module init time.
1706 The output of the file /proc/net/bonding/bondX has changed so the output Queue
1707 ID is now printed for each slave::
1709 Bonding Mode: fault-tolerance (active-backup)
1711 Currently Active Slave: eth0
1713 MII Polling Interval (ms): 0
1717 Slave Interface: eth0
1719 Link Failure Count: 0
1720 Permanent HW addr: 00:1a:a0:12:8f:cb
1723 Slave Interface: eth1
1725 Link Failure Count: 0
1726 Permanent HW addr: 00:1a:a0:12:8f:cc
1729 The queue_id for a slave can be set using the command::
1731 # echo "eth1:2" > /sys/class/net/bond0/bonding/queue_id
1733 Any interface that needs a queue_id set should set it with multiple calls
1734 like the one above until proper priorities are set for all interfaces. On
1735 distributions that allow configuration via initscripts, multiple 'queue_id'
1736 arguments can be added to BONDING_OPTS to set all needed slave queues.
1738 These queue id's can be used in conjunction with the tc utility to configure
1739 a multiqueue qdisc and filters to bias certain traffic to transmit on certain
1740 slave devices. For instance, say we wanted, in the above configuration to
1741 force all traffic bound to 192.168.1.100 to use eth1 in the bond as its output
1742 device. The following commands would accomplish this::
1744 # tc qdisc add dev bond0 handle 1 root multiq
1746 # tc filter add dev bond0 protocol ip parent 1: prio 1 u32 match ip \
1747 dst 192.168.1.100 action skbedit queue_mapping 2
1749 These commands tell the kernel to attach a multiqueue queue discipline to the
1750 bond0 interface and filter traffic enqueued to it, such that packets with a dst
1751 ip of 192.168.1.100 have their output queue mapping value overwritten to 2.
1752 This value is then passed into the driver, causing the normal output path
1753 selection policy to be overridden, selecting instead qid 2, which maps to eth1.
1755 Note that qid values begin at 1. Qid 0 is reserved to initiate to the driver
1756 that normal output policy selection should take place. One benefit to simply
1757 leaving the qid for a slave to 0 is the multiqueue awareness in the bonding
1758 driver that is now present. This awareness allows tc filters to be placed on
1759 slave devices as well as bond devices and the bonding driver will simply act as
1760 a pass-through for selecting output queues on the slave device rather than
1761 output port selection.
1763 This feature first appeared in bonding driver version 3.7.0 and support for
1764 output slave selection was limited to round-robin and active-backup modes.
1766 3.7 Configuring LACP for 802.3ad mode in a more secure way
1767 ----------------------------------------------------------
1769 When using 802.3ad bonding mode, the Actor (host) and Partner (switch)
1770 exchange LACPDUs. These LACPDUs cannot be sniffed, because they are
1771 destined to link local mac addresses (which switches/bridges are not
1772 supposed to forward). However, most of the values are easily predictable
1773 or are simply the machine's MAC address (which is trivially known to all
1774 other hosts in the same L2). This implies that other machines in the L2
1775 domain can spoof LACPDU packets from other hosts to the switch and potentially
1776 cause mayhem by joining (from the point of view of the switch) another
1777 machine's aggregate, thus receiving a portion of that hosts incoming
1778 traffic and / or spoofing traffic from that machine themselves (potentially
1779 even successfully terminating some portion of flows). Though this is not
1780 a likely scenario, one could avoid this possibility by simply configuring
1781 few bonding parameters:
1783 (a) ad_actor_system : You can set a random mac-address that can be used for
1784 these LACPDU exchanges. The value can not be either NULL or Multicast.
1785 Also it's preferable to set the local-admin bit. Following shell code
1786 generates a random mac-address as described above::
1788 # sys_mac_addr=$(printf '%02x:%02x:%02x:%02x:%02x:%02x' \
1789 $(( (RANDOM & 0xFE) | 0x02 )) \
1790 $(( RANDOM & 0xFF )) \
1791 $(( RANDOM & 0xFF )) \
1792 $(( RANDOM & 0xFF )) \
1793 $(( RANDOM & 0xFF )) \
1794 $(( RANDOM & 0xFF )))
1795 # echo $sys_mac_addr > /sys/class/net/bond0/bonding/ad_actor_system
1797 (b) ad_actor_sys_prio : Randomize the system priority. The default value
1798 is 65535, but system can take the value from 1 - 65535. Following shell
1799 code generates random priority and sets it::
1801 # sys_prio=$(( 1 + RANDOM + RANDOM ))
1802 # echo $sys_prio > /sys/class/net/bond0/bonding/ad_actor_sys_prio
1804 (c) ad_user_port_key : Use the user portion of the port-key. The default
1805 keeps this empty. These are the upper 10 bits of the port-key and value
1806 ranges from 0 - 1023. Following shell code generates these 10 bits and
1809 # usr_port_key=$(( RANDOM & 0x3FF ))
1810 # echo $usr_port_key > /sys/class/net/bond0/bonding/ad_user_port_key
1813 4 Querying Bonding Configuration
1814 =================================
1816 4.1 Bonding Configuration
1817 -------------------------
1819 Each bonding device has a read-only file residing in the
1820 /proc/net/bonding directory. The file contents include information
1821 about the bonding configuration, options and state of each slave.
1823 For example, the contents of /proc/net/bonding/bond0 after the
1824 driver is loaded with parameters of mode=0 and miimon=1000 is
1825 generally as follows::
1827 Ethernet Channel Bonding Driver: 2.6.1 (October 29, 2004)
1828 Bonding Mode: load balancing (round-robin)
1829 Currently Active Slave: eth0
1831 MII Polling Interval (ms): 1000
1835 Slave Interface: eth1
1837 Link Failure Count: 1
1839 Slave Interface: eth0
1841 Link Failure Count: 1
1843 The precise format and contents will change depending upon the
1844 bonding configuration, state, and version of the bonding driver.
1846 4.2 Network configuration
1847 -------------------------
1849 The network configuration can be inspected using the ifconfig
1850 command. Bonding devices will have the MASTER flag set; Bonding slave
1851 devices will have the SLAVE flag set. The ifconfig output does not
1852 contain information on which slaves are associated with which masters.
1854 In the example below, the bond0 interface is the master
1855 (MASTER) while eth0 and eth1 are slaves (SLAVE). Notice all slaves of
1856 bond0 have the same MAC address (HWaddr) as bond0 for all modes except
1857 TLB and ALB that require a unique MAC address for each slave::
1860 bond0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1861 inet addr:XXX.XXX.XXX.YYY Bcast:XXX.XXX.XXX.255 Mask:255.255.252.0
1862 UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
1863 RX packets:7224794 errors:0 dropped:0 overruns:0 frame:0
1864 TX packets:3286647 errors:1 dropped:0 overruns:1 carrier:0
1865 collisions:0 txqueuelen:0
1867 eth0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1868 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1869 RX packets:3573025 errors:0 dropped:0 overruns:0 frame:0
1870 TX packets:1643167 errors:1 dropped:0 overruns:1 carrier:0
1871 collisions:0 txqueuelen:100
1872 Interrupt:10 Base address:0x1080
1874 eth1 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1875 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1876 RX packets:3651769 errors:0 dropped:0 overruns:0 frame:0
1877 TX packets:1643480 errors:0 dropped:0 overruns:0 carrier:0
1878 collisions:0 txqueuelen:100
1879 Interrupt:9 Base address:0x1400
1881 5. Switch Configuration
1882 =======================
1884 For this section, "switch" refers to whatever system the
1885 bonded devices are directly connected to (i.e., where the other end of
1886 the cable plugs into). This may be an actual dedicated switch device,
1887 or it may be another regular system (e.g., another computer running
1890 The active-backup, balance-tlb and balance-alb modes do not
1891 require any specific configuration of the switch.
1893 The 802.3ad mode requires that the switch have the appropriate
1894 ports configured as an 802.3ad aggregation. The precise method used
1895 to configure this varies from switch to switch, but, for example, a
1896 Cisco 3550 series switch requires that the appropriate ports first be
1897 grouped together in a single etherchannel instance, then that
1898 etherchannel is set to mode "lacp" to enable 802.3ad (instead of
1899 standard EtherChannel).
1901 The balance-rr, balance-xor and broadcast modes generally
1902 require that the switch have the appropriate ports grouped together.
1903 The nomenclature for such a group differs between switches, it may be
1904 called an "etherchannel" (as in the Cisco example, above), a "trunk
1905 group" or some other similar variation. For these modes, each switch
1906 will also have its own configuration options for the switch's transmit
1907 policy to the bond. Typical choices include XOR of either the MAC or
1908 IP addresses. The transmit policy of the two peers does not need to
1909 match. For these three modes, the bonding mode really selects a
1910 transmit policy for an EtherChannel group; all three will interoperate
1911 with another EtherChannel group.
1914 6. 802.1q VLAN Support
1915 ======================
1917 It is possible to configure VLAN devices over a bond interface
1918 using the 8021q driver. However, only packets coming from the 8021q
1919 driver and passing through bonding will be tagged by default. Self
1920 generated packets, for example, bonding's learning packets or ARP
1921 packets generated by either ALB mode or the ARP monitor mechanism, are
1922 tagged internally by bonding itself. As a result, bonding must
1923 "learn" the VLAN IDs configured above it, and use those IDs to tag
1924 self generated packets.
1926 For reasons of simplicity, and to support the use of adapters
1927 that can do VLAN hardware acceleration offloading, the bonding
1928 interface declares itself as fully hardware offloading capable, it gets
1929 the add_vid/kill_vid notifications to gather the necessary
1930 information, and it propagates those actions to the slaves. In case
1931 of mixed adapter types, hardware accelerated tagged packets that
1932 should go through an adapter that is not offloading capable are
1933 "un-accelerated" by the bonding driver so the VLAN tag sits in the
1936 VLAN interfaces *must* be added on top of a bonding interface
1937 only after enslaving at least one slave. The bonding interface has a
1938 hardware address of 00:00:00:00:00:00 until the first slave is added.
1939 If the VLAN interface is created prior to the first enslavement, it
1940 would pick up the all-zeroes hardware address. Once the first slave
1941 is attached to the bond, the bond device itself will pick up the
1942 slave's hardware address, which is then available for the VLAN device.
1944 Also, be aware that a similar problem can occur if all slaves
1945 are released from a bond that still has one or more VLAN interfaces on
1946 top of it. When a new slave is added, the bonding interface will
1947 obtain its hardware address from the first slave, which might not
1948 match the hardware address of the VLAN interfaces (which was
1949 ultimately copied from an earlier slave).
1951 There are two methods to insure that the VLAN device operates
1952 with the correct hardware address if all slaves are removed from a
1955 1. Remove all VLAN interfaces then recreate them
1957 2. Set the bonding interface's hardware address so that it
1958 matches the hardware address of the VLAN interfaces.
1960 Note that changing a VLAN interface's HW address would set the
1961 underlying device -- i.e. the bonding interface -- to promiscuous
1962 mode, which might not be what you want.
1968 The bonding driver at present supports two schemes for
1969 monitoring a slave device's link state: the ARP monitor and the MII
1972 At the present time, due to implementation restrictions in the
1973 bonding driver itself, it is not possible to enable both ARP and MII
1974 monitoring simultaneously.
1976 7.1 ARP Monitor Operation
1977 -------------------------
1979 The ARP monitor operates as its name suggests: it sends ARP
1980 queries to one or more designated peer systems on the network, and
1981 uses the response as an indication that the link is operating. This
1982 gives some assurance that traffic is actually flowing to and from one
1983 or more peers on the local network.
1985 The ARP monitor relies on the device driver itself to verify
1986 that traffic is flowing. In particular, the driver must keep up to
1987 date the last receive time, dev->last_rx. Drivers that use NETIF_F_LLTX
1988 flag must also update netdev_queue->trans_start. If they do not, then the
1989 ARP monitor will immediately fail any slaves using that driver, and
1990 those slaves will stay down. If networking monitoring (tcpdump, etc)
1991 shows the ARP requests and replies on the network, then it may be that
1992 your device driver is not updating last_rx and trans_start.
1994 7.2 Configuring Multiple ARP Targets
1995 ------------------------------------
1997 While ARP monitoring can be done with just one target, it can
1998 be useful in a High Availability setup to have several targets to
1999 monitor. In the case of just one target, the target itself may go
2000 down or have a problem making it unresponsive to ARP requests. Having
2001 an additional target (or several) increases the reliability of the ARP
2004 Multiple ARP targets must be separated by commas as follows::
2006 # example options for ARP monitoring with three targets
2008 options bond0 arp_interval=60 arp_ip_target=192.168.0.1,192.168.0.3,192.168.0.9
2010 For just a single target the options would resemble::
2012 # example options for ARP monitoring with one target
2014 options bond0 arp_interval=60 arp_ip_target=192.168.0.100
2017 7.3 MII Monitor Operation
2018 -------------------------
2020 The MII monitor monitors only the carrier state of the local
2021 network interface. It accomplishes this in one of three ways: by
2022 depending upon the device driver to maintain its carrier state, by
2023 querying the device's MII registers, or by making an ethtool query to
2026 If the use_carrier module parameter is 1 (the default value),
2027 then the MII monitor will rely on the driver for carrier state
2028 information (via the netif_carrier subsystem). As explained in the
2029 use_carrier parameter information, above, if the MII monitor fails to
2030 detect carrier loss on the device (e.g., when the cable is physically
2031 disconnected), it may be that the driver does not support
2034 If use_carrier is 0, then the MII monitor will first query the
2035 device's (via ioctl) MII registers and check the link state. If that
2036 request fails (not just that it returns carrier down), then the MII
2037 monitor will make an ethtool ETHTOOL_GLINK request to attempt to obtain
2038 the same information. If both methods fail (i.e., the driver either
2039 does not support or had some error in processing both the MII register
2040 and ethtool requests), then the MII monitor will assume the link is
2043 8. Potential Sources of Trouble
2044 ===============================
2046 8.1 Adventures in Routing
2047 -------------------------
2049 When bonding is configured, it is important that the slave
2050 devices not have routes that supersede routes of the master (or,
2051 generally, not have routes at all). For example, suppose the bonding
2052 device bond0 has two slaves, eth0 and eth1, and the routing table is
2055 Kernel IP routing table
2056 Destination Gateway Genmask Flags MSS Window irtt Iface
2057 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth0
2058 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth1
2059 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 bond0
2060 127.0.0.0 0.0.0.0 255.0.0.0 U 40 0 0 lo
2062 This routing configuration will likely still update the
2063 receive/transmit times in the driver (needed by the ARP monitor), but
2064 may bypass the bonding driver (because outgoing traffic to, in this
2065 case, another host on network 10 would use eth0 or eth1 before bond0).
2067 The ARP monitor (and ARP itself) may become confused by this
2068 configuration, because ARP requests (generated by the ARP monitor)
2069 will be sent on one interface (bond0), but the corresponding reply
2070 will arrive on a different interface (eth0). This reply looks to ARP
2071 as an unsolicited ARP reply (because ARP matches replies on an
2072 interface basis), and is discarded. The MII monitor is not affected
2073 by the state of the routing table.
2075 The solution here is simply to insure that slaves do not have
2076 routes of their own, and if for some reason they must, those routes do
2077 not supersede routes of their master. This should generally be the
2078 case, but unusual configurations or errant manual or automatic static
2079 route additions may cause trouble.
2081 8.2 Ethernet Device Renaming
2082 ----------------------------
2084 On systems with network configuration scripts that do not
2085 associate physical devices directly with network interface names (so
2086 that the same physical device always has the same "ethX" name), it may
2087 be necessary to add some special logic to config files in
2090 For example, given a modules.conf containing the following::
2093 options bond0 mode=some-mode miimon=50
2099 If neither eth0 and eth1 are slaves to bond0, then when the
2100 bond0 interface comes up, the devices may end up reordered. This
2101 happens because bonding is loaded first, then its slave device's
2102 drivers are loaded next. Since no other drivers have been loaded,
2103 when the e1000 driver loads, it will receive eth0 and eth1 for its
2104 devices, but the bonding configuration tries to enslave eth2 and eth3
2105 (which may later be assigned to the tg3 devices).
2107 Adding the following::
2109 add above bonding e1000 tg3
2111 causes modprobe to load e1000 then tg3, in that order, when
2112 bonding is loaded. This command is fully documented in the
2113 modules.conf manual page.
2115 On systems utilizing modprobe an equivalent problem can occur.
2116 In this case, the following can be added to config files in
2117 /etc/modprobe.d/ as::
2119 softdep bonding pre: tg3 e1000
2121 This will load tg3 and e1000 modules before loading the bonding one.
2122 Full documentation on this can be found in the modprobe.d and modprobe
2125 8.3. Painfully Slow Or No Failed Link Detection By Miimon
2126 ---------------------------------------------------------
2128 By default, bonding enables the use_carrier option, which
2129 instructs bonding to trust the driver to maintain carrier state.
2131 As discussed in the options section, above, some drivers do
2132 not support the netif_carrier_on/_off link state tracking system.
2133 With use_carrier enabled, bonding will always see these links as up,
2134 regardless of their actual state.
2136 Additionally, other drivers do support netif_carrier, but do
2137 not maintain it in real time, e.g., only polling the link state at
2138 some fixed interval. In this case, miimon will detect failures, but
2139 only after some long period of time has expired. If it appears that
2140 miimon is very slow in detecting link failures, try specifying
2141 use_carrier=0 to see if that improves the failure detection time. If
2142 it does, then it may be that the driver checks the carrier state at a
2143 fixed interval, but does not cache the MII register values (so the
2144 use_carrier=0 method of querying the registers directly works). If
2145 use_carrier=0 does not improve the failover, then the driver may cache
2146 the registers, or the problem may be elsewhere.
2148 Also, remember that miimon only checks for the device's
2149 carrier state. It has no way to determine the state of devices on or
2150 beyond other ports of a switch, or if a switch is refusing to pass
2151 traffic while still maintaining carrier on.
2156 If running SNMP agents, the bonding driver should be loaded
2157 before any network drivers participating in a bond. This requirement
2158 is due to the interface index (ipAdEntIfIndex) being associated to
2159 the first interface found with a given IP address. That is, there is
2160 only one ipAdEntIfIndex for each IP address. For example, if eth0 and
2161 eth1 are slaves of bond0 and the driver for eth0 is loaded before the
2162 bonding driver, the interface for the IP address will be associated
2163 with the eth0 interface. This configuration is shown below, the IP
2164 address 192.168.1.1 has an interface index of 2 which indexes to eth0
2165 in the ifDescr table (ifDescr.2).
2169 interfaces.ifTable.ifEntry.ifDescr.1 = lo
2170 interfaces.ifTable.ifEntry.ifDescr.2 = eth0
2171 interfaces.ifTable.ifEntry.ifDescr.3 = eth1
2172 interfaces.ifTable.ifEntry.ifDescr.4 = eth2
2173 interfaces.ifTable.ifEntry.ifDescr.5 = eth3
2174 interfaces.ifTable.ifEntry.ifDescr.6 = bond0
2175 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 5
2176 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
2177 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 4
2178 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
2180 This problem is avoided by loading the bonding driver before
2181 any network drivers participating in a bond. Below is an example of
2182 loading the bonding driver first, the IP address 192.168.1.1 is
2183 correctly associated with ifDescr.2.
2185 interfaces.ifTable.ifEntry.ifDescr.1 = lo
2186 interfaces.ifTable.ifEntry.ifDescr.2 = bond0
2187 interfaces.ifTable.ifEntry.ifDescr.3 = eth0
2188 interfaces.ifTable.ifEntry.ifDescr.4 = eth1
2189 interfaces.ifTable.ifEntry.ifDescr.5 = eth2
2190 interfaces.ifTable.ifEntry.ifDescr.6 = eth3
2191 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 6
2192 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
2193 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 5
2194 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
2196 While some distributions may not report the interface name in
2197 ifDescr, the association between the IP address and IfIndex remains
2198 and SNMP functions such as Interface_Scan_Next will report that
2201 10. Promiscuous mode
2202 ====================
2204 When running network monitoring tools, e.g., tcpdump, it is
2205 common to enable promiscuous mode on the device, so that all traffic
2206 is seen (instead of seeing only traffic destined for the local host).
2207 The bonding driver handles promiscuous mode changes to the bonding
2208 master device (e.g., bond0), and propagates the setting to the slave
2211 For the balance-rr, balance-xor, broadcast, and 802.3ad modes,
2212 the promiscuous mode setting is propagated to all slaves.
2214 For the active-backup, balance-tlb and balance-alb modes, the
2215 promiscuous mode setting is propagated only to the active slave.
2217 For balance-tlb mode, the active slave is the slave currently
2218 receiving inbound traffic.
2220 For balance-alb mode, the active slave is the slave used as a
2221 "primary." This slave is used for mode-specific control traffic, for
2222 sending to peers that are unassigned or if the load is unbalanced.
2224 For the active-backup, balance-tlb and balance-alb modes, when
2225 the active slave changes (e.g., due to a link failure), the
2226 promiscuous setting will be propagated to the new active slave.
2228 11. Configuring Bonding for High Availability
2229 =============================================
2231 High Availability refers to configurations that provide
2232 maximum network availability by having redundant or backup devices,
2233 links or switches between the host and the rest of the world. The
2234 goal is to provide the maximum availability of network connectivity
2235 (i.e., the network always works), even though other configurations
2236 could provide higher throughput.
2238 11.1 High Availability in a Single Switch Topology
2239 --------------------------------------------------
2241 If two hosts (or a host and a single switch) are directly
2242 connected via multiple physical links, then there is no availability
2243 penalty to optimizing for maximum bandwidth. In this case, there is
2244 only one switch (or peer), so if it fails, there is no alternative
2245 access to fail over to. Additionally, the bonding load balance modes
2246 support link monitoring of their members, so if individual links fail,
2247 the load will be rebalanced across the remaining devices.
2249 See Section 12, "Configuring Bonding for Maximum Throughput"
2250 for information on configuring bonding with one peer device.
2252 11.2 High Availability in a Multiple Switch Topology
2253 ----------------------------------------------------
2255 With multiple switches, the configuration of bonding and the
2256 network changes dramatically. In multiple switch topologies, there is
2257 a trade off between network availability and usable bandwidth.
2259 Below is a sample network, configured to maximize the
2260 availability of the network::
2264 +-----+----+ +-----+----+
2265 | |port2 ISL port2| |
2266 | switch A +--------------------------+ switch B |
2268 +-----+----+ +-----++---+
2271 +-------------+ host1 +---------------+
2274 In this configuration, there is a link between the two
2275 switches (ISL, or inter switch link), and multiple ports connecting to
2276 the outside world ("port3" on each switch). There is no technical
2277 reason that this could not be extended to a third switch.
2279 11.2.1 HA Bonding Mode Selection for Multiple Switch Topology
2280 -------------------------------------------------------------
2282 In a topology such as the example above, the active-backup and
2283 broadcast modes are the only useful bonding modes when optimizing for
2284 availability; the other modes require all links to terminate on the
2285 same peer for them to behave rationally.
2288 This is generally the preferred mode, particularly if
2289 the switches have an ISL and play together well. If the
2290 network configuration is such that one switch is specifically
2291 a backup switch (e.g., has lower capacity, higher cost, etc),
2292 then the primary option can be used to insure that the
2293 preferred link is always used when it is available.
2296 This mode is really a special purpose mode, and is suitable
2297 only for very specific needs. For example, if the two
2298 switches are not connected (no ISL), and the networks beyond
2299 them are totally independent. In this case, if it is
2300 necessary for some specific one-way traffic to reach both
2301 independent networks, then the broadcast mode may be suitable.
2303 11.2.2 HA Link Monitoring Selection for Multiple Switch Topology
2304 ----------------------------------------------------------------
2306 The choice of link monitoring ultimately depends upon your
2307 switch. If the switch can reliably fail ports in response to other
2308 failures, then either the MII or ARP monitors should work. For
2309 example, in the above example, if the "port3" link fails at the remote
2310 end, the MII monitor has no direct means to detect this. The ARP
2311 monitor could be configured with a target at the remote end of port3,
2312 thus detecting that failure without switch support.
2314 In general, however, in a multiple switch topology, the ARP
2315 monitor can provide a higher level of reliability in detecting end to
2316 end connectivity failures (which may be caused by the failure of any
2317 individual component to pass traffic for any reason). Additionally,
2318 the ARP monitor should be configured with multiple targets (at least
2319 one for each switch in the network). This will insure that,
2320 regardless of which switch is active, the ARP monitor has a suitable
2323 Note, also, that of late many switches now support a functionality
2324 generally referred to as "trunk failover." This is a feature of the
2325 switch that causes the link state of a particular switch port to be set
2326 down (or up) when the state of another switch port goes down (or up).
2327 Its purpose is to propagate link failures from logically "exterior" ports
2328 to the logically "interior" ports that bonding is able to monitor via
2329 miimon. Availability and configuration for trunk failover varies by
2330 switch, but this can be a viable alternative to the ARP monitor when using
2333 12. Configuring Bonding for Maximum Throughput
2334 ==============================================
2336 12.1 Maximizing Throughput in a Single Switch Topology
2337 ------------------------------------------------------
2339 In a single switch configuration, the best method to maximize
2340 throughput depends upon the application and network environment. The
2341 various load balancing modes each have strengths and weaknesses in
2342 different environments, as detailed below.
2344 For this discussion, we will break down the topologies into
2345 two categories. Depending upon the destination of most traffic, we
2346 categorize them into either "gatewayed" or "local" configurations.
2348 In a gatewayed configuration, the "switch" is acting primarily
2349 as a router, and the majority of traffic passes through this router to
2350 other networks. An example would be the following::
2353 +----------+ +----------+
2354 | |eth0 port1| | to other networks
2355 | Host A +---------------------+ router +------------------->
2356 | +---------------------+ | Hosts B and C are out
2357 | |eth1 port2| | here somewhere
2358 +----------+ +----------+
2360 The router may be a dedicated router device, or another host
2361 acting as a gateway. For our discussion, the important point is that
2362 the majority of traffic from Host A will pass through the router to
2363 some other network before reaching its final destination.
2365 In a gatewayed network configuration, although Host A may
2366 communicate with many other systems, all of its traffic will be sent
2367 and received via one other peer on the local network, the router.
2369 Note that the case of two systems connected directly via
2370 multiple physical links is, for purposes of configuring bonding, the
2371 same as a gatewayed configuration. In that case, it happens that all
2372 traffic is destined for the "gateway" itself, not some other network
2375 In a local configuration, the "switch" is acting primarily as
2376 a switch, and the majority of traffic passes through this switch to
2377 reach other stations on the same network. An example would be the
2380 +----------+ +----------+ +--------+
2381 | |eth0 port1| +-------+ Host B |
2382 | Host A +------------+ switch |port3 +--------+
2383 | +------------+ | +--------+
2384 | |eth1 port2| +------------------+ Host C |
2385 +----------+ +----------+port4 +--------+
2388 Again, the switch may be a dedicated switch device, or another
2389 host acting as a gateway. For our discussion, the important point is
2390 that the majority of traffic from Host A is destined for other hosts
2391 on the same local network (Hosts B and C in the above example).
2393 In summary, in a gatewayed configuration, traffic to and from
2394 the bonded device will be to the same MAC level peer on the network
2395 (the gateway itself, i.e., the router), regardless of its final
2396 destination. In a local configuration, traffic flows directly to and
2397 from the final destinations, thus, each destination (Host B, Host C)
2398 will be addressed directly by their individual MAC addresses.
2400 This distinction between a gatewayed and a local network
2401 configuration is important because many of the load balancing modes
2402 available use the MAC addresses of the local network source and
2403 destination to make load balancing decisions. The behavior of each
2404 mode is described below.
2407 12.1.1 MT Bonding Mode Selection for Single Switch Topology
2408 -----------------------------------------------------------
2410 This configuration is the easiest to set up and to understand,
2411 although you will have to decide which bonding mode best suits your
2412 needs. The trade offs for each mode are detailed below:
2415 This mode is the only mode that will permit a single
2416 TCP/IP connection to stripe traffic across multiple
2417 interfaces. It is therefore the only mode that will allow a
2418 single TCP/IP stream to utilize more than one interface's
2419 worth of throughput. This comes at a cost, however: the
2420 striping generally results in peer systems receiving packets out
2421 of order, causing TCP/IP's congestion control system to kick
2422 in, often by retransmitting segments.
2424 It is possible to adjust TCP/IP's congestion limits by
2425 altering the net.ipv4.tcp_reordering sysctl parameter. The
2426 usual default value is 3. But keep in mind TCP stack is able
2427 to automatically increase this when it detects reorders.
2429 Note that the fraction of packets that will be delivered out of
2430 order is highly variable, and is unlikely to be zero. The level
2431 of reordering depends upon a variety of factors, including the
2432 networking interfaces, the switch, and the topology of the
2433 configuration. Speaking in general terms, higher speed network
2434 cards produce more reordering (due to factors such as packet
2435 coalescing), and a "many to many" topology will reorder at a
2436 higher rate than a "many slow to one fast" configuration.
2438 Many switches do not support any modes that stripe traffic
2439 (instead choosing a port based upon IP or MAC level addresses);
2440 for those devices, traffic for a particular connection flowing
2441 through the switch to a balance-rr bond will not utilize greater
2442 than one interface's worth of bandwidth.
2444 If you are utilizing protocols other than TCP/IP, UDP for
2445 example, and your application can tolerate out of order
2446 delivery, then this mode can allow for single stream datagram
2447 performance that scales near linearly as interfaces are added
2450 This mode requires the switch to have the appropriate ports
2451 configured for "etherchannel" or "trunking."
2454 There is not much advantage in this network topology to
2455 the active-backup mode, as the inactive backup devices are all
2456 connected to the same peer as the primary. In this case, a
2457 load balancing mode (with link monitoring) will provide the
2458 same level of network availability, but with increased
2459 available bandwidth. On the plus side, active-backup mode
2460 does not require any configuration of the switch, so it may
2461 have value if the hardware available does not support any of
2462 the load balance modes.
2465 This mode will limit traffic such that packets destined
2466 for specific peers will always be sent over the same
2467 interface. Since the destination is determined by the MAC
2468 addresses involved, this mode works best in a "local" network
2469 configuration (as described above), with destinations all on
2470 the same local network. This mode is likely to be suboptimal
2471 if all your traffic is passed through a single router (i.e., a
2472 "gatewayed" network configuration, as described above).
2474 As with balance-rr, the switch ports need to be configured for
2475 "etherchannel" or "trunking."
2478 Like active-backup, there is not much advantage to this
2479 mode in this type of network topology.
2482 This mode can be a good choice for this type of network
2483 topology. The 802.3ad mode is an IEEE standard, so all peers
2484 that implement 802.3ad should interoperate well. The 802.3ad
2485 protocol includes automatic configuration of the aggregates,
2486 so minimal manual configuration of the switch is needed
2487 (typically only to designate that some set of devices is
2488 available for 802.3ad). The 802.3ad standard also mandates
2489 that frames be delivered in order (within certain limits), so
2490 in general single connections will not see misordering of
2491 packets. The 802.3ad mode does have some drawbacks: the
2492 standard mandates that all devices in the aggregate operate at
2493 the same speed and duplex. Also, as with all bonding load
2494 balance modes other than balance-rr, no single connection will
2495 be able to utilize more than a single interface's worth of
2498 Additionally, the linux bonding 802.3ad implementation
2499 distributes traffic by peer (using an XOR of MAC addresses
2500 and packet type ID), so in a "gatewayed" configuration, all
2501 outgoing traffic will generally use the same device. Incoming
2502 traffic may also end up on a single device, but that is
2503 dependent upon the balancing policy of the peer's 802.3ad
2504 implementation. In a "local" configuration, traffic will be
2505 distributed across the devices in the bond.
2507 Finally, the 802.3ad mode mandates the use of the MII monitor,
2508 therefore, the ARP monitor is not available in this mode.
2511 The balance-tlb mode balances outgoing traffic by peer.
2512 Since the balancing is done according to MAC address, in a
2513 "gatewayed" configuration (as described above), this mode will
2514 send all traffic across a single device. However, in a
2515 "local" network configuration, this mode balances multiple
2516 local network peers across devices in a vaguely intelligent
2517 manner (not a simple XOR as in balance-xor or 802.3ad mode),
2518 so that mathematically unlucky MAC addresses (i.e., ones that
2519 XOR to the same value) will not all "bunch up" on a single
2522 Unlike 802.3ad, interfaces may be of differing speeds, and no
2523 special switch configuration is required. On the down side,
2524 in this mode all incoming traffic arrives over a single
2525 interface, this mode requires certain ethtool support in the
2526 network device driver of the slave interfaces, and the ARP
2527 monitor is not available.
2530 This mode is everything that balance-tlb is, and more.
2531 It has all of the features (and restrictions) of balance-tlb,
2532 and will also balance incoming traffic from local network
2533 peers (as described in the Bonding Module Options section,
2536 The only additional down side to this mode is that the network
2537 device driver must support changing the hardware address while
2540 12.1.2 MT Link Monitoring for Single Switch Topology
2541 ----------------------------------------------------
2543 The choice of link monitoring may largely depend upon which
2544 mode you choose to use. The more advanced load balancing modes do not
2545 support the use of the ARP monitor, and are thus restricted to using
2546 the MII monitor (which does not provide as high a level of end to end
2547 assurance as the ARP monitor).
2549 12.2 Maximum Throughput in a Multiple Switch Topology
2550 -----------------------------------------------------
2552 Multiple switches may be utilized to optimize for throughput
2553 when they are configured in parallel as part of an isolated network
2554 between two or more systems, for example::
2560 +--------+ | +---------+
2562 +------+---+ +-----+----+ +-----+----+
2563 | Switch A | | Switch B | | Switch C |
2564 +------+---+ +-----+----+ +-----+----+
2566 +--------+ | +---------+
2572 In this configuration, the switches are isolated from one
2573 another. One reason to employ a topology such as this is for an
2574 isolated network with many hosts (a cluster configured for high
2575 performance, for example), using multiple smaller switches can be more
2576 cost effective than a single larger switch, e.g., on a network with 24
2577 hosts, three 24 port switches can be significantly less expensive than
2578 a single 72 port switch.
2580 If access beyond the network is required, an individual host
2581 can be equipped with an additional network device connected to an
2582 external network; this host then additionally acts as a gateway.
2584 12.2.1 MT Bonding Mode Selection for Multiple Switch Topology
2585 -------------------------------------------------------------
2587 In actual practice, the bonding mode typically employed in
2588 configurations of this type is balance-rr. Historically, in this
2589 network configuration, the usual caveats about out of order packet
2590 delivery are mitigated by the use of network adapters that do not do
2591 any kind of packet coalescing (via the use of NAPI, or because the
2592 device itself does not generate interrupts until some number of
2593 packets has arrived). When employed in this fashion, the balance-rr
2594 mode allows individual connections between two hosts to effectively
2595 utilize greater than one interface's bandwidth.
2597 12.2.2 MT Link Monitoring for Multiple Switch Topology
2598 ------------------------------------------------------
2600 Again, in actual practice, the MII monitor is most often used
2601 in this configuration, as performance is given preference over
2602 availability. The ARP monitor will function in this topology, but its
2603 advantages over the MII monitor are mitigated by the volume of probes
2604 needed as the number of systems involved grows (remember that each
2605 host in the network is configured with bonding).
2607 13. Switch Behavior Issues
2608 ==========================
2610 13.1 Link Establishment and Failover Delays
2611 -------------------------------------------
2613 Some switches exhibit undesirable behavior with regard to the
2614 timing of link up and down reporting by the switch.
2616 First, when a link comes up, some switches may indicate that
2617 the link is up (carrier available), but not pass traffic over the
2618 interface for some period of time. This delay is typically due to
2619 some type of autonegotiation or routing protocol, but may also occur
2620 during switch initialization (e.g., during recovery after a switch
2621 failure). If you find this to be a problem, specify an appropriate
2622 value to the updelay bonding module option to delay the use of the
2623 relevant interface(s).
2625 Second, some switches may "bounce" the link state one or more
2626 times while a link is changing state. This occurs most commonly while
2627 the switch is initializing. Again, an appropriate updelay value may
2630 Note that when a bonding interface has no active links, the
2631 driver will immediately reuse the first link that goes up, even if the
2632 updelay parameter has been specified (the updelay is ignored in this
2633 case). If there are slave interfaces waiting for the updelay timeout
2634 to expire, the interface that first went into that state will be
2635 immediately reused. This reduces down time of the network if the
2636 value of updelay has been overestimated, and since this occurs only in
2637 cases with no connectivity, there is no additional penalty for
2638 ignoring the updelay.
2640 In addition to the concerns about switch timings, if your
2641 switches take a long time to go into backup mode, it may be desirable
2642 to not activate a backup interface immediately after a link goes down.
2643 Failover may be delayed via the downdelay bonding module option.
2645 13.2 Duplicated Incoming Packets
2646 --------------------------------
2648 NOTE: Starting with version 3.0.2, the bonding driver has logic to
2649 suppress duplicate packets, which should largely eliminate this problem.
2650 The following description is kept for reference.
2652 It is not uncommon to observe a short burst of duplicated
2653 traffic when the bonding device is first used, or after it has been
2654 idle for some period of time. This is most easily observed by issuing
2655 a "ping" to some other host on the network, and noticing that the
2656 output from ping flags duplicates (typically one per slave).
2658 For example, on a bond in active-backup mode with five slaves
2659 all connected to one switch, the output may appear as follows::
2662 PING 10.0.4.2 (10.0.4.2) from 10.0.3.10 : 56(84) bytes of data.
2663 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.7 ms
2664 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2665 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2666 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2667 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2668 64 bytes from 10.0.4.2: icmp_seq=2 ttl=64 time=0.216 ms
2669 64 bytes from 10.0.4.2: icmp_seq=3 ttl=64 time=0.267 ms
2670 64 bytes from 10.0.4.2: icmp_seq=4 ttl=64 time=0.222 ms
2672 This is not due to an error in the bonding driver, rather, it
2673 is a side effect of how many switches update their MAC forwarding
2674 tables. Initially, the switch does not associate the MAC address in
2675 the packet with a particular switch port, and so it may send the
2676 traffic to all ports until its MAC forwarding table is updated. Since
2677 the interfaces attached to the bond may occupy multiple ports on a
2678 single switch, when the switch (temporarily) floods the traffic to all
2679 ports, the bond device receives multiple copies of the same packet
2680 (one per slave device).
2682 The duplicated packet behavior is switch dependent, some
2683 switches exhibit this, and some do not. On switches that display this
2684 behavior, it can be induced by clearing the MAC forwarding table (on
2685 most Cisco switches, the privileged command "clear mac address-table
2686 dynamic" will accomplish this).
2688 14. Hardware Specific Considerations
2689 ====================================
2691 This section contains additional information for configuring
2692 bonding on specific hardware platforms, or for interfacing bonding
2693 with particular switches or other devices.
2695 14.1 IBM BladeCenter
2696 --------------------
2698 This applies to the JS20 and similar systems.
2700 On the JS20 blades, the bonding driver supports only
2701 balance-rr, active-backup, balance-tlb and balance-alb modes. This is
2702 largely due to the network topology inside the BladeCenter, detailed
2705 JS20 network adapter information
2706 --------------------------------
2708 All JS20s come with two Broadcom Gigabit Ethernet ports
2709 integrated on the planar (that's "motherboard" in IBM-speak). In the
2710 BladeCenter chassis, the eth0 port of all JS20 blades is hard wired to
2711 I/O Module #1; similarly, all eth1 ports are wired to I/O Module #2.
2712 An add-on Broadcom daughter card can be installed on a JS20 to provide
2713 two more Gigabit Ethernet ports. These ports, eth2 and eth3, are
2714 wired to I/O Modules 3 and 4, respectively.
2716 Each I/O Module may contain either a switch or a passthrough
2717 module (which allows ports to be directly connected to an external
2718 switch). Some bonding modes require a specific BladeCenter internal
2719 network topology in order to function; these are detailed below.
2721 Additional BladeCenter-specific networking information can be
2722 found in two IBM Redbooks (www.ibm.com/redbooks):
2724 - "IBM eServer BladeCenter Networking Options"
2725 - "IBM eServer BladeCenter Layer 2-7 Network Switching"
2727 BladeCenter networking configuration
2728 ------------------------------------
2730 Because a BladeCenter can be configured in a very large number
2731 of ways, this discussion will be confined to describing basic
2734 Normally, Ethernet Switch Modules (ESMs) are used in I/O
2735 modules 1 and 2. In this configuration, the eth0 and eth1 ports of a
2736 JS20 will be connected to different internal switches (in the
2737 respective I/O modules).
2739 A passthrough module (OPM or CPM, optical or copper,
2740 passthrough module) connects the I/O module directly to an external
2741 switch. By using PMs in I/O module #1 and #2, the eth0 and eth1
2742 interfaces of a JS20 can be redirected to the outside world and
2743 connected to a common external switch.
2745 Depending upon the mix of ESMs and PMs, the network will
2746 appear to bonding as either a single switch topology (all PMs) or as a
2747 multiple switch topology (one or more ESMs, zero or more PMs). It is
2748 also possible to connect ESMs together, resulting in a configuration
2749 much like the example in "High Availability in a Multiple Switch
2752 Requirements for specific modes
2753 -------------------------------
2755 The balance-rr mode requires the use of passthrough modules
2756 for devices in the bond, all connected to an common external switch.
2757 That switch must be configured for "etherchannel" or "trunking" on the
2758 appropriate ports, as is usual for balance-rr.
2760 The balance-alb and balance-tlb modes will function with
2761 either switch modules or passthrough modules (or a mix). The only
2762 specific requirement for these modes is that all network interfaces
2763 must be able to reach all destinations for traffic sent over the
2764 bonding device (i.e., the network must converge at some point outside
2767 The active-backup mode has no additional requirements.
2769 Link monitoring issues
2770 ----------------------
2772 When an Ethernet Switch Module is in place, only the ARP
2773 monitor will reliably detect link loss to an external switch. This is
2774 nothing unusual, but examination of the BladeCenter cabinet would
2775 suggest that the "external" network ports are the ethernet ports for
2776 the system, when it fact there is a switch between these "external"
2777 ports and the devices on the JS20 system itself. The MII monitor is
2778 only able to detect link failures between the ESM and the JS20 system.
2780 When a passthrough module is in place, the MII monitor does
2781 detect failures to the "external" port, which is then directly
2782 connected to the JS20 system.
2787 The Serial Over LAN (SoL) link is established over the primary
2788 ethernet (eth0) only, therefore, any loss of link to eth0 will result
2789 in losing your SoL connection. It will not fail over with other
2790 network traffic, as the SoL system is beyond the control of the
2793 It may be desirable to disable spanning tree on the switch
2794 (either the internal Ethernet Switch Module, or an external switch) to
2795 avoid fail-over delay issues when using bonding.
2798 15. Frequently Asked Questions
2799 ==============================
2804 Yes. The old 2.0.xx channel bonding patch was not SMP safe.
2805 The new driver was designed to be SMP safe from the start.
2807 2. What type of cards will work with it?
2808 -----------------------------------------
2810 Any Ethernet type cards (you can even mix cards - a Intel
2811 EtherExpress PRO/100 and a 3com 3c905b, for example). For most modes,
2812 devices need not be of the same speed.
2814 Starting with version 3.2.1, bonding also supports Infiniband
2815 slaves in active-backup mode.
2817 3. How many bonding devices can I have?
2818 ----------------------------------------
2822 4. How many slaves can a bonding device have?
2823 ----------------------------------------------
2825 This is limited only by the number of network interfaces Linux
2826 supports and/or the number of network cards you can place in your
2829 5. What happens when a slave link dies?
2830 ----------------------------------------
2832 If link monitoring is enabled, then the failing device will be
2833 disabled. The active-backup mode will fail over to a backup link, and
2834 other modes will ignore the failed link. The link will continue to be
2835 monitored, and should it recover, it will rejoin the bond (in whatever
2836 manner is appropriate for the mode). See the sections on High
2837 Availability and the documentation for each mode for additional
2840 Link monitoring can be enabled via either the miimon or
2841 arp_interval parameters (described in the module parameters section,
2842 above). In general, miimon monitors the carrier state as sensed by
2843 the underlying network device, and the arp monitor (arp_interval)
2844 monitors connectivity to another host on the local network.
2846 If no link monitoring is configured, the bonding driver will
2847 be unable to detect link failures, and will assume that all links are
2848 always available. This will likely result in lost packets, and a
2849 resulting degradation of performance. The precise performance loss
2850 depends upon the bonding mode and network configuration.
2852 6. Can bonding be used for High Availability?
2853 ----------------------------------------------
2855 Yes. See the section on High Availability for details.
2857 7. Which switches/systems does it work with?
2858 ---------------------------------------------
2860 The full answer to this depends upon the desired mode.
2862 In the basic balance modes (balance-rr and balance-xor), it
2863 works with any system that supports etherchannel (also called
2864 trunking). Most managed switches currently available have such
2865 support, and many unmanaged switches as well.
2867 The advanced balance modes (balance-tlb and balance-alb) do
2868 not have special switch requirements, but do need device drivers that
2869 support specific features (described in the appropriate section under
2870 module parameters, above).
2872 In 802.3ad mode, it works with systems that support IEEE
2873 802.3ad Dynamic Link Aggregation. Most managed and many unmanaged
2874 switches currently available support 802.3ad.
2876 The active-backup mode should work with any Layer-II switch.
2878 8. Where does a bonding device get its MAC address from?
2879 ---------------------------------------------------------
2881 When using slave devices that have fixed MAC addresses, or when
2882 the fail_over_mac option is enabled, the bonding device's MAC address is
2883 the MAC address of the active slave.
2885 For other configurations, if not explicitly configured (with
2886 ifconfig or ip link), the MAC address of the bonding device is taken from
2887 its first slave device. This MAC address is then passed to all following
2888 slaves and remains persistent (even if the first slave is removed) until
2889 the bonding device is brought down or reconfigured.
2891 If you wish to change the MAC address, you can set it with
2892 ifconfig or ip link::
2894 # ifconfig bond0 hw ether 00:11:22:33:44:55
2896 # ip link set bond0 address 66:77:88:99:aa:bb
2898 The MAC address can be also changed by bringing down/up the
2899 device and then changing its slaves (or their order)::
2901 # ifconfig bond0 down ; modprobe -r bonding
2902 # ifconfig bond0 .... up
2903 # ifenslave bond0 eth...
2905 This method will automatically take the address from the next
2906 slave that is added.
2908 To restore your slaves' MAC addresses, you need to detach them
2909 from the bond (``ifenslave -d bond0 eth0``). The bonding driver will
2910 then restore the MAC addresses that the slaves had before they were
2913 16. Resources and Links
2914 =======================
2916 The latest version of the bonding driver can be found in the latest
2917 version of the linux kernel, found on http://kernel.org
2919 The latest version of this document can be found in the latest kernel
2920 source (named Documentation/networking/bonding.rst).
2922 Discussions regarding the development of the bonding driver take place
2923 on the main Linux network mailing list, hosted at vger.kernel.org. The list
2926 netdev@vger.kernel.org
2928 The administrative interface (to subscribe or unsubscribe) can
2931 http://vger.kernel.org/vger-lists.html#netdev