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
785 A string (eth0, eth2, etc) specifying which slave is the
786 primary device. The specified device will always be the
787 active slave while it is available. Only when the primary is
788 off-line will alternate devices be used. This is useful when
789 one slave is preferred over another, e.g., when one slave has
790 higher throughput than another.
792 The primary option is only valid for active-backup(1),
793 balance-tlb (5) and balance-alb (6) mode.
797 Specifies the reselection policy for the primary slave. This
798 affects how the primary slave is chosen to become the active slave
799 when failure of the active slave or recovery of the primary slave
800 occurs. This option is designed to prevent flip-flopping between
801 the primary slave and other slaves. Possible values are:
803 always or 0 (default)
805 The primary slave becomes the active slave whenever it
810 The primary slave becomes the active slave when it comes
811 back up, if the speed and duplex of the primary slave is
812 better than the speed and duplex of the current active
817 The primary slave becomes the active slave only if the
818 current active slave fails and the primary slave is up.
820 The primary_reselect setting is ignored in two cases:
822 If no slaves are active, the first slave to recover is
823 made the active slave.
825 When initially enslaved, the primary slave is always made
828 Changing the primary_reselect policy via sysfs will cause an
829 immediate selection of the best active slave according to the new
830 policy. This may or may not result in a change of the active
831 slave, depending upon the circumstances.
833 This option was added for bonding version 3.6.0.
837 Specifies if dynamic shuffling of flows is enabled in tlb
838 mode. The value has no effect on any other modes.
840 The default behavior of tlb mode is to shuffle active flows across
841 slaves based on the load in that interval. This gives nice lb
842 characteristics but can cause packet reordering. If re-ordering is
843 a concern use this variable to disable flow shuffling and rely on
844 load balancing provided solely by the hash distribution.
845 xmit-hash-policy can be used to select the appropriate hashing for
848 The sysfs entry can be used to change the setting per bond device
849 and the initial value is derived from the module parameter. The
850 sysfs entry is allowed to be changed only if the bond device is
853 The default value is "1" that enables flow shuffling while value "0"
854 disables it. This option was added in bonding driver 3.7.1
859 Specifies the time, in milliseconds, to wait before enabling a
860 slave after a link recovery has been detected. This option is
861 only valid for the miimon link monitor. The updelay value
862 should be a multiple of the miimon value; if not, it will be
863 rounded down to the nearest multiple. The default value is 0.
867 Specifies whether or not miimon should use MII or ETHTOOL
868 ioctls vs. netif_carrier_ok() to determine the link
869 status. The MII or ETHTOOL ioctls are less efficient and
870 utilize a deprecated calling sequence within the kernel. The
871 netif_carrier_ok() relies on the device driver to maintain its
872 state with netif_carrier_on/off; at this writing, most, but
873 not all, device drivers support this facility.
875 If bonding insists that the link is up when it should not be,
876 it may be that your network device driver does not support
877 netif_carrier_on/off. The default state for netif_carrier is
878 "carrier on," so if a driver does not support netif_carrier,
879 it will appear as if the link is always up. In this case,
880 setting use_carrier to 0 will cause bonding to revert to the
881 MII / ETHTOOL ioctl method to determine the link state.
883 A value of 1 enables the use of netif_carrier_ok(), a value of
884 0 will use the deprecated MII / ETHTOOL ioctls. The default
889 Selects the transmit hash policy to use for slave selection in
890 balance-xor, 802.3ad, and tlb modes. Possible values are:
894 Uses XOR of hardware MAC addresses and packet type ID
895 field to generate the hash. The formula is
897 hash = source MAC[5] XOR destination MAC[5] XOR packet type ID
898 slave number = hash modulo slave count
900 This algorithm will place all traffic to a particular
901 network peer on the same slave.
903 This algorithm is 802.3ad compliant.
907 This policy uses a combination of layer2 and layer3
908 protocol information to generate the hash.
910 Uses XOR of hardware MAC addresses and IP addresses to
911 generate the hash. The formula is
913 hash = source MAC[5] XOR destination MAC[5] XOR packet type ID
914 hash = hash XOR source IP XOR destination IP
915 hash = hash XOR (hash RSHIFT 16)
916 hash = hash XOR (hash RSHIFT 8)
917 And then hash is reduced modulo slave count.
919 If the protocol is IPv6 then the source and destination
920 addresses are first hashed using ipv6_addr_hash.
922 This algorithm will place all traffic to a particular
923 network peer on the same slave. For non-IP traffic,
924 the formula is the same as for the layer2 transmit
927 This policy is intended to provide a more balanced
928 distribution of traffic than layer2 alone, especially
929 in environments where a layer3 gateway device is
930 required to reach most destinations.
932 This algorithm is 802.3ad compliant.
936 This policy uses upper layer protocol information,
937 when available, to generate the hash. This allows for
938 traffic to a particular network peer to span multiple
939 slaves, although a single connection will not span
942 The formula for unfragmented TCP and UDP packets is
944 hash = source port, destination port (as in the header)
945 hash = hash XOR source IP XOR destination IP
946 hash = hash XOR (hash RSHIFT 16)
947 hash = hash XOR (hash RSHIFT 8)
948 And then hash is reduced modulo slave count.
950 If the protocol is IPv6 then the source and destination
951 addresses are first hashed using ipv6_addr_hash.
953 For fragmented TCP or UDP packets and all other IPv4 and
954 IPv6 protocol traffic, the source and destination port
955 information is omitted. For non-IP traffic, the
956 formula is the same as for the layer2 transmit hash
959 This algorithm is not fully 802.3ad compliant. A
960 single TCP or UDP conversation containing both
961 fragmented and unfragmented packets will see packets
962 striped across two interfaces. This may result in out
963 of order delivery. Most traffic types will not meet
964 this criteria, as TCP rarely fragments traffic, and
965 most UDP traffic is not involved in extended
966 conversations. Other implementations of 802.3ad may
967 or may not tolerate this noncompliance.
971 This policy uses the same formula as layer2+3 but it
972 relies on skb_flow_dissect to obtain the header fields
973 which might result in the use of inner headers if an
974 encapsulation protocol is used. For example this will
975 improve the performance for tunnel users because the
976 packets will be distributed according to the encapsulated
981 This policy uses the same formula as layer3+4 but it
982 relies on skb_flow_dissect to obtain the header fields
983 which might result in the use of inner headers if an
984 encapsulation protocol is used. For example this will
985 improve the performance for tunnel users because the
986 packets will be distributed according to the encapsulated
991 This policy uses a very rudimentary vlan ID and source mac
992 hash to load-balance traffic per-vlan, with failover
993 should one leg fail. The intended use case is for a bond
994 shared by multiple virtual machines, all configured to
995 use their own vlan, to give lacp-like functionality
996 without requiring lacp-capable switching hardware.
998 The formula for the hash is simply
1000 hash = (vlan ID) XOR (source MAC vendor) XOR (source MAC dev)
1002 The default value is layer2. This option was added in bonding
1003 version 2.6.3. In earlier versions of bonding, this parameter
1004 does not exist, and the layer2 policy is the only policy. The
1005 layer2+3 value was added for bonding version 3.2.2.
1009 Specifies the number of IGMP membership reports to be issued after
1010 a failover event. One membership report is issued immediately after
1011 the failover, subsequent packets are sent in each 200ms interval.
1013 The valid range is 0 - 255; the default value is 1. A value of 0
1014 prevents the IGMP membership report from being issued in response
1015 to the failover event.
1017 This option is useful for bonding modes balance-rr (0), active-backup
1018 (1), balance-tlb (5) and balance-alb (6), in which a failover can
1019 switch the IGMP traffic from one slave to another. Therefore a fresh
1020 IGMP report must be issued to cause the switch to forward the incoming
1021 IGMP traffic over the newly selected slave.
1023 This option was added for bonding version 3.7.0.
1027 Specifies the number of seconds between instances where the bonding
1028 driver sends learning packets to each slaves peer switch.
1030 The valid range is 1 - 0x7fffffff; the default value is 1. This Option
1031 has effect only in balance-tlb and balance-alb modes.
1033 3. Configuring Bonding Devices
1034 ==============================
1036 You can configure bonding using either your distro's network
1037 initialization scripts, or manually using either iproute2 or the
1038 sysfs interface. Distros generally use one of three packages for the
1039 network initialization scripts: initscripts, sysconfig or interfaces.
1040 Recent versions of these packages have support for bonding, while older
1043 We will first describe the options for configuring bonding for
1044 distros using versions of initscripts, sysconfig and interfaces with full
1045 or partial support for bonding, then provide information on enabling
1046 bonding without support from the network initialization scripts (i.e.,
1047 older versions of initscripts or sysconfig).
1049 If you're unsure whether your distro uses sysconfig,
1050 initscripts or interfaces, or don't know if it's new enough, have no fear.
1051 Determining this is fairly straightforward.
1053 First, look for a file called interfaces in /etc/network directory.
1054 If this file is present in your system, then your system use interfaces. See
1055 Configuration with Interfaces Support.
1057 Else, issue the command::
1059 $ rpm -qf /sbin/ifup
1061 It will respond with a line of text starting with either
1062 "initscripts" or "sysconfig," followed by some numbers. This is the
1063 package that provides your network initialization scripts.
1065 Next, to determine if your installation supports bonding,
1068 $ grep ifenslave /sbin/ifup
1070 If this returns any matches, then your initscripts or
1071 sysconfig has support for bonding.
1073 3.1 Configuration with Sysconfig Support
1074 ----------------------------------------
1076 This section applies to distros using a version of sysconfig
1077 with bonding support, for example, SuSE Linux Enterprise Server 9.
1079 SuSE SLES 9's networking configuration system does support
1080 bonding, however, at this writing, the YaST system configuration
1081 front end does not provide any means to work with bonding devices.
1082 Bonding devices can be managed by hand, however, as follows.
1084 First, if they have not already been configured, configure the
1085 slave devices. On SLES 9, this is most easily done by running the
1086 yast2 sysconfig configuration utility. The goal is for to create an
1087 ifcfg-id file for each slave device. The simplest way to accomplish
1088 this is to configure the devices for DHCP (this is only to get the
1089 file ifcfg-id file created; see below for some issues with DHCP). The
1090 name of the configuration file for each device will be of the form::
1092 ifcfg-id-xx:xx:xx:xx:xx:xx
1094 Where the "xx" portion will be replaced with the digits from
1095 the device's permanent MAC address.
1097 Once the set of ifcfg-id-xx:xx:xx:xx:xx:xx files has been
1098 created, it is necessary to edit the configuration files for the slave
1099 devices (the MAC addresses correspond to those of the slave devices).
1100 Before editing, the file will contain multiple lines, and will look
1101 something like this::
1106 UNIQUE='XNzu.WeZGOGF+4wE'
1107 _nm_name='bus-pci-0001:61:01.0'
1109 Change the BOOTPROTO and STARTMODE lines to the following::
1114 Do not alter the UNIQUE or _nm_name lines. Remove any other
1115 lines (USERCTL, etc).
1117 Once the ifcfg-id-xx:xx:xx:xx:xx:xx files have been modified,
1118 it's time to create the configuration file for the bonding device
1119 itself. This file is named ifcfg-bondX, where X is the number of the
1120 bonding device to create, starting at 0. The first such file is
1121 ifcfg-bond0, the second is ifcfg-bond1, and so on. The sysconfig
1122 network configuration system will correctly start multiple instances
1125 The contents of the ifcfg-bondX file is as follows::
1128 BROADCAST="10.0.2.255"
1130 NETMASK="255.255.0.0"
1134 BONDING_MASTER="yes"
1135 BONDING_MODULE_OPTS="mode=active-backup miimon=100"
1136 BONDING_SLAVE0="eth0"
1137 BONDING_SLAVE1="bus-pci-0000:06:08.1"
1139 Replace the sample BROADCAST, IPADDR, NETMASK and NETWORK
1140 values with the appropriate values for your network.
1142 The STARTMODE specifies when the device is brought online.
1143 The possible values are:
1145 ======== ======================================================
1146 onboot The device is started at boot time. If you're not
1147 sure, this is probably what you want.
1149 manual The device is started only when ifup is called
1150 manually. Bonding devices may be configured this
1151 way if you do not wish them to start automatically
1152 at boot for some reason.
1154 hotplug The device is started by a hotplug event. This is not
1155 a valid choice for a bonding device.
1157 off or The device configuration is ignored.
1159 ======== ======================================================
1161 The line BONDING_MASTER='yes' indicates that the device is a
1162 bonding master device. The only useful value is "yes."
1164 The contents of BONDING_MODULE_OPTS are supplied to the
1165 instance of the bonding module for this device. Specify the options
1166 for the bonding mode, link monitoring, and so on here. Do not include
1167 the max_bonds bonding parameter; this will confuse the configuration
1168 system if you have multiple bonding devices.
1170 Finally, supply one BONDING_SLAVEn="slave device" for each
1171 slave. where "n" is an increasing value, one for each slave. The
1172 "slave device" is either an interface name, e.g., "eth0", or a device
1173 specifier for the network device. The interface name is easier to
1174 find, but the ethN names are subject to change at boot time if, e.g.,
1175 a device early in the sequence has failed. The device specifiers
1176 (bus-pci-0000:06:08.1 in the example above) specify the physical
1177 network device, and will not change unless the device's bus location
1178 changes (for example, it is moved from one PCI slot to another). The
1179 example above uses one of each type for demonstration purposes; most
1180 configurations will choose one or the other for all slave devices.
1182 When all configuration files have been modified or created,
1183 networking must be restarted for the configuration changes to take
1184 effect. This can be accomplished via the following::
1186 # /etc/init.d/network restart
1188 Note that the network control script (/sbin/ifdown) will
1189 remove the bonding module as part of the network shutdown processing,
1190 so it is not necessary to remove the module by hand if, e.g., the
1191 module parameters have changed.
1193 Also, at this writing, YaST/YaST2 will not manage bonding
1194 devices (they do not show bonding interfaces on its list of network
1195 devices). It is necessary to edit the configuration file by hand to
1196 change the bonding configuration.
1198 Additional general options and details of the ifcfg file
1199 format can be found in an example ifcfg template file::
1201 /etc/sysconfig/network/ifcfg.template
1203 Note that the template does not document the various ``BONDING_*``
1204 settings described above, but does describe many of the other options.
1206 3.1.1 Using DHCP with Sysconfig
1207 -------------------------------
1209 Under sysconfig, configuring a device with BOOTPROTO='dhcp'
1210 will cause it to query DHCP for its IP address information. At this
1211 writing, this does not function for bonding devices; the scripts
1212 attempt to obtain the device address from DHCP prior to adding any of
1213 the slave devices. Without active slaves, the DHCP requests are not
1214 sent to the network.
1216 3.1.2 Configuring Multiple Bonds with Sysconfig
1217 -----------------------------------------------
1219 The sysconfig network initialization system is capable of
1220 handling multiple bonding devices. All that is necessary is for each
1221 bonding instance to have an appropriately configured ifcfg-bondX file
1222 (as described above). Do not specify the "max_bonds" parameter to any
1223 instance of bonding, as this will confuse sysconfig. If you require
1224 multiple bonding devices with identical parameters, create multiple
1227 Because the sysconfig scripts supply the bonding module
1228 options in the ifcfg-bondX file, it is not necessary to add them to
1229 the system ``/etc/modules.d/*.conf`` configuration files.
1231 3.2 Configuration with Initscripts Support
1232 ------------------------------------------
1234 This section applies to distros using a recent version of
1235 initscripts with bonding support, for example, Red Hat Enterprise Linux
1236 version 3 or later, Fedora, etc. On these systems, the network
1237 initialization scripts have knowledge of bonding, and can be configured to
1238 control bonding devices. Note that older versions of the initscripts
1239 package have lower levels of support for bonding; this will be noted where
1242 These distros will not automatically load the network adapter
1243 driver unless the ethX device is configured with an IP address.
1244 Because of this constraint, users must manually configure a
1245 network-script file for all physical adapters that will be members of
1246 a bondX link. Network script files are located in the directory:
1248 /etc/sysconfig/network-scripts
1250 The file name must be prefixed with "ifcfg-eth" and suffixed
1251 with the adapter's physical adapter number. For example, the script
1252 for eth0 would be named /etc/sysconfig/network-scripts/ifcfg-eth0.
1253 Place the following text in the file::
1262 The DEVICE= line will be different for every ethX device and
1263 must correspond with the name of the file, i.e., ifcfg-eth1 must have
1264 a device line of DEVICE=eth1. The setting of the MASTER= line will
1265 also depend on the final bonding interface name chosen for your bond.
1266 As with other network devices, these typically start at 0, and go up
1267 one for each device, i.e., the first bonding instance is bond0, the
1268 second is bond1, and so on.
1270 Next, create a bond network script. The file name for this
1271 script will be /etc/sysconfig/network-scripts/ifcfg-bondX where X is
1272 the number of the bond. For bond0 the file is named "ifcfg-bond0",
1273 for bond1 it is named "ifcfg-bond1", and so on. Within that file,
1274 place the following text::
1278 NETMASK=255.255.255.0
1280 BROADCAST=192.168.1.255
1285 Be sure to change the networking specific lines (IPADDR,
1286 NETMASK, NETWORK and BROADCAST) to match your network configuration.
1288 For later versions of initscripts, such as that found with Fedora
1289 7 (or later) and Red Hat Enterprise Linux version 5 (or later), it is possible,
1290 and, indeed, preferable, to specify the bonding options in the ifcfg-bond0
1291 file, e.g. a line of the format::
1293 BONDING_OPTS="mode=active-backup arp_interval=60 arp_ip_target=192.168.1.254"
1295 will configure the bond with the specified options. The options
1296 specified in BONDING_OPTS are identical to the bonding module parameters
1297 except for the arp_ip_target field when using versions of initscripts older
1298 than and 8.57 (Fedora 8) and 8.45.19 (Red Hat Enterprise Linux 5.2). When
1299 using older versions each target should be included as a separate option and
1300 should be preceded by a '+' to indicate it should be added to the list of
1301 queried targets, e.g.,::
1303 arp_ip_target=+192.168.1.1 arp_ip_target=+192.168.1.2
1305 is the proper syntax to specify multiple targets. When specifying
1306 options via BONDING_OPTS, it is not necessary to edit
1307 ``/etc/modprobe.d/*.conf``.
1309 For even older versions of initscripts that do not support
1310 BONDING_OPTS, it is necessary to edit /etc/modprobe.d/*.conf, depending upon
1311 your distro) to load the bonding module with your desired options when the
1312 bond0 interface is brought up. The following lines in /etc/modprobe.d/*.conf
1313 will load the bonding module, and select its options:
1316 options bond0 mode=balance-alb miimon=100
1318 Replace the sample parameters with the appropriate set of
1319 options for your configuration.
1321 Finally run "/etc/rc.d/init.d/network restart" as root. This
1322 will restart the networking subsystem and your bond link should be now
1325 3.2.1 Using DHCP with Initscripts
1326 ---------------------------------
1328 Recent versions of initscripts (the versions supplied with Fedora
1329 Core 3 and Red Hat Enterprise Linux 4, or later versions, are reported to
1330 work) have support for assigning IP information to bonding devices via
1333 To configure bonding for DHCP, configure it as described
1334 above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp"
1335 and add a line consisting of "TYPE=Bonding". Note that the TYPE value
1338 3.2.2 Configuring Multiple Bonds with Initscripts
1339 -------------------------------------------------
1341 Initscripts packages that are included with Fedora 7 and Red Hat
1342 Enterprise Linux 5 support multiple bonding interfaces by simply
1343 specifying the appropriate BONDING_OPTS= in ifcfg-bondX where X is the
1344 number of the bond. This support requires sysfs support in the kernel,
1345 and a bonding driver of version 3.0.0 or later. Other configurations may
1346 not support this method for specifying multiple bonding interfaces; for
1347 those instances, see the "Configuring Multiple Bonds Manually" section,
1350 3.3 Configuring Bonding Manually with iproute2
1351 -----------------------------------------------
1353 This section applies to distros whose network initialization
1354 scripts (the sysconfig or initscripts package) do not have specific
1355 knowledge of bonding. One such distro is SuSE Linux Enterprise Server
1358 The general method for these systems is to place the bonding
1359 module parameters into a config file in /etc/modprobe.d/ (as
1360 appropriate for the installed distro), then add modprobe and/or
1361 `ip link` commands to the system's global init script. The name of
1362 the global init script differs; for sysconfig, it is
1363 /etc/init.d/boot.local and for initscripts it is /etc/rc.d/rc.local.
1365 For example, if you wanted to make a simple bond of two e100
1366 devices (presumed to be eth0 and eth1), and have it persist across
1367 reboots, edit the appropriate file (/etc/init.d/boot.local or
1368 /etc/rc.d/rc.local), and add the following::
1370 modprobe bonding mode=balance-alb miimon=100
1372 ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
1373 ip link set eth0 master bond0
1374 ip link set eth1 master bond0
1376 Replace the example bonding module parameters and bond0
1377 network configuration (IP address, netmask, etc) with the appropriate
1378 values for your configuration.
1380 Unfortunately, this method will not provide support for the
1381 ifup and ifdown scripts on the bond devices. To reload the bonding
1382 configuration, it is necessary to run the initialization script, e.g.,::
1384 # /etc/init.d/boot.local
1388 # /etc/rc.d/rc.local
1390 It may be desirable in such a case to create a separate script
1391 which only initializes the bonding configuration, then call that
1392 separate script from within boot.local. This allows for bonding to be
1393 enabled without re-running the entire global init script.
1395 To shut down the bonding devices, it is necessary to first
1396 mark the bonding device itself as being down, then remove the
1397 appropriate device driver modules. For our example above, you can do
1400 # ifconfig bond0 down
1404 Again, for convenience, it may be desirable to create a script
1405 with these commands.
1408 3.3.1 Configuring Multiple Bonds Manually
1409 -----------------------------------------
1411 This section contains information on configuring multiple
1412 bonding devices with differing options for those systems whose network
1413 initialization scripts lack support for configuring multiple bonds.
1415 If you require multiple bonding devices, but all with the same
1416 options, you may wish to use the "max_bonds" module parameter,
1419 To create multiple bonding devices with differing options, it is
1420 preferable to use bonding parameters exported by sysfs, documented in the
1423 For versions of bonding without sysfs support, the only means to
1424 provide multiple instances of bonding with differing options is to load
1425 the bonding driver multiple times. Note that current versions of the
1426 sysconfig network initialization scripts handle this automatically; if
1427 your distro uses these scripts, no special action is needed. See the
1428 section Configuring Bonding Devices, above, if you're not sure about your
1429 network initialization scripts.
1431 To load multiple instances of the module, it is necessary to
1432 specify a different name for each instance (the module loading system
1433 requires that every loaded module, even multiple instances of the same
1434 module, have a unique name). This is accomplished by supplying multiple
1435 sets of bonding options in ``/etc/modprobe.d/*.conf``, for example::
1438 options bond0 -o bond0 mode=balance-rr miimon=100
1441 options bond1 -o bond1 mode=balance-alb miimon=50
1443 will load the bonding module two times. The first instance is
1444 named "bond0" and creates the bond0 device in balance-rr mode with an
1445 miimon of 100. The second instance is named "bond1" and creates the
1446 bond1 device in balance-alb mode with an miimon of 50.
1448 In some circumstances (typically with older distributions),
1449 the above does not work, and the second bonding instance never sees
1450 its options. In that case, the second options line can be substituted
1453 install bond1 /sbin/modprobe --ignore-install bonding -o bond1 \
1454 mode=balance-alb miimon=50
1456 This may be repeated any number of times, specifying a new and
1457 unique name in place of bond1 for each subsequent instance.
1459 It has been observed that some Red Hat supplied kernels are unable
1460 to rename modules at load time (the "-o bond1" part). Attempts to pass
1461 that option to modprobe will produce an "Operation not permitted" error.
1462 This has been reported on some Fedora Core kernels, and has been seen on
1463 RHEL 4 as well. On kernels exhibiting this problem, it will be impossible
1464 to configure multiple bonds with differing parameters (as they are older
1465 kernels, and also lack sysfs support).
1467 3.4 Configuring Bonding Manually via Sysfs
1468 ------------------------------------------
1470 Starting with version 3.0.0, Channel Bonding may be configured
1471 via the sysfs interface. This interface allows dynamic configuration
1472 of all bonds in the system without unloading the module. It also
1473 allows for adding and removing bonds at runtime. Ifenslave is no
1474 longer required, though it is still supported.
1476 Use of the sysfs interface allows you to use multiple bonds
1477 with different configurations without having to reload the module.
1478 It also allows you to use multiple, differently configured bonds when
1479 bonding is compiled into the kernel.
1481 You must have the sysfs filesystem mounted to configure
1482 bonding this way. The examples in this document assume that you
1483 are using the standard mount point for sysfs, e.g. /sys. If your
1484 sysfs filesystem is mounted elsewhere, you will need to adjust the
1485 example paths accordingly.
1487 Creating and Destroying Bonds
1488 -----------------------------
1489 To add a new bond foo::
1491 # echo +foo > /sys/class/net/bonding_masters
1493 To remove an existing bond bar::
1495 # echo -bar > /sys/class/net/bonding_masters
1497 To show all existing bonds::
1499 # cat /sys/class/net/bonding_masters
1503 due to 4K size limitation of sysfs files, this list may be
1504 truncated if you have more than a few hundred bonds. This is unlikely
1505 to occur under normal operating conditions.
1507 Adding and Removing Slaves
1508 --------------------------
1509 Interfaces may be enslaved to a bond using the file
1510 /sys/class/net/<bond>/bonding/slaves. The semantics for this file
1511 are the same as for the bonding_masters file.
1513 To enslave interface eth0 to bond bond0::
1516 # echo +eth0 > /sys/class/net/bond0/bonding/slaves
1518 To free slave eth0 from bond bond0::
1520 # echo -eth0 > /sys/class/net/bond0/bonding/slaves
1522 When an interface is enslaved to a bond, symlinks between the
1523 two are created in the sysfs filesystem. In this case, you would get
1524 /sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and
1525 /sys/class/net/eth0/master pointing to /sys/class/net/bond0.
1527 This means that you can tell quickly whether or not an
1528 interface is enslaved by looking for the master symlink. Thus:
1529 # echo -eth0 > /sys/class/net/eth0/master/bonding/slaves
1530 will free eth0 from whatever bond it is enslaved to, regardless of
1531 the name of the bond interface.
1533 Changing a Bond's Configuration
1534 -------------------------------
1535 Each bond may be configured individually by manipulating the
1536 files located in /sys/class/net/<bond name>/bonding
1538 The names of these files correspond directly with the command-
1539 line parameters described elsewhere in this file, and, with the
1540 exception of arp_ip_target, they accept the same values. To see the
1541 current setting, simply cat the appropriate file.
1543 A few examples will be given here; for specific usage
1544 guidelines for each parameter, see the appropriate section in this
1547 To configure bond0 for balance-alb mode::
1549 # ifconfig bond0 down
1550 # echo 6 > /sys/class/net/bond0/bonding/mode
1552 # echo balance-alb > /sys/class/net/bond0/bonding/mode
1556 The bond interface must be down before the mode can be changed.
1558 To enable MII monitoring on bond0 with a 1 second interval::
1560 # echo 1000 > /sys/class/net/bond0/bonding/miimon
1564 If ARP monitoring is enabled, it will disabled when MII
1565 monitoring is enabled, and vice-versa.
1567 To add ARP targets::
1569 # echo +192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1570 # echo +192.168.0.101 > /sys/class/net/bond0/bonding/arp_ip_target
1574 up to 16 target addresses may be specified.
1576 To remove an ARP target::
1578 # echo -192.168.0.100 > /sys/class/net/bond0/bonding/arp_ip_target
1580 To configure the interval between learning packet transmits::
1582 # echo 12 > /sys/class/net/bond0/bonding/lp_interval
1586 the lp_interval is the number of seconds between instances where
1587 the bonding driver sends learning packets to each slaves peer switch. The
1588 default interval is 1 second.
1590 Example Configuration
1591 ---------------------
1592 We begin with the same example that is shown in section 3.3,
1593 executed with sysfs, and without using ifenslave.
1595 To make a simple bond of two e100 devices (presumed to be eth0
1596 and eth1), and have it persist across reboots, edit the appropriate
1597 file (/etc/init.d/boot.local or /etc/rc.d/rc.local), and add the
1602 echo balance-alb > /sys/class/net/bond0/bonding/mode
1603 ifconfig bond0 192.168.1.1 netmask 255.255.255.0 up
1604 echo 100 > /sys/class/net/bond0/bonding/miimon
1605 echo +eth0 > /sys/class/net/bond0/bonding/slaves
1606 echo +eth1 > /sys/class/net/bond0/bonding/slaves
1608 To add a second bond, with two e1000 interfaces in
1609 active-backup mode, using ARP monitoring, add the following lines to
1613 echo +bond1 > /sys/class/net/bonding_masters
1614 echo active-backup > /sys/class/net/bond1/bonding/mode
1615 ifconfig bond1 192.168.2.1 netmask 255.255.255.0 up
1616 echo +192.168.2.100 /sys/class/net/bond1/bonding/arp_ip_target
1617 echo 2000 > /sys/class/net/bond1/bonding/arp_interval
1618 echo +eth2 > /sys/class/net/bond1/bonding/slaves
1619 echo +eth3 > /sys/class/net/bond1/bonding/slaves
1621 3.5 Configuration with Interfaces Support
1622 -----------------------------------------
1624 This section applies to distros which use /etc/network/interfaces file
1625 to describe network interface configuration, most notably Debian and it's
1628 The ifup and ifdown commands on Debian don't support bonding out of
1629 the box. The ifenslave-2.6 package should be installed to provide bonding
1630 support. Once installed, this package will provide ``bond-*`` options
1631 to be used into /etc/network/interfaces.
1633 Note that ifenslave-2.6 package will load the bonding module and use
1634 the ifenslave command when appropriate.
1636 Example Configurations
1637 ----------------------
1639 In /etc/network/interfaces, the following stanza will configure bond0, in
1640 active-backup mode, with eth0 and eth1 as slaves::
1643 iface bond0 inet dhcp
1644 bond-slaves eth0 eth1
1645 bond-mode active-backup
1647 bond-primary eth0 eth1
1649 If the above configuration doesn't work, you might have a system using
1650 upstart for system startup. This is most notably true for recent
1651 Ubuntu versions. The following stanza in /etc/network/interfaces will
1652 produce the same result on those systems::
1655 iface bond0 inet dhcp
1657 bond-mode active-backup
1661 iface eth0 inet manual
1663 bond-primary eth0 eth1
1666 iface eth1 inet manual
1668 bond-primary eth0 eth1
1670 For a full list of ``bond-*`` supported options in /etc/network/interfaces and
1671 some more advanced examples tailored to you particular distros, see the files in
1672 /usr/share/doc/ifenslave-2.6.
1674 3.6 Overriding Configuration for Special Cases
1675 ----------------------------------------------
1677 When using the bonding driver, the physical port which transmits a frame is
1678 typically selected by the bonding driver, and is not relevant to the user or
1679 system administrator. The output port is simply selected using the policies of
1680 the selected bonding mode. On occasion however, it is helpful to direct certain
1681 classes of traffic to certain physical interfaces on output to implement
1682 slightly more complex policies. For example, to reach a web server over a
1683 bonded interface in which eth0 connects to a private network, while eth1
1684 connects via a public network, it may be desirous to bias the bond to send said
1685 traffic over eth0 first, using eth1 only as a fall back, while all other traffic
1686 can safely be sent over either interface. Such configurations may be achieved
1687 using the traffic control utilities inherent in linux.
1689 By default the bonding driver is multiqueue aware and 16 queues are created
1690 when the driver initializes (see Documentation/networking/multiqueue.rst
1691 for details). If more or less queues are desired the module parameter
1692 tx_queues can be used to change this value. There is no sysfs parameter
1693 available as the allocation is done at module init time.
1695 The output of the file /proc/net/bonding/bondX has changed so the output Queue
1696 ID is now printed for each slave::
1698 Bonding Mode: fault-tolerance (active-backup)
1700 Currently Active Slave: eth0
1702 MII Polling Interval (ms): 0
1706 Slave Interface: eth0
1708 Link Failure Count: 0
1709 Permanent HW addr: 00:1a:a0:12:8f:cb
1712 Slave Interface: eth1
1714 Link Failure Count: 0
1715 Permanent HW addr: 00:1a:a0:12:8f:cc
1718 The queue_id for a slave can be set using the command::
1720 # echo "eth1:2" > /sys/class/net/bond0/bonding/queue_id
1722 Any interface that needs a queue_id set should set it with multiple calls
1723 like the one above until proper priorities are set for all interfaces. On
1724 distributions that allow configuration via initscripts, multiple 'queue_id'
1725 arguments can be added to BONDING_OPTS to set all needed slave queues.
1727 These queue id's can be used in conjunction with the tc utility to configure
1728 a multiqueue qdisc and filters to bias certain traffic to transmit on certain
1729 slave devices. For instance, say we wanted, in the above configuration to
1730 force all traffic bound to 192.168.1.100 to use eth1 in the bond as its output
1731 device. The following commands would accomplish this::
1733 # tc qdisc add dev bond0 handle 1 root multiq
1735 # tc filter add dev bond0 protocol ip parent 1: prio 1 u32 match ip \
1736 dst 192.168.1.100 action skbedit queue_mapping 2
1738 These commands tell the kernel to attach a multiqueue queue discipline to the
1739 bond0 interface and filter traffic enqueued to it, such that packets with a dst
1740 ip of 192.168.1.100 have their output queue mapping value overwritten to 2.
1741 This value is then passed into the driver, causing the normal output path
1742 selection policy to be overridden, selecting instead qid 2, which maps to eth1.
1744 Note that qid values begin at 1. Qid 0 is reserved to initiate to the driver
1745 that normal output policy selection should take place. One benefit to simply
1746 leaving the qid for a slave to 0 is the multiqueue awareness in the bonding
1747 driver that is now present. This awareness allows tc filters to be placed on
1748 slave devices as well as bond devices and the bonding driver will simply act as
1749 a pass-through for selecting output queues on the slave device rather than
1750 output port selection.
1752 This feature first appeared in bonding driver version 3.7.0 and support for
1753 output slave selection was limited to round-robin and active-backup modes.
1755 3.7 Configuring LACP for 802.3ad mode in a more secure way
1756 ----------------------------------------------------------
1758 When using 802.3ad bonding mode, the Actor (host) and Partner (switch)
1759 exchange LACPDUs. These LACPDUs cannot be sniffed, because they are
1760 destined to link local mac addresses (which switches/bridges are not
1761 supposed to forward). However, most of the values are easily predictable
1762 or are simply the machine's MAC address (which is trivially known to all
1763 other hosts in the same L2). This implies that other machines in the L2
1764 domain can spoof LACPDU packets from other hosts to the switch and potentially
1765 cause mayhem by joining (from the point of view of the switch) another
1766 machine's aggregate, thus receiving a portion of that hosts incoming
1767 traffic and / or spoofing traffic from that machine themselves (potentially
1768 even successfully terminating some portion of flows). Though this is not
1769 a likely scenario, one could avoid this possibility by simply configuring
1770 few bonding parameters:
1772 (a) ad_actor_system : You can set a random mac-address that can be used for
1773 these LACPDU exchanges. The value can not be either NULL or Multicast.
1774 Also it's preferable to set the local-admin bit. Following shell code
1775 generates a random mac-address as described above::
1777 # sys_mac_addr=$(printf '%02x:%02x:%02x:%02x:%02x:%02x' \
1778 $(( (RANDOM & 0xFE) | 0x02 )) \
1779 $(( RANDOM & 0xFF )) \
1780 $(( RANDOM & 0xFF )) \
1781 $(( RANDOM & 0xFF )) \
1782 $(( RANDOM & 0xFF )) \
1783 $(( RANDOM & 0xFF )))
1784 # echo $sys_mac_addr > /sys/class/net/bond0/bonding/ad_actor_system
1786 (b) ad_actor_sys_prio : Randomize the system priority. The default value
1787 is 65535, but system can take the value from 1 - 65535. Following shell
1788 code generates random priority and sets it::
1790 # sys_prio=$(( 1 + RANDOM + RANDOM ))
1791 # echo $sys_prio > /sys/class/net/bond0/bonding/ad_actor_sys_prio
1793 (c) ad_user_port_key : Use the user portion of the port-key. The default
1794 keeps this empty. These are the upper 10 bits of the port-key and value
1795 ranges from 0 - 1023. Following shell code generates these 10 bits and
1798 # usr_port_key=$(( RANDOM & 0x3FF ))
1799 # echo $usr_port_key > /sys/class/net/bond0/bonding/ad_user_port_key
1802 4 Querying Bonding Configuration
1803 =================================
1805 4.1 Bonding Configuration
1806 -------------------------
1808 Each bonding device has a read-only file residing in the
1809 /proc/net/bonding directory. The file contents include information
1810 about the bonding configuration, options and state of each slave.
1812 For example, the contents of /proc/net/bonding/bond0 after the
1813 driver is loaded with parameters of mode=0 and miimon=1000 is
1814 generally as follows::
1816 Ethernet Channel Bonding Driver: 2.6.1 (October 29, 2004)
1817 Bonding Mode: load balancing (round-robin)
1818 Currently Active Slave: eth0
1820 MII Polling Interval (ms): 1000
1824 Slave Interface: eth1
1826 Link Failure Count: 1
1828 Slave Interface: eth0
1830 Link Failure Count: 1
1832 The precise format and contents will change depending upon the
1833 bonding configuration, state, and version of the bonding driver.
1835 4.2 Network configuration
1836 -------------------------
1838 The network configuration can be inspected using the ifconfig
1839 command. Bonding devices will have the MASTER flag set; Bonding slave
1840 devices will have the SLAVE flag set. The ifconfig output does not
1841 contain information on which slaves are associated with which masters.
1843 In the example below, the bond0 interface is the master
1844 (MASTER) while eth0 and eth1 are slaves (SLAVE). Notice all slaves of
1845 bond0 have the same MAC address (HWaddr) as bond0 for all modes except
1846 TLB and ALB that require a unique MAC address for each slave::
1849 bond0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1850 inet addr:XXX.XXX.XXX.YYY Bcast:XXX.XXX.XXX.255 Mask:255.255.252.0
1851 UP BROADCAST RUNNING MASTER MULTICAST MTU:1500 Metric:1
1852 RX packets:7224794 errors:0 dropped:0 overruns:0 frame:0
1853 TX packets:3286647 errors:1 dropped:0 overruns:1 carrier:0
1854 collisions:0 txqueuelen:0
1856 eth0 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1857 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1858 RX packets:3573025 errors:0 dropped:0 overruns:0 frame:0
1859 TX packets:1643167 errors:1 dropped:0 overruns:1 carrier:0
1860 collisions:0 txqueuelen:100
1861 Interrupt:10 Base address:0x1080
1863 eth1 Link encap:Ethernet HWaddr 00:C0:F0:1F:37:B4
1864 UP BROADCAST RUNNING SLAVE MULTICAST MTU:1500 Metric:1
1865 RX packets:3651769 errors:0 dropped:0 overruns:0 frame:0
1866 TX packets:1643480 errors:0 dropped:0 overruns:0 carrier:0
1867 collisions:0 txqueuelen:100
1868 Interrupt:9 Base address:0x1400
1870 5. Switch Configuration
1871 =======================
1873 For this section, "switch" refers to whatever system the
1874 bonded devices are directly connected to (i.e., where the other end of
1875 the cable plugs into). This may be an actual dedicated switch device,
1876 or it may be another regular system (e.g., another computer running
1879 The active-backup, balance-tlb and balance-alb modes do not
1880 require any specific configuration of the switch.
1882 The 802.3ad mode requires that the switch have the appropriate
1883 ports configured as an 802.3ad aggregation. The precise method used
1884 to configure this varies from switch to switch, but, for example, a
1885 Cisco 3550 series switch requires that the appropriate ports first be
1886 grouped together in a single etherchannel instance, then that
1887 etherchannel is set to mode "lacp" to enable 802.3ad (instead of
1888 standard EtherChannel).
1890 The balance-rr, balance-xor and broadcast modes generally
1891 require that the switch have the appropriate ports grouped together.
1892 The nomenclature for such a group differs between switches, it may be
1893 called an "etherchannel" (as in the Cisco example, above), a "trunk
1894 group" or some other similar variation. For these modes, each switch
1895 will also have its own configuration options for the switch's transmit
1896 policy to the bond. Typical choices include XOR of either the MAC or
1897 IP addresses. The transmit policy of the two peers does not need to
1898 match. For these three modes, the bonding mode really selects a
1899 transmit policy for an EtherChannel group; all three will interoperate
1900 with another EtherChannel group.
1903 6. 802.1q VLAN Support
1904 ======================
1906 It is possible to configure VLAN devices over a bond interface
1907 using the 8021q driver. However, only packets coming from the 8021q
1908 driver and passing through bonding will be tagged by default. Self
1909 generated packets, for example, bonding's learning packets or ARP
1910 packets generated by either ALB mode or the ARP monitor mechanism, are
1911 tagged internally by bonding itself. As a result, bonding must
1912 "learn" the VLAN IDs configured above it, and use those IDs to tag
1913 self generated packets.
1915 For reasons of simplicity, and to support the use of adapters
1916 that can do VLAN hardware acceleration offloading, the bonding
1917 interface declares itself as fully hardware offloading capable, it gets
1918 the add_vid/kill_vid notifications to gather the necessary
1919 information, and it propagates those actions to the slaves. In case
1920 of mixed adapter types, hardware accelerated tagged packets that
1921 should go through an adapter that is not offloading capable are
1922 "un-accelerated" by the bonding driver so the VLAN tag sits in the
1925 VLAN interfaces *must* be added on top of a bonding interface
1926 only after enslaving at least one slave. The bonding interface has a
1927 hardware address of 00:00:00:00:00:00 until the first slave is added.
1928 If the VLAN interface is created prior to the first enslavement, it
1929 would pick up the all-zeroes hardware address. Once the first slave
1930 is attached to the bond, the bond device itself will pick up the
1931 slave's hardware address, which is then available for the VLAN device.
1933 Also, be aware that a similar problem can occur if all slaves
1934 are released from a bond that still has one or more VLAN interfaces on
1935 top of it. When a new slave is added, the bonding interface will
1936 obtain its hardware address from the first slave, which might not
1937 match the hardware address of the VLAN interfaces (which was
1938 ultimately copied from an earlier slave).
1940 There are two methods to insure that the VLAN device operates
1941 with the correct hardware address if all slaves are removed from a
1944 1. Remove all VLAN interfaces then recreate them
1946 2. Set the bonding interface's hardware address so that it
1947 matches the hardware address of the VLAN interfaces.
1949 Note that changing a VLAN interface's HW address would set the
1950 underlying device -- i.e. the bonding interface -- to promiscuous
1951 mode, which might not be what you want.
1957 The bonding driver at present supports two schemes for
1958 monitoring a slave device's link state: the ARP monitor and the MII
1961 At the present time, due to implementation restrictions in the
1962 bonding driver itself, it is not possible to enable both ARP and MII
1963 monitoring simultaneously.
1965 7.1 ARP Monitor Operation
1966 -------------------------
1968 The ARP monitor operates as its name suggests: it sends ARP
1969 queries to one or more designated peer systems on the network, and
1970 uses the response as an indication that the link is operating. This
1971 gives some assurance that traffic is actually flowing to and from one
1972 or more peers on the local network.
1974 The ARP monitor relies on the device driver itself to verify
1975 that traffic is flowing. In particular, the driver must keep up to
1976 date the last receive time, dev->last_rx. Drivers that use NETIF_F_LLTX
1977 flag must also update netdev_queue->trans_start. If they do not, then the
1978 ARP monitor will immediately fail any slaves using that driver, and
1979 those slaves will stay down. If networking monitoring (tcpdump, etc)
1980 shows the ARP requests and replies on the network, then it may be that
1981 your device driver is not updating last_rx and trans_start.
1983 7.2 Configuring Multiple ARP Targets
1984 ------------------------------------
1986 While ARP monitoring can be done with just one target, it can
1987 be useful in a High Availability setup to have several targets to
1988 monitor. In the case of just one target, the target itself may go
1989 down or have a problem making it unresponsive to ARP requests. Having
1990 an additional target (or several) increases the reliability of the ARP
1993 Multiple ARP targets must be separated by commas as follows::
1995 # example options for ARP monitoring with three targets
1997 options bond0 arp_interval=60 arp_ip_target=192.168.0.1,192.168.0.3,192.168.0.9
1999 For just a single target the options would resemble::
2001 # example options for ARP monitoring with one target
2003 options bond0 arp_interval=60 arp_ip_target=192.168.0.100
2006 7.3 MII Monitor Operation
2007 -------------------------
2009 The MII monitor monitors only the carrier state of the local
2010 network interface. It accomplishes this in one of three ways: by
2011 depending upon the device driver to maintain its carrier state, by
2012 querying the device's MII registers, or by making an ethtool query to
2015 If the use_carrier module parameter is 1 (the default value),
2016 then the MII monitor will rely on the driver for carrier state
2017 information (via the netif_carrier subsystem). As explained in the
2018 use_carrier parameter information, above, if the MII monitor fails to
2019 detect carrier loss on the device (e.g., when the cable is physically
2020 disconnected), it may be that the driver does not support
2023 If use_carrier is 0, then the MII monitor will first query the
2024 device's (via ioctl) MII registers and check the link state. If that
2025 request fails (not just that it returns carrier down), then the MII
2026 monitor will make an ethtool ETHTOOL_GLINK request to attempt to obtain
2027 the same information. If both methods fail (i.e., the driver either
2028 does not support or had some error in processing both the MII register
2029 and ethtool requests), then the MII monitor will assume the link is
2032 8. Potential Sources of Trouble
2033 ===============================
2035 8.1 Adventures in Routing
2036 -------------------------
2038 When bonding is configured, it is important that the slave
2039 devices not have routes that supersede routes of the master (or,
2040 generally, not have routes at all). For example, suppose the bonding
2041 device bond0 has two slaves, eth0 and eth1, and the routing table is
2044 Kernel IP routing table
2045 Destination Gateway Genmask Flags MSS Window irtt Iface
2046 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth0
2047 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 eth1
2048 10.0.0.0 0.0.0.0 255.255.0.0 U 40 0 0 bond0
2049 127.0.0.0 0.0.0.0 255.0.0.0 U 40 0 0 lo
2051 This routing configuration will likely still update the
2052 receive/transmit times in the driver (needed by the ARP monitor), but
2053 may bypass the bonding driver (because outgoing traffic to, in this
2054 case, another host on network 10 would use eth0 or eth1 before bond0).
2056 The ARP monitor (and ARP itself) may become confused by this
2057 configuration, because ARP requests (generated by the ARP monitor)
2058 will be sent on one interface (bond0), but the corresponding reply
2059 will arrive on a different interface (eth0). This reply looks to ARP
2060 as an unsolicited ARP reply (because ARP matches replies on an
2061 interface basis), and is discarded. The MII monitor is not affected
2062 by the state of the routing table.
2064 The solution here is simply to insure that slaves do not have
2065 routes of their own, and if for some reason they must, those routes do
2066 not supersede routes of their master. This should generally be the
2067 case, but unusual configurations or errant manual or automatic static
2068 route additions may cause trouble.
2070 8.2 Ethernet Device Renaming
2071 ----------------------------
2073 On systems with network configuration scripts that do not
2074 associate physical devices directly with network interface names (so
2075 that the same physical device always has the same "ethX" name), it may
2076 be necessary to add some special logic to config files in
2079 For example, given a modules.conf containing the following::
2082 options bond0 mode=some-mode miimon=50
2088 If neither eth0 and eth1 are slaves to bond0, then when the
2089 bond0 interface comes up, the devices may end up reordered. This
2090 happens because bonding is loaded first, then its slave device's
2091 drivers are loaded next. Since no other drivers have been loaded,
2092 when the e1000 driver loads, it will receive eth0 and eth1 for its
2093 devices, but the bonding configuration tries to enslave eth2 and eth3
2094 (which may later be assigned to the tg3 devices).
2096 Adding the following::
2098 add above bonding e1000 tg3
2100 causes modprobe to load e1000 then tg3, in that order, when
2101 bonding is loaded. This command is fully documented in the
2102 modules.conf manual page.
2104 On systems utilizing modprobe an equivalent problem can occur.
2105 In this case, the following can be added to config files in
2106 /etc/modprobe.d/ as::
2108 softdep bonding pre: tg3 e1000
2110 This will load tg3 and e1000 modules before loading the bonding one.
2111 Full documentation on this can be found in the modprobe.d and modprobe
2114 8.3. Painfully Slow Or No Failed Link Detection By Miimon
2115 ---------------------------------------------------------
2117 By default, bonding enables the use_carrier option, which
2118 instructs bonding to trust the driver to maintain carrier state.
2120 As discussed in the options section, above, some drivers do
2121 not support the netif_carrier_on/_off link state tracking system.
2122 With use_carrier enabled, bonding will always see these links as up,
2123 regardless of their actual state.
2125 Additionally, other drivers do support netif_carrier, but do
2126 not maintain it in real time, e.g., only polling the link state at
2127 some fixed interval. In this case, miimon will detect failures, but
2128 only after some long period of time has expired. If it appears that
2129 miimon is very slow in detecting link failures, try specifying
2130 use_carrier=0 to see if that improves the failure detection time. If
2131 it does, then it may be that the driver checks the carrier state at a
2132 fixed interval, but does not cache the MII register values (so the
2133 use_carrier=0 method of querying the registers directly works). If
2134 use_carrier=0 does not improve the failover, then the driver may cache
2135 the registers, or the problem may be elsewhere.
2137 Also, remember that miimon only checks for the device's
2138 carrier state. It has no way to determine the state of devices on or
2139 beyond other ports of a switch, or if a switch is refusing to pass
2140 traffic while still maintaining carrier on.
2145 If running SNMP agents, the bonding driver should be loaded
2146 before any network drivers participating in a bond. This requirement
2147 is due to the interface index (ipAdEntIfIndex) being associated to
2148 the first interface found with a given IP address. That is, there is
2149 only one ipAdEntIfIndex for each IP address. For example, if eth0 and
2150 eth1 are slaves of bond0 and the driver for eth0 is loaded before the
2151 bonding driver, the interface for the IP address will be associated
2152 with the eth0 interface. This configuration is shown below, the IP
2153 address 192.168.1.1 has an interface index of 2 which indexes to eth0
2154 in the ifDescr table (ifDescr.2).
2158 interfaces.ifTable.ifEntry.ifDescr.1 = lo
2159 interfaces.ifTable.ifEntry.ifDescr.2 = eth0
2160 interfaces.ifTable.ifEntry.ifDescr.3 = eth1
2161 interfaces.ifTable.ifEntry.ifDescr.4 = eth2
2162 interfaces.ifTable.ifEntry.ifDescr.5 = eth3
2163 interfaces.ifTable.ifEntry.ifDescr.6 = bond0
2164 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 5
2165 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
2166 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 4
2167 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
2169 This problem is avoided by loading the bonding driver before
2170 any network drivers participating in a bond. Below is an example of
2171 loading the bonding driver first, the IP address 192.168.1.1 is
2172 correctly associated with ifDescr.2.
2174 interfaces.ifTable.ifEntry.ifDescr.1 = lo
2175 interfaces.ifTable.ifEntry.ifDescr.2 = bond0
2176 interfaces.ifTable.ifEntry.ifDescr.3 = eth0
2177 interfaces.ifTable.ifEntry.ifDescr.4 = eth1
2178 interfaces.ifTable.ifEntry.ifDescr.5 = eth2
2179 interfaces.ifTable.ifEntry.ifDescr.6 = eth3
2180 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.10.10.10 = 6
2181 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.192.168.1.1 = 2
2182 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.10.74.20.94 = 5
2183 ip.ipAddrTable.ipAddrEntry.ipAdEntIfIndex.127.0.0.1 = 1
2185 While some distributions may not report the interface name in
2186 ifDescr, the association between the IP address and IfIndex remains
2187 and SNMP functions such as Interface_Scan_Next will report that
2190 10. Promiscuous mode
2191 ====================
2193 When running network monitoring tools, e.g., tcpdump, it is
2194 common to enable promiscuous mode on the device, so that all traffic
2195 is seen (instead of seeing only traffic destined for the local host).
2196 The bonding driver handles promiscuous mode changes to the bonding
2197 master device (e.g., bond0), and propagates the setting to the slave
2200 For the balance-rr, balance-xor, broadcast, and 802.3ad modes,
2201 the promiscuous mode setting is propagated to all slaves.
2203 For the active-backup, balance-tlb and balance-alb modes, the
2204 promiscuous mode setting is propagated only to the active slave.
2206 For balance-tlb mode, the active slave is the slave currently
2207 receiving inbound traffic.
2209 For balance-alb mode, the active slave is the slave used as a
2210 "primary." This slave is used for mode-specific control traffic, for
2211 sending to peers that are unassigned or if the load is unbalanced.
2213 For the active-backup, balance-tlb and balance-alb modes, when
2214 the active slave changes (e.g., due to a link failure), the
2215 promiscuous setting will be propagated to the new active slave.
2217 11. Configuring Bonding for High Availability
2218 =============================================
2220 High Availability refers to configurations that provide
2221 maximum network availability by having redundant or backup devices,
2222 links or switches between the host and the rest of the world. The
2223 goal is to provide the maximum availability of network connectivity
2224 (i.e., the network always works), even though other configurations
2225 could provide higher throughput.
2227 11.1 High Availability in a Single Switch Topology
2228 --------------------------------------------------
2230 If two hosts (or a host and a single switch) are directly
2231 connected via multiple physical links, then there is no availability
2232 penalty to optimizing for maximum bandwidth. In this case, there is
2233 only one switch (or peer), so if it fails, there is no alternative
2234 access to fail over to. Additionally, the bonding load balance modes
2235 support link monitoring of their members, so if individual links fail,
2236 the load will be rebalanced across the remaining devices.
2238 See Section 12, "Configuring Bonding for Maximum Throughput"
2239 for information on configuring bonding with one peer device.
2241 11.2 High Availability in a Multiple Switch Topology
2242 ----------------------------------------------------
2244 With multiple switches, the configuration of bonding and the
2245 network changes dramatically. In multiple switch topologies, there is
2246 a trade off between network availability and usable bandwidth.
2248 Below is a sample network, configured to maximize the
2249 availability of the network::
2253 +-----+----+ +-----+----+
2254 | |port2 ISL port2| |
2255 | switch A +--------------------------+ switch B |
2257 +-----+----+ +-----++---+
2260 +-------------+ host1 +---------------+
2263 In this configuration, there is a link between the two
2264 switches (ISL, or inter switch link), and multiple ports connecting to
2265 the outside world ("port3" on each switch). There is no technical
2266 reason that this could not be extended to a third switch.
2268 11.2.1 HA Bonding Mode Selection for Multiple Switch Topology
2269 -------------------------------------------------------------
2271 In a topology such as the example above, the active-backup and
2272 broadcast modes are the only useful bonding modes when optimizing for
2273 availability; the other modes require all links to terminate on the
2274 same peer for them to behave rationally.
2277 This is generally the preferred mode, particularly if
2278 the switches have an ISL and play together well. If the
2279 network configuration is such that one switch is specifically
2280 a backup switch (e.g., has lower capacity, higher cost, etc),
2281 then the primary option can be used to insure that the
2282 preferred link is always used when it is available.
2285 This mode is really a special purpose mode, and is suitable
2286 only for very specific needs. For example, if the two
2287 switches are not connected (no ISL), and the networks beyond
2288 them are totally independent. In this case, if it is
2289 necessary for some specific one-way traffic to reach both
2290 independent networks, then the broadcast mode may be suitable.
2292 11.2.2 HA Link Monitoring Selection for Multiple Switch Topology
2293 ----------------------------------------------------------------
2295 The choice of link monitoring ultimately depends upon your
2296 switch. If the switch can reliably fail ports in response to other
2297 failures, then either the MII or ARP monitors should work. For
2298 example, in the above example, if the "port3" link fails at the remote
2299 end, the MII monitor has no direct means to detect this. The ARP
2300 monitor could be configured with a target at the remote end of port3,
2301 thus detecting that failure without switch support.
2303 In general, however, in a multiple switch topology, the ARP
2304 monitor can provide a higher level of reliability in detecting end to
2305 end connectivity failures (which may be caused by the failure of any
2306 individual component to pass traffic for any reason). Additionally,
2307 the ARP monitor should be configured with multiple targets (at least
2308 one for each switch in the network). This will insure that,
2309 regardless of which switch is active, the ARP monitor has a suitable
2312 Note, also, that of late many switches now support a functionality
2313 generally referred to as "trunk failover." This is a feature of the
2314 switch that causes the link state of a particular switch port to be set
2315 down (or up) when the state of another switch port goes down (or up).
2316 Its purpose is to propagate link failures from logically "exterior" ports
2317 to the logically "interior" ports that bonding is able to monitor via
2318 miimon. Availability and configuration for trunk failover varies by
2319 switch, but this can be a viable alternative to the ARP monitor when using
2322 12. Configuring Bonding for Maximum Throughput
2323 ==============================================
2325 12.1 Maximizing Throughput in a Single Switch Topology
2326 ------------------------------------------------------
2328 In a single switch configuration, the best method to maximize
2329 throughput depends upon the application and network environment. The
2330 various load balancing modes each have strengths and weaknesses in
2331 different environments, as detailed below.
2333 For this discussion, we will break down the topologies into
2334 two categories. Depending upon the destination of most traffic, we
2335 categorize them into either "gatewayed" or "local" configurations.
2337 In a gatewayed configuration, the "switch" is acting primarily
2338 as a router, and the majority of traffic passes through this router to
2339 other networks. An example would be the following::
2342 +----------+ +----------+
2343 | |eth0 port1| | to other networks
2344 | Host A +---------------------+ router +------------------->
2345 | +---------------------+ | Hosts B and C are out
2346 | |eth1 port2| | here somewhere
2347 +----------+ +----------+
2349 The router may be a dedicated router device, or another host
2350 acting as a gateway. For our discussion, the important point is that
2351 the majority of traffic from Host A will pass through the router to
2352 some other network before reaching its final destination.
2354 In a gatewayed network configuration, although Host A may
2355 communicate with many other systems, all of its traffic will be sent
2356 and received via one other peer on the local network, the router.
2358 Note that the case of two systems connected directly via
2359 multiple physical links is, for purposes of configuring bonding, the
2360 same as a gatewayed configuration. In that case, it happens that all
2361 traffic is destined for the "gateway" itself, not some other network
2364 In a local configuration, the "switch" is acting primarily as
2365 a switch, and the majority of traffic passes through this switch to
2366 reach other stations on the same network. An example would be the
2369 +----------+ +----------+ +--------+
2370 | |eth0 port1| +-------+ Host B |
2371 | Host A +------------+ switch |port3 +--------+
2372 | +------------+ | +--------+
2373 | |eth1 port2| +------------------+ Host C |
2374 +----------+ +----------+port4 +--------+
2377 Again, the switch may be a dedicated switch device, or another
2378 host acting as a gateway. For our discussion, the important point is
2379 that the majority of traffic from Host A is destined for other hosts
2380 on the same local network (Hosts B and C in the above example).
2382 In summary, in a gatewayed configuration, traffic to and from
2383 the bonded device will be to the same MAC level peer on the network
2384 (the gateway itself, i.e., the router), regardless of its final
2385 destination. In a local configuration, traffic flows directly to and
2386 from the final destinations, thus, each destination (Host B, Host C)
2387 will be addressed directly by their individual MAC addresses.
2389 This distinction between a gatewayed and a local network
2390 configuration is important because many of the load balancing modes
2391 available use the MAC addresses of the local network source and
2392 destination to make load balancing decisions. The behavior of each
2393 mode is described below.
2396 12.1.1 MT Bonding Mode Selection for Single Switch Topology
2397 -----------------------------------------------------------
2399 This configuration is the easiest to set up and to understand,
2400 although you will have to decide which bonding mode best suits your
2401 needs. The trade offs for each mode are detailed below:
2404 This mode is the only mode that will permit a single
2405 TCP/IP connection to stripe traffic across multiple
2406 interfaces. It is therefore the only mode that will allow a
2407 single TCP/IP stream to utilize more than one interface's
2408 worth of throughput. This comes at a cost, however: the
2409 striping generally results in peer systems receiving packets out
2410 of order, causing TCP/IP's congestion control system to kick
2411 in, often by retransmitting segments.
2413 It is possible to adjust TCP/IP's congestion limits by
2414 altering the net.ipv4.tcp_reordering sysctl parameter. The
2415 usual default value is 3. But keep in mind TCP stack is able
2416 to automatically increase this when it detects reorders.
2418 Note that the fraction of packets that will be delivered out of
2419 order is highly variable, and is unlikely to be zero. The level
2420 of reordering depends upon a variety of factors, including the
2421 networking interfaces, the switch, and the topology of the
2422 configuration. Speaking in general terms, higher speed network
2423 cards produce more reordering (due to factors such as packet
2424 coalescing), and a "many to many" topology will reorder at a
2425 higher rate than a "many slow to one fast" configuration.
2427 Many switches do not support any modes that stripe traffic
2428 (instead choosing a port based upon IP or MAC level addresses);
2429 for those devices, traffic for a particular connection flowing
2430 through the switch to a balance-rr bond will not utilize greater
2431 than one interface's worth of bandwidth.
2433 If you are utilizing protocols other than TCP/IP, UDP for
2434 example, and your application can tolerate out of order
2435 delivery, then this mode can allow for single stream datagram
2436 performance that scales near linearly as interfaces are added
2439 This mode requires the switch to have the appropriate ports
2440 configured for "etherchannel" or "trunking."
2443 There is not much advantage in this network topology to
2444 the active-backup mode, as the inactive backup devices are all
2445 connected to the same peer as the primary. In this case, a
2446 load balancing mode (with link monitoring) will provide the
2447 same level of network availability, but with increased
2448 available bandwidth. On the plus side, active-backup mode
2449 does not require any configuration of the switch, so it may
2450 have value if the hardware available does not support any of
2451 the load balance modes.
2454 This mode will limit traffic such that packets destined
2455 for specific peers will always be sent over the same
2456 interface. Since the destination is determined by the MAC
2457 addresses involved, this mode works best in a "local" network
2458 configuration (as described above), with destinations all on
2459 the same local network. This mode is likely to be suboptimal
2460 if all your traffic is passed through a single router (i.e., a
2461 "gatewayed" network configuration, as described above).
2463 As with balance-rr, the switch ports need to be configured for
2464 "etherchannel" or "trunking."
2467 Like active-backup, there is not much advantage to this
2468 mode in this type of network topology.
2471 This mode can be a good choice for this type of network
2472 topology. The 802.3ad mode is an IEEE standard, so all peers
2473 that implement 802.3ad should interoperate well. The 802.3ad
2474 protocol includes automatic configuration of the aggregates,
2475 so minimal manual configuration of the switch is needed
2476 (typically only to designate that some set of devices is
2477 available for 802.3ad). The 802.3ad standard also mandates
2478 that frames be delivered in order (within certain limits), so
2479 in general single connections will not see misordering of
2480 packets. The 802.3ad mode does have some drawbacks: the
2481 standard mandates that all devices in the aggregate operate at
2482 the same speed and duplex. Also, as with all bonding load
2483 balance modes other than balance-rr, no single connection will
2484 be able to utilize more than a single interface's worth of
2487 Additionally, the linux bonding 802.3ad implementation
2488 distributes traffic by peer (using an XOR of MAC addresses
2489 and packet type ID), so in a "gatewayed" configuration, all
2490 outgoing traffic will generally use the same device. Incoming
2491 traffic may also end up on a single device, but that is
2492 dependent upon the balancing policy of the peer's 802.3ad
2493 implementation. In a "local" configuration, traffic will be
2494 distributed across the devices in the bond.
2496 Finally, the 802.3ad mode mandates the use of the MII monitor,
2497 therefore, the ARP monitor is not available in this mode.
2500 The balance-tlb mode balances outgoing traffic by peer.
2501 Since the balancing is done according to MAC address, in a
2502 "gatewayed" configuration (as described above), this mode will
2503 send all traffic across a single device. However, in a
2504 "local" network configuration, this mode balances multiple
2505 local network peers across devices in a vaguely intelligent
2506 manner (not a simple XOR as in balance-xor or 802.3ad mode),
2507 so that mathematically unlucky MAC addresses (i.e., ones that
2508 XOR to the same value) will not all "bunch up" on a single
2511 Unlike 802.3ad, interfaces may be of differing speeds, and no
2512 special switch configuration is required. On the down side,
2513 in this mode all incoming traffic arrives over a single
2514 interface, this mode requires certain ethtool support in the
2515 network device driver of the slave interfaces, and the ARP
2516 monitor is not available.
2519 This mode is everything that balance-tlb is, and more.
2520 It has all of the features (and restrictions) of balance-tlb,
2521 and will also balance incoming traffic from local network
2522 peers (as described in the Bonding Module Options section,
2525 The only additional down side to this mode is that the network
2526 device driver must support changing the hardware address while
2529 12.1.2 MT Link Monitoring for Single Switch Topology
2530 ----------------------------------------------------
2532 The choice of link monitoring may largely depend upon which
2533 mode you choose to use. The more advanced load balancing modes do not
2534 support the use of the ARP monitor, and are thus restricted to using
2535 the MII monitor (which does not provide as high a level of end to end
2536 assurance as the ARP monitor).
2538 12.2 Maximum Throughput in a Multiple Switch Topology
2539 -----------------------------------------------------
2541 Multiple switches may be utilized to optimize for throughput
2542 when they are configured in parallel as part of an isolated network
2543 between two or more systems, for example::
2549 +--------+ | +---------+
2551 +------+---+ +-----+----+ +-----+----+
2552 | Switch A | | Switch B | | Switch C |
2553 +------+---+ +-----+----+ +-----+----+
2555 +--------+ | +---------+
2561 In this configuration, the switches are isolated from one
2562 another. One reason to employ a topology such as this is for an
2563 isolated network with many hosts (a cluster configured for high
2564 performance, for example), using multiple smaller switches can be more
2565 cost effective than a single larger switch, e.g., on a network with 24
2566 hosts, three 24 port switches can be significantly less expensive than
2567 a single 72 port switch.
2569 If access beyond the network is required, an individual host
2570 can be equipped with an additional network device connected to an
2571 external network; this host then additionally acts as a gateway.
2573 12.2.1 MT Bonding Mode Selection for Multiple Switch Topology
2574 -------------------------------------------------------------
2576 In actual practice, the bonding mode typically employed in
2577 configurations of this type is balance-rr. Historically, in this
2578 network configuration, the usual caveats about out of order packet
2579 delivery are mitigated by the use of network adapters that do not do
2580 any kind of packet coalescing (via the use of NAPI, or because the
2581 device itself does not generate interrupts until some number of
2582 packets has arrived). When employed in this fashion, the balance-rr
2583 mode allows individual connections between two hosts to effectively
2584 utilize greater than one interface's bandwidth.
2586 12.2.2 MT Link Monitoring for Multiple Switch Topology
2587 ------------------------------------------------------
2589 Again, in actual practice, the MII monitor is most often used
2590 in this configuration, as performance is given preference over
2591 availability. The ARP monitor will function in this topology, but its
2592 advantages over the MII monitor are mitigated by the volume of probes
2593 needed as the number of systems involved grows (remember that each
2594 host in the network is configured with bonding).
2596 13. Switch Behavior Issues
2597 ==========================
2599 13.1 Link Establishment and Failover Delays
2600 -------------------------------------------
2602 Some switches exhibit undesirable behavior with regard to the
2603 timing of link up and down reporting by the switch.
2605 First, when a link comes up, some switches may indicate that
2606 the link is up (carrier available), but not pass traffic over the
2607 interface for some period of time. This delay is typically due to
2608 some type of autonegotiation or routing protocol, but may also occur
2609 during switch initialization (e.g., during recovery after a switch
2610 failure). If you find this to be a problem, specify an appropriate
2611 value to the updelay bonding module option to delay the use of the
2612 relevant interface(s).
2614 Second, some switches may "bounce" the link state one or more
2615 times while a link is changing state. This occurs most commonly while
2616 the switch is initializing. Again, an appropriate updelay value may
2619 Note that when a bonding interface has no active links, the
2620 driver will immediately reuse the first link that goes up, even if the
2621 updelay parameter has been specified (the updelay is ignored in this
2622 case). If there are slave interfaces waiting for the updelay timeout
2623 to expire, the interface that first went into that state will be
2624 immediately reused. This reduces down time of the network if the
2625 value of updelay has been overestimated, and since this occurs only in
2626 cases with no connectivity, there is no additional penalty for
2627 ignoring the updelay.
2629 In addition to the concerns about switch timings, if your
2630 switches take a long time to go into backup mode, it may be desirable
2631 to not activate a backup interface immediately after a link goes down.
2632 Failover may be delayed via the downdelay bonding module option.
2634 13.2 Duplicated Incoming Packets
2635 --------------------------------
2637 NOTE: Starting with version 3.0.2, the bonding driver has logic to
2638 suppress duplicate packets, which should largely eliminate this problem.
2639 The following description is kept for reference.
2641 It is not uncommon to observe a short burst of duplicated
2642 traffic when the bonding device is first used, or after it has been
2643 idle for some period of time. This is most easily observed by issuing
2644 a "ping" to some other host on the network, and noticing that the
2645 output from ping flags duplicates (typically one per slave).
2647 For example, on a bond in active-backup mode with five slaves
2648 all connected to one switch, the output may appear as follows::
2651 PING 10.0.4.2 (10.0.4.2) from 10.0.3.10 : 56(84) bytes of data.
2652 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.7 ms
2653 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2654 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2655 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2656 64 bytes from 10.0.4.2: icmp_seq=1 ttl=64 time=13.8 ms (DUP!)
2657 64 bytes from 10.0.4.2: icmp_seq=2 ttl=64 time=0.216 ms
2658 64 bytes from 10.0.4.2: icmp_seq=3 ttl=64 time=0.267 ms
2659 64 bytes from 10.0.4.2: icmp_seq=4 ttl=64 time=0.222 ms
2661 This is not due to an error in the bonding driver, rather, it
2662 is a side effect of how many switches update their MAC forwarding
2663 tables. Initially, the switch does not associate the MAC address in
2664 the packet with a particular switch port, and so it may send the
2665 traffic to all ports until its MAC forwarding table is updated. Since
2666 the interfaces attached to the bond may occupy multiple ports on a
2667 single switch, when the switch (temporarily) floods the traffic to all
2668 ports, the bond device receives multiple copies of the same packet
2669 (one per slave device).
2671 The duplicated packet behavior is switch dependent, some
2672 switches exhibit this, and some do not. On switches that display this
2673 behavior, it can be induced by clearing the MAC forwarding table (on
2674 most Cisco switches, the privileged command "clear mac address-table
2675 dynamic" will accomplish this).
2677 14. Hardware Specific Considerations
2678 ====================================
2680 This section contains additional information for configuring
2681 bonding on specific hardware platforms, or for interfacing bonding
2682 with particular switches or other devices.
2684 14.1 IBM BladeCenter
2685 --------------------
2687 This applies to the JS20 and similar systems.
2689 On the JS20 blades, the bonding driver supports only
2690 balance-rr, active-backup, balance-tlb and balance-alb modes. This is
2691 largely due to the network topology inside the BladeCenter, detailed
2694 JS20 network adapter information
2695 --------------------------------
2697 All JS20s come with two Broadcom Gigabit Ethernet ports
2698 integrated on the planar (that's "motherboard" in IBM-speak). In the
2699 BladeCenter chassis, the eth0 port of all JS20 blades is hard wired to
2700 I/O Module #1; similarly, all eth1 ports are wired to I/O Module #2.
2701 An add-on Broadcom daughter card can be installed on a JS20 to provide
2702 two more Gigabit Ethernet ports. These ports, eth2 and eth3, are
2703 wired to I/O Modules 3 and 4, respectively.
2705 Each I/O Module may contain either a switch or a passthrough
2706 module (which allows ports to be directly connected to an external
2707 switch). Some bonding modes require a specific BladeCenter internal
2708 network topology in order to function; these are detailed below.
2710 Additional BladeCenter-specific networking information can be
2711 found in two IBM Redbooks (www.ibm.com/redbooks):
2713 - "IBM eServer BladeCenter Networking Options"
2714 - "IBM eServer BladeCenter Layer 2-7 Network Switching"
2716 BladeCenter networking configuration
2717 ------------------------------------
2719 Because a BladeCenter can be configured in a very large number
2720 of ways, this discussion will be confined to describing basic
2723 Normally, Ethernet Switch Modules (ESMs) are used in I/O
2724 modules 1 and 2. In this configuration, the eth0 and eth1 ports of a
2725 JS20 will be connected to different internal switches (in the
2726 respective I/O modules).
2728 A passthrough module (OPM or CPM, optical or copper,
2729 passthrough module) connects the I/O module directly to an external
2730 switch. By using PMs in I/O module #1 and #2, the eth0 and eth1
2731 interfaces of a JS20 can be redirected to the outside world and
2732 connected to a common external switch.
2734 Depending upon the mix of ESMs and PMs, the network will
2735 appear to bonding as either a single switch topology (all PMs) or as a
2736 multiple switch topology (one or more ESMs, zero or more PMs). It is
2737 also possible to connect ESMs together, resulting in a configuration
2738 much like the example in "High Availability in a Multiple Switch
2741 Requirements for specific modes
2742 -------------------------------
2744 The balance-rr mode requires the use of passthrough modules
2745 for devices in the bond, all connected to an common external switch.
2746 That switch must be configured for "etherchannel" or "trunking" on the
2747 appropriate ports, as is usual for balance-rr.
2749 The balance-alb and balance-tlb modes will function with
2750 either switch modules or passthrough modules (or a mix). The only
2751 specific requirement for these modes is that all network interfaces
2752 must be able to reach all destinations for traffic sent over the
2753 bonding device (i.e., the network must converge at some point outside
2756 The active-backup mode has no additional requirements.
2758 Link monitoring issues
2759 ----------------------
2761 When an Ethernet Switch Module is in place, only the ARP
2762 monitor will reliably detect link loss to an external switch. This is
2763 nothing unusual, but examination of the BladeCenter cabinet would
2764 suggest that the "external" network ports are the ethernet ports for
2765 the system, when it fact there is a switch between these "external"
2766 ports and the devices on the JS20 system itself. The MII monitor is
2767 only able to detect link failures between the ESM and the JS20 system.
2769 When a passthrough module is in place, the MII monitor does
2770 detect failures to the "external" port, which is then directly
2771 connected to the JS20 system.
2776 The Serial Over LAN (SoL) link is established over the primary
2777 ethernet (eth0) only, therefore, any loss of link to eth0 will result
2778 in losing your SoL connection. It will not fail over with other
2779 network traffic, as the SoL system is beyond the control of the
2782 It may be desirable to disable spanning tree on the switch
2783 (either the internal Ethernet Switch Module, or an external switch) to
2784 avoid fail-over delay issues when using bonding.
2787 15. Frequently Asked Questions
2788 ==============================
2793 Yes. The old 2.0.xx channel bonding patch was not SMP safe.
2794 The new driver was designed to be SMP safe from the start.
2796 2. What type of cards will work with it?
2797 -----------------------------------------
2799 Any Ethernet type cards (you can even mix cards - a Intel
2800 EtherExpress PRO/100 and a 3com 3c905b, for example). For most modes,
2801 devices need not be of the same speed.
2803 Starting with version 3.2.1, bonding also supports Infiniband
2804 slaves in active-backup mode.
2806 3. How many bonding devices can I have?
2807 ----------------------------------------
2811 4. How many slaves can a bonding device have?
2812 ----------------------------------------------
2814 This is limited only by the number of network interfaces Linux
2815 supports and/or the number of network cards you can place in your
2818 5. What happens when a slave link dies?
2819 ----------------------------------------
2821 If link monitoring is enabled, then the failing device will be
2822 disabled. The active-backup mode will fail over to a backup link, and
2823 other modes will ignore the failed link. The link will continue to be
2824 monitored, and should it recover, it will rejoin the bond (in whatever
2825 manner is appropriate for the mode). See the sections on High
2826 Availability and the documentation for each mode for additional
2829 Link monitoring can be enabled via either the miimon or
2830 arp_interval parameters (described in the module parameters section,
2831 above). In general, miimon monitors the carrier state as sensed by
2832 the underlying network device, and the arp monitor (arp_interval)
2833 monitors connectivity to another host on the local network.
2835 If no link monitoring is configured, the bonding driver will
2836 be unable to detect link failures, and will assume that all links are
2837 always available. This will likely result in lost packets, and a
2838 resulting degradation of performance. The precise performance loss
2839 depends upon the bonding mode and network configuration.
2841 6. Can bonding be used for High Availability?
2842 ----------------------------------------------
2844 Yes. See the section on High Availability for details.
2846 7. Which switches/systems does it work with?
2847 ---------------------------------------------
2849 The full answer to this depends upon the desired mode.
2851 In the basic balance modes (balance-rr and balance-xor), it
2852 works with any system that supports etherchannel (also called
2853 trunking). Most managed switches currently available have such
2854 support, and many unmanaged switches as well.
2856 The advanced balance modes (balance-tlb and balance-alb) do
2857 not have special switch requirements, but do need device drivers that
2858 support specific features (described in the appropriate section under
2859 module parameters, above).
2861 In 802.3ad mode, it works with systems that support IEEE
2862 802.3ad Dynamic Link Aggregation. Most managed and many unmanaged
2863 switches currently available support 802.3ad.
2865 The active-backup mode should work with any Layer-II switch.
2867 8. Where does a bonding device get its MAC address from?
2868 ---------------------------------------------------------
2870 When using slave devices that have fixed MAC addresses, or when
2871 the fail_over_mac option is enabled, the bonding device's MAC address is
2872 the MAC address of the active slave.
2874 For other configurations, if not explicitly configured (with
2875 ifconfig or ip link), the MAC address of the bonding device is taken from
2876 its first slave device. This MAC address is then passed to all following
2877 slaves and remains persistent (even if the first slave is removed) until
2878 the bonding device is brought down or reconfigured.
2880 If you wish to change the MAC address, you can set it with
2881 ifconfig or ip link::
2883 # ifconfig bond0 hw ether 00:11:22:33:44:55
2885 # ip link set bond0 address 66:77:88:99:aa:bb
2887 The MAC address can be also changed by bringing down/up the
2888 device and then changing its slaves (or their order)::
2890 # ifconfig bond0 down ; modprobe -r bonding
2891 # ifconfig bond0 .... up
2892 # ifenslave bond0 eth...
2894 This method will automatically take the address from the next
2895 slave that is added.
2897 To restore your slaves' MAC addresses, you need to detach them
2898 from the bond (``ifenslave -d bond0 eth0``). The bonding driver will
2899 then restore the MAC addresses that the slaves had before they were
2902 16. Resources and Links
2903 =======================
2905 The latest version of the bonding driver can be found in the latest
2906 version of the linux kernel, found on http://kernel.org
2908 The latest version of this document can be found in the latest kernel
2909 source (named Documentation/networking/bonding.rst).
2911 Discussions regarding the development of the bonding driver take place
2912 on the main Linux network mailing list, hosted at vger.kernel.org. The list
2915 netdev@vger.kernel.org
2917 The administrative interface (to subscribe or unsubscribe) can
2920 http://vger.kernel.org/vger-lists.html#netdev
projects / linux-2.6-microblaze.git / blob