aboutsummaryrefslogtreecommitdiffstats
path: root/Documentation/networking
diff options
context:
space:
mode:
authorJay Vosburgh <fubar@us.ibm.com>2007-11-13 23:25:48 -0500
committerDavid S. Miller <davem@davemloft.net>2008-01-28 18:03:48 -0500
commit9a6c686799346b6c95405c9e051f5023873504fa (patch)
treef1a8414921b8535a5fb4acdf8d57e69459c48289 /Documentation/networking
parent7a47dd7a2f178cc4e87d584b0469eef4b58b7aea (diff)
[BONDING]: Documentation update
Update the bonding documentation: more discussion on initialization and configuration, changes to discussion of packet reordering in balance-rr, update some out of date information. Based in part on input from Rick Jones <rick.jones2@hp.com> and Andy Gospodarek <andy@greyhouse.net>. Signed-off-by: Jay Vosburgh <fubar@us.ibm.com> Acked-by: Andy Gospodarek <andy@greyhouse.net> Signed-off-by: David S. Miller <davem@davemloft.net>
Diffstat (limited to 'Documentation/networking')
-rw-r--r--Documentation/networking/bonding.txt204
1 files changed, 143 insertions, 61 deletions
diff --git a/Documentation/networking/bonding.txt b/Documentation/networking/bonding.txt
index 6cc30e0d5795..a0cda062bc33 100644
--- a/Documentation/networking/bonding.txt
+++ b/Documentation/networking/bonding.txt
@@ -1,7 +1,7 @@
1 1
2 Linux Ethernet Bonding Driver HOWTO 2 Linux Ethernet Bonding Driver HOWTO
3 3
4 Latest update: 24 April 2006 4 Latest update: 12 November 2007
5 5
6Initial release : Thomas Davis <tadavis at lbl.gov> 6Initial release : Thomas Davis <tadavis at lbl.gov>
7Corrections, HA extensions : 2000/10/03-15 : 7Corrections, HA extensions : 2000/10/03-15 :
@@ -166,12 +166,17 @@ to use ifenslave.
1662. Bonding Driver Options 1662. Bonding Driver Options
167========================= 167=========================
168 168
169 Options for the bonding driver are supplied as parameters to 169 Options for the bonding driver are supplied as parameters to the
170the bonding module at load time. They may be given as command line 170bonding module at load time, or are specified via sysfs.
171arguments to the insmod or modprobe command, but are usually specified 171
172in either the /etc/modules.conf or /etc/modprobe.conf configuration 172 Module options may be given as command line arguments to the
173file, or in a distro-specific configuration file (some of which are 173insmod or modprobe command, but are usually specified in either the
174detailed in the next section). 174/etc/modules.conf or /etc/modprobe.conf configuration file, or in a
175distro-specific configuration file (some of which are detailed in the next
176section).
177
178 Details on bonding support for sysfs is provided in the
179"Configuring Bonding Manually via Sysfs" section, below.
175 180
176 The available bonding driver parameters are listed below. If a 181 The available bonding driver parameters are listed below. If a
177parameter is not specified the default value is used. When initially 182parameter is not specified the default value is used. When initially
@@ -812,11 +817,13 @@ the system /etc/modules.conf or /etc/modprobe.conf configuration file.
8123.2 Configuration with Initscripts Support 8173.2 Configuration with Initscripts Support
813------------------------------------------ 818------------------------------------------
814 819
815 This section applies to distros using a version of initscripts 820 This section applies to distros using a recent version of
816with bonding support, for example, Red Hat Linux 9 or Red Hat 821initscripts with bonding support, for example, Red Hat Enterprise Linux
817Enterprise Linux version 3 or 4. On these systems, the network 822version 3 or later, Fedora, etc. On these systems, the network
818initialization scripts have some knowledge of bonding, and can be 823initialization scripts have knowledge of bonding, and can be configured to
819configured to control bonding devices. 824control bonding devices. Note that older versions of the initscripts
825package have lower levels of support for bonding; this will be noted where
826applicable.
820 827
821 These distros will not automatically load the network adapter 828 These distros will not automatically load the network adapter
822driver unless the ethX device is configured with an IP address. 829driver unless the ethX device is configured with an IP address.
@@ -864,11 +871,31 @@ USERCTL=no
864 Be sure to change the networking specific lines (IPADDR, 871 Be sure to change the networking specific lines (IPADDR,
865NETMASK, NETWORK and BROADCAST) to match your network configuration. 872NETMASK, NETWORK and BROADCAST) to match your network configuration.
866 873
867 Finally, it is necessary to edit /etc/modules.conf (or 874 For later versions of initscripts, such as that found with Fedora
868/etc/modprobe.conf, depending upon your distro) to load the bonding 8757 and Red Hat Enterprise Linux version 5 (or later), it is possible, and,
869module with your desired options when the bond0 interface is brought 876indeed, preferable, to specify the bonding options in the ifcfg-bond0
870up. The following lines in /etc/modules.conf (or modprobe.conf) will 877file, e.g. a line of the format:
871load the bonding module, and select its options: 878
879BONDING_OPTS="mode=active-backup arp_interval=60 arp_ip_target=+192.168.1.254"
880
881 will configure the bond with the specified options. The options
882specified in BONDING_OPTS are identical to the bonding module parameters
883except for the arp_ip_target field. Each target should be included as a
884separate option and should be preceded by a '+' to indicate it should be
885added to the list of queried targets, e.g.,
886
887 arp_ip_target=+192.168.1.1 arp_ip_target=+192.168.1.2
888
889 is the proper syntax to specify multiple targets. When specifying
890options via BONDING_OPTS, it is not necessary to edit /etc/modules.conf or
891/etc/modprobe.conf.
892
893 For older versions of initscripts that do not support
894BONDING_OPTS, it is necessary to edit /etc/modules.conf (or
895/etc/modprobe.conf, depending upon your distro) to load the bonding module
896with your desired options when the bond0 interface is brought up. The
897following lines in /etc/modules.conf (or modprobe.conf) will load the
898bonding module, and select its options:
872 899
873alias bond0 bonding 900alias bond0 bonding
874options bond0 mode=balance-alb miimon=100 901options bond0 mode=balance-alb miimon=100
@@ -883,9 +910,10 @@ up and running.
8833.2.1 Using DHCP with Initscripts 9103.2.1 Using DHCP with Initscripts
884--------------------------------- 911---------------------------------
885 912
886 Recent versions of initscripts (the version supplied with 913 Recent versions of initscripts (the versions supplied with Fedora
887Fedora Core 3 and Red Hat Enterprise Linux 4 is reported to work) do 914Core 3 and Red Hat Enterprise Linux 4, or later versions, are reported to
888have support for assigning IP information to bonding devices via DHCP. 915work) have support for assigning IP information to bonding devices via
916DHCP.
889 917
890 To configure bonding for DHCP, configure it as described 918 To configure bonding for DHCP, configure it as described
891above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp" 919above, except replace the line "BOOTPROTO=none" with "BOOTPROTO=dhcp"
@@ -895,18 +923,14 @@ is case sensitive.
8953.2.2 Configuring Multiple Bonds with Initscripts 9233.2.2 Configuring Multiple Bonds with Initscripts
896------------------------------------------------- 924-------------------------------------------------
897 925
898 At this writing, the initscripts package does not directly 926 Initscripts packages that are included with Fedora 7 and Red Hat
899support loading the bonding driver multiple times, so the process for 927Enterprise Linux 5 support multiple bonding interfaces by simply
900doing so is the same as described in the "Configuring Multiple Bonds 928specifying the appropriate BONDING_OPTS= in ifcfg-bondX where X is the
901Manually" section, below. 929number of the bond. This support requires sysfs support in the kernel,
902 930and a bonding driver of version 3.0.0 or later. Other configurations may
903 NOTE: It has been observed that some Red Hat supplied kernels 931not support this method for specifying multiple bonding interfaces; for
904are apparently unable to rename modules at load time (the "-o bond1" 932those instances, see the "Configuring Multiple Bonds Manually" section,
905part). Attempts to pass that option to modprobe will produce an 933below.
906"Operation not permitted" error. This has been reported on some
907Fedora Core kernels, and has been seen on RHEL 4 as well. On kernels
908exhibiting this problem, it will be impossible to configure multiple
909bonds with differing parameters.
910 934
9113.3 Configuring Bonding Manually with Ifenslave 9353.3 Configuring Bonding Manually with Ifenslave
912----------------------------------------------- 936-----------------------------------------------
@@ -977,15 +1001,58 @@ initialization scripts lack support for configuring multiple bonds.
977options, you may wish to use the "max_bonds" module parameter, 1001options, you may wish to use the "max_bonds" module parameter,
978documented above. 1002documented above.
979 1003
980 To create multiple bonding devices with differing options, it 1004 To create multiple bonding devices with differing options, it is
981is necessary to use bonding parameters exported by sysfs, documented 1005preferrable to use bonding parameters exported by sysfs, documented in the
982in the section below. 1006section below.
1007
1008 For versions of bonding without sysfs support, the only means to
1009provide multiple instances of bonding with differing options is to load
1010the bonding driver multiple times. Note that current versions of the
1011sysconfig network initialization scripts handle this automatically; if
1012your distro uses these scripts, no special action is needed. See the
1013section Configuring Bonding Devices, above, if you're not sure about your
1014network initialization scripts.
1015
1016 To load multiple instances of the module, it is necessary to
1017specify a different name for each instance (the module loading system
1018requires that every loaded module, even multiple instances of the same
1019module, have a unique name). This is accomplished by supplying multiple
1020sets of bonding options in /etc/modprobe.conf, for example:
1021
1022alias bond0 bonding
1023options bond0 -o bond0 mode=balance-rr miimon=100
1024
1025alias bond1 bonding
1026options bond1 -o bond1 mode=balance-alb miimon=50
1027
1028 will load the bonding module two times. The first instance is
1029named "bond0" and creates the bond0 device in balance-rr mode with an
1030miimon of 100. The second instance is named "bond1" and creates the
1031bond1 device in balance-alb mode with an miimon of 50.
1032
1033 In some circumstances (typically with older distributions),
1034the above does not work, and the second bonding instance never sees
1035its options. In that case, the second options line can be substituted
1036as follows:
1037
1038install bond1 /sbin/modprobe --ignore-install bonding -o bond1 \
1039 mode=balance-alb miimon=50
983 1040
1041 This may be repeated any number of times, specifying a new and
1042unique name in place of bond1 for each subsequent instance.
1043
1044 It has been observed that some Red Hat supplied kernels are unable
1045to rename modules at load time (the "-o bond1" part). Attempts to pass
1046that option to modprobe will produce an "Operation not permitted" error.
1047This has been reported on some Fedora Core kernels, and has been seen on
1048RHEL 4 as well. On kernels exhibiting this problem, it will be impossible
1049to configure multiple bonds with differing parameters (as they are older
1050kernels, and also lack sysfs support).
984 1051
9853.4 Configuring Bonding Manually via Sysfs 10523.4 Configuring Bonding Manually via Sysfs
986------------------------------------------ 1053------------------------------------------
987 1054
988 Starting with version 3.0, Channel Bonding may be configured 1055 Starting with version 3.0.0, Channel Bonding may be configured
989via the sysfs interface. This interface allows dynamic configuration 1056via the sysfs interface. This interface allows dynamic configuration
990of all bonds in the system without unloading the module. It also 1057of all bonds in the system without unloading the module. It also
991allows for adding and removing bonds at runtime. Ifenslave is no 1058allows for adding and removing bonds at runtime. Ifenslave is no
@@ -1030,9 +1097,6 @@ To enslave interface eth0 to bond bond0:
1030To free slave eth0 from bond bond0: 1097To free slave eth0 from bond bond0:
1031# echo -eth0 > /sys/class/net/bond0/bonding/slaves 1098# echo -eth0 > /sys/class/net/bond0/bonding/slaves
1032 1099
1033 NOTE: The bond must be up before slaves can be added. All
1034slaves are freed when the interface is brought down.
1035
1036 When an interface is enslaved to a bond, symlinks between the 1100 When an interface is enslaved to a bond, symlinks between the
1037two are created in the sysfs filesystem. In this case, you would get 1101two are created in the sysfs filesystem. In this case, you would get
1038/sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and 1102/sys/class/net/bond0/slave_eth0 pointing to /sys/class/net/eth0, and
@@ -1622,6 +1686,15 @@ one for each switch in the network). This will insure that,
1622regardless of which switch is active, the ARP monitor has a suitable 1686regardless of which switch is active, the ARP monitor has a suitable
1623target to query. 1687target to query.
1624 1688
1689 Note, also, that of late many switches now support a functionality
1690generally referred to as "trunk failover." This is a feature of the
1691switch that causes the link state of a particular switch port to be set
1692down (or up) when the state of another switch port goes down (or up).
1693It's purpose is to propogate link failures from logically "exterior" ports
1694to the logically "interior" ports that bonding is able to monitor via
1695miimon. Availability and configuration for trunk failover varies by
1696switch, but this can be a viable alternative to the ARP monitor when using
1697suitable switches.
1625 1698
162612. Configuring Bonding for Maximum Throughput 169912. Configuring Bonding for Maximum Throughput
1627============================================== 1700==============================================
@@ -1709,7 +1782,7 @@ balance-rr: This mode is the only mode that will permit a single
1709 interfaces. It is therefore the only mode that will allow a 1782 interfaces. It is therefore the only mode that will allow a
1710 single TCP/IP stream to utilize more than one interface's 1783 single TCP/IP stream to utilize more than one interface's
1711 worth of throughput. This comes at a cost, however: the 1784 worth of throughput. This comes at a cost, however: the
1712 striping often results in peer systems receiving packets out 1785 striping generally results in peer systems receiving packets out
1713 of order, causing TCP/IP's congestion control system to kick 1786 of order, causing TCP/IP's congestion control system to kick
1714 in, often by retransmitting segments. 1787 in, often by retransmitting segments.
1715 1788
@@ -1721,22 +1794,20 @@ balance-rr: This mode is the only mode that will permit a single
1721 interface's worth of throughput, even after adjusting 1794 interface's worth of throughput, even after adjusting
1722 tcp_reordering. 1795 tcp_reordering.
1723 1796
1724 Note that this out of order delivery occurs when both the 1797 Note that the fraction of packets that will be delivered out of
1725 sending and receiving systems are utilizing a multiple 1798 order is highly variable, and is unlikely to be zero. The level
1726 interface bond. Consider a configuration in which a 1799 of reordering depends upon a variety of factors, including the
1727 balance-rr bond feeds into a single higher capacity network 1800 networking interfaces, the switch, and the topology of the
1728 channel (e.g., multiple 100Mb/sec ethernets feeding a single 1801 configuration. Speaking in general terms, higher speed network
1729 gigabit ethernet via an etherchannel capable switch). In this 1802 cards produce more reordering (due to factors such as packet
1730 configuration, traffic sent from the multiple 100Mb devices to 1803 coalescing), and a "many to many" topology will reorder at a
1731 a destination connected to the gigabit device will not see 1804 higher rate than a "many slow to one fast" configuration.
1732 packets out of order. However, traffic sent from the gigabit 1805
1733 device to the multiple 100Mb devices may or may not see 1806 Many switches do not support any modes that stripe traffic
1734 traffic out of order, depending upon the balance policy of the 1807 (instead choosing a port based upon IP or MAC level addresses);
1735 switch. Many switches do not support any modes that stripe 1808 for those devices, traffic for a particular connection flowing
1736 traffic (instead choosing a port based upon IP or MAC level 1809 through the switch to a balance-rr bond will not utilize greater
1737 addresses); for those devices, traffic flowing from the 1810 than one interface's worth of bandwidth.
1738 gigabit device to the many 100Mb devices will only utilize one
1739 interface.
1740 1811
1741 If you are utilizing protocols other than TCP/IP, UDP for 1812 If you are utilizing protocols other than TCP/IP, UDP for
1742 example, and your application can tolerate out of order 1813 example, and your application can tolerate out of order
@@ -1936,6 +2007,10 @@ Failover may be delayed via the downdelay bonding module option.
193613.2 Duplicated Incoming Packets 200713.2 Duplicated Incoming Packets
1937-------------------------------- 2008--------------------------------
1938 2009
2010 NOTE: Starting with version 3.0.2, the bonding driver has logic to
2011suppress duplicate packets, which should largely eliminate this problem.
2012The following description is kept for reference.
2013
1939 It is not uncommon to observe a short burst of duplicated 2014 It is not uncommon to observe a short burst of duplicated
1940traffic when the bonding device is first used, or after it has been 2015traffic when the bonding device is first used, or after it has been
1941idle for some period of time. This is most easily observed by issuing 2016idle for some period of time. This is most easily observed by issuing
@@ -2096,6 +2171,9 @@ The new driver was designed to be SMP safe from the start.
2096EtherExpress PRO/100 and a 3com 3c905b, for example). For most modes, 2171EtherExpress PRO/100 and a 3com 3c905b, for example). For most modes,
2097devices need not be of the same speed. 2172devices need not be of the same speed.
2098 2173
2174 Starting with version 3.2.1, bonding also supports Infiniband
2175slaves in active-backup mode.
2176
20993. How many bonding devices can I have? 21773. How many bonding devices can I have?
2100 2178
2101 There is no limit. 2179 There is no limit.
@@ -2154,11 +2232,15 @@ switches currently available support 802.3ad.
2154 2232
21558. Where does a bonding device get its MAC address from? 22338. Where does a bonding device get its MAC address from?
2156 2234
2157 If not explicitly configured (with ifconfig or ip link), the 2235 When using slave devices that have fixed MAC addresses, or when
2158MAC address of the bonding device is taken from its first slave 2236the fail_over_mac option is enabled, the bonding device's MAC address is
2159device. This MAC address is then passed to all following slaves and 2237the MAC address of the active slave.
2160remains persistent (even if the first slave is removed) until the 2238
2161bonding device is brought down or reconfigured. 2239 For other configurations, if not explicitly configured (with
2240ifconfig or ip link), the MAC address of the bonding device is taken from
2241its first slave device. This MAC address is then passed to all following
2242slaves and remains persistent (even if the first slave is removed) until
2243the bonding device is brought down or reconfigured.
2162 2244
2163 If you wish to change the MAC address, you can set it with 2245 If you wish to change the MAC address, you can set it with
2164ifconfig or ip link: 2246ifconfig or ip link: