Link aggregation and IPMP are not the same and do not serve the same purpose. Some of the features overlap, but both can be used together for both higher performance (throughput) and better HA. I recommend the following blog for good information on the capabilities of the two:
http://blogs.sun.com/droux/entry/link_aggregation_vs_ip_multipathing An example of using the two together is as follows (T5240): - aggregate nxge0 (p...@400) and nxge4 (p...@500) as aggr1 - cable aggr1 to switchA (ether-channel on switch) - set aggr1 as the vsw0 switch (map vsw0 to switchA) - aggregate nxge1 (p...@400) and nxge5 (p...@500) as aggr2 - cable aggr2 to switchB (ether-channel on switch) - set aggr2 as the vsw1 switch (map vsw0 to switchA) - create vnet0 (mapping: vnet0 -> vsw0 -> aggr1 -> switchA) - create vnet1 (mapping: vnet1 -> vsw1 -> aggr2 -> switchB) - setup IPMP on vnet0 and vnet1 (both link based and probe based IPMP works in LDoms 1.3) Sample output showing IPMP (probe & link based monitoring) in the control domain: <code> [r...@control] # dladm show-aggr key: 1 (0x0001) policy: L4 address: <MAC> (auto) device address speed duplex link state nxge0 <MAC> 1000 Mbps full up attached nxge5 <MAC> 1000 Mbps full up attached key: 2 (0x0002) policy: L4 address: <MAC> (auto) device address speed duplex link state nxge1 <MAC> 1000 Mbps full up attached nxge4 <MAC> 1000 Mbps full up attached [r...@control] # ldm ls -o network primary NAME primary MAC <MAC> VSW NAME MAC NET-DEV ID DEVICE LINKPROP DEFAULT-VLAN-ID PVID VID MTU MODE primary-vsw0 00:14:4f:fb:e0:40 aggr1 0 swi...@0 1 1 1500 primary-vsw1 00:14:4f:f9:9e:de aggr2 1 swi...@1 1 1 1500 NETWORK NAME SERVICE ID DEVICE MAC MODE PVID VID MTU LINKPROP vnet0 primary-v...@primary 0 netw...@0 00:14:4f:f9:25:39 1 1500 phys-state vnet1 primary-v...@primary 1 netw...@1 00:14:4f:f8:08:a9 1 1500 phys-state [r...@control] # ifconfig -a lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 vnet0: flags=9040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER> mtu 1500 index 2 inet 10.10.10.24 netmask ffffffc0 broadcast 10.10.10.63 groupname IPMP ether 0:14:4f:f9:25:39 vnet0:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2 inet 10.10.10.5 netmask ffffffc0 broadcast 10.10.10.63 vnet1: flags=69040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER,STANDBY,INACTIVE> mtu 1500 index 3 inet 10.10.10.25 netmask ffffffc0 broadcast 10.10.10.63 groupname IPMP ether 0:14:4f:f8:8:a9 [r...@control] # </code> Note: The "phys-state" property must be set on any vnet devices for link-based monitoring. The following shows a guest LDom using IPMP (link-based only) from the above control domain: <code> [r...@control] # ldm ls -o network test-domain NAME test-domain MAC 00:14:4f:f8:de:ce NETWORK NAME SERVICE ID DEVICE MAC MODE PVID VID MTU LINKPROP vnet0 primary-v...@primary 0 netw...@0 00:14:4f:fa:8c:9f 1 1500 phys-state vnet1 primary-v...@primary 1 netw...@1 00:14:4f:f8:17:41 1 1500 phys-state # ifconfig -a lo0: flags=2001000849<UP,LOOPBACK,RUNNING,MULTICAST,IPv4,VIRTUAL> mtu 8232 index 1 inet 127.0.0.1 netmask ff000000 vnet0: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2 inet 10.10.10.27 netmask ffffffc0 broadcast 10.10.10.63 groupname grp1 ether 0:14:4f:fa:8c:9f vnet0:1: flags=1000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4> mtu 1500 index 2 inet 0.0.0.0 netmask ff000000 broadcast 0.255.255.255 vnet1: flags=69000842<BROADCAST,RUNNING,MULTICAST,IPv4,NOFAILOVER,STANDBY,INACTIVE> mtu 0 index 3 inet 0.0.0.0 netmask 0 groupname grp1 ether 0:14:4f:f8:17:41 # </code> -- This message posted from opensolaris.org _______________________________________________ ldoms-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/ldoms-discuss
