IN response to #3, Can you maybe do LAG/LACP?
Take a look at dladm, it creates an aggregate interface that support
(primitave) load balancing and fail over.
# dladm create-aggr -l active -d e1000g0 -d e1000g1 1
# dladm show-aggr
key: 1 (0x0001) policy: L4 address: 0:14:4f:82:55:be (auto)
device address speed duplex link
state
e1000g0 0:14:4f:82:55:be 1000 Mbps full up
attached
e1000g1 0:14:4f:82:55:bf 1000 Mbps full up
attached
# ldm ls-bindings primary
Name: primary
State: active
Flags: normal,control,vio service
OS: Solaris running
Util: 0.7%
Uptime: 1d 16h 48m
[snip]
Vsw: my-infrasvc-vsw
mac-addr=0:14:4f:f8:82:d9
net-dev=aggr111001
mode=prog,promisc
Vsw: my-admin-vsw
mac-addr=0:14:4f:f9:30:57
net-dev=aggr110001
[LDC: 0xe]
[LDom hrc-admin-global, name: vnet1, mac-addr: 0:14:4f:fa:f:43]
mode=prog,promisc
[snip]
In this case I have an interface "aggr1" that is both e1000g0 and e1000g1. Each
interface is plugged into a different blade of my force 10 switch with LACP
enabled on the ports. I have interfaces on vlan 110 and vlan 111 using 802.1q
vlan tagging on the switch.
Tommy
--
This message was posted from opensolaris.org
