I have the following IPMP configuration:

r...@dsshare02:~# cat /etc/hostname.e1000g0 
group ipmp0 -failover dsshare02-net0 up
r...@dsshare02:~# cat /etc/hostname.e1000g1
group ipmp0 -failover dsshare02-net1 up
r...@dsshare02:~# cat /etc/hostname.ipmp0 
ipmp group ipmp0 dsshare02 up

I have set the MTU for my physical interfaces according to this document: 
http://www.sun.com/bigadmin/sundocs/articles/nicddladmconf.jsp

Here is the output from ifconfig -a (sans lo0):

ipmp0: flags=9001000843<UP,BROADCAST,RUNNING,MULTICAST,IPv4,FIXEDMTU,IPMP> mtu 
1500 index 2
        inet 10.93.7.56 netmask ffff0000 broadcast 10.93.255.255
        groupname ipmp0
e1000g0: 
flags=1009040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER,FIXEDMTU>
 mtu 9216 index 5
        inet 10.93.7.54 netmask ff000000 broadcast 10.255.255.255
        groupname ipmp0
        ether 0:14:4f:20:da:cc 
e1000g1: 
flags=1009040843<UP,BROADCAST,RUNNING,MULTICAST,DEPRECATED,IPv4,NOFAILOVER,FIXEDMTU>
 mtu 9216 index 6
        inet 10.93.7.55 netmask ff000000 broadcast 10.255.255.255
        groupname ipmp0
        ether 0:14:4f:20:da:cd


The MTU is set correctly for the physical interfaces, howver the iPMP interface 
isn't picking it up. I trues adding 'mtu 9216' to /etc/hostname.ipmp0 but that 
didn't work.

According to a PDF i found on this site about ipmp,

"The MTU of data packets will be controlled by the IPMP interface’s MTU, which 
will be set to the smallest MTU supported by all of the interfaces currently in 
the group. Thus, if an IPMP group is comprised of two interfaces, one with a 
1400 byte MTU, and another with a 1500 byte MTU, then the IPMP interface will 
have
a 1400 byte MTU. If the first interface is removed from the group, the MTU will 
be automatically
raised to 1500 bytes. "

I would think that ifconfig -a would reflect the ipmp interface changing MTU to 
the appropriate one.

Can someone provide assistance with modifying my ipmp mtu?
-- 
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]

Reply via email to