Im assuming that the default route on the filer is being used to send out traffic destined for 151.142.223.0/24.
Not on a good day it won't! :-) The filer's networking code should be more than sharp enough to see that it has an interface directly connected to your 151.142.223.0 network and route packets targeted to this network to said interface accordingly. In fact, a "netstat -r" on the console should show you an explicit zero-metric route for each network, directing traffic to the interface on which that network is to be found (look for gateway targets that look like "link#1", "link#2" in the "Gateway" column).
What should I do to make the e1a interface active to the 151.142.223.0/24 network?
I read your e-mail twice and can't spot a configuration mistake. Everything looked fine in fact. Is it possible that there is a more fundamental problem? A bad cable connecting the filer to the switch? A VLAN configuration within the switch that is preventing the filer's port from communicating with the Sun ports?
A complete examination of the routing tables on filer, Suns and Cisco might reveal something peculiar. As an act of desperation, bringing the filer up without the FDDI interface configured (such that it is no more than a single Ethernet interface system connected via a switch to three other single Ethernet interface systems) might also be an experiment worthy of a shot.
Keith
Not on a good day it won't! :-) The filer's networking code should be more than sharp enough to see that it has an interface directly connected to your 151.142.223.0 network and route packets targeted to this network to said interface accordingly. In fact, a "netstat -r" on the console should show you an explicit zero-metric route for each network, directing traffic to the interface on which that network is to be found (look for gateway targets that look like "link#1", "link#2" in the "Gateway" column). ---
That too is what I thought.
*hmm*
The filer can ping itself on 223.30, the router and suns can talk to each other, but thats it for anything else.
Another cable might fix this, but the cable on there now, is new and certified...will test on a working sun server.
Oook.
Found what it was.
Ive been using the GUI exclusivly for setting simple things, although I can drive the CLI as well.
What I found, was that somehow lo got set to the IP address of 151.142.223.30, the one I was trying to set for e0, or e1a-d with no sucess.
I wasnt able to remove the address from lo, so I just booted the sucker...and did it all by hand.
No problems.
Im yet to recreate the problem, but on my spare filer heads that are waiting patiently for cluster-failover, I will try to reproduce before I make any bug reports.
Also, Ive noticed a max input rate of about 20Mb/sec into the filter, then the CPU tops out on my F760.
I understand that the NVRAM can only process 16Mb at a time before the system really gums up, but I did expect a little more input performance.
Does anyone have any official or un-official comments on this?
-
Jeff Mohler -
Keith Brown