SUMMARY: V480 with ce NIC's

Hello,

For those who are interested, here is the information that I received.  Much
thanks to all of those who replied:
	Aaron Hirsch
	Chris Price
	Bob Cummings
	John Riddoch
	mmcmanus@houston.geoquest.slb.com

Look for Infodoc # 41665
GigaSwift Gigabit Ethernet - Further Information
>From the FAQ section:
=====================
2.How do I see the actual Operational Mode settings ? 
  Check link_ parameters of netstat -k ce[your_interface.number]
     link_up - 0 down, 1 up 
     link_speed - speed in Mbit/s 
     link_duplex - 1 half duplex, 2 full duplex, 0 down
  This parameters must show the same values on both sides.
  If not, Autonegotiation does not work. 
You'll have to set for both cards if you want to get them both to 100/FD.
For example, here's what I set for a QFE card in a server:
	for i in 0 1 2 3; do
	  ndd -set /dev/qfe instance $i
	  ndd -set /dev/qfe adv_100T4_cap 0
	  ndd -set /dev/qfe adv_100fdx_cap 1
	  ndd -set /dev/qfe adv_100hdx_cap 0
	  ndd -set /dev/qfe adv_10fdx_cap 0
	  ndd -set /dev/qfe adv_10hdx_cap 0
	  ndd -set /dev/qfe adv_autoneg_cap 0
	done
Alternatively, if you set it in /etc/system, it should carry for all cards.
Did you turn off gigabit capability for the devices (ce0,ce1)? That one
stumped me for a few days.

Thanks again,
John

>  -----Original Message-----
> From: 	Olson, John C  
> Sent:	Monday, November 25, 2002 9:39 AM
> To:	'sunmanagers@sunmanagers.org'
> Subject:	V480 with ce NIC's
> 
> Hello,
> 
> Had an unusual problem the other night and was hoping that someone else
> could explain what happened.  
> 
> Implementing 4 V480's with 2 NIC's each into a an environment where they
> are each connected to the same two switches (which are hard-configured to
> 100/full duplex).  I have a startup script that runs at boot that send ndd
> commands to turn off autonegotiate and turn off all settings but 100fdx to
> /dev/ce.  That seemed to work fine and ce1 was working perfectly.  The
> problem was ce0 on all four servers was not talking to anyone.  As they
> were all on the same subnet, talking to the same switch on the same VLAN,
> there was no routing issues to contend with.  After a while of not
> figuring anything out, I re-ran the startup script on only one of the four
> servers.  At that point, they all started talking to each other.  Any
> ideas why that would happen?
> 
> Also, ran the 'netstat -k' command and was looking at the results for the
> ce0 and ce1 portions and could not translate a few of the options (for
> example link_duplex=2).  Does anyone know of documentation on the ce
> interface that explains all of the status codes?  I have the 'Platform
> Notes: Sun(tm) GigaSwift Ethernet Device Driver' pdf that has some info
> but not all...
> 
> Thanks,
> John
_______________________________________________
sunmanagers mailing list
sunmanagers@sunmanagers.org
http://www.sunmanagers.org/mailman/listinfo/sunmanagers