Hi managers -- got a couple of responses to my question. Much thanks to:
Scott F. Woods <sfw@adc.idt.com>
Jay Morgan <jmorgan@qualix.com>
The consensus is that this is probably a driver problem leading to failure
of the auto-negotiation. The fix is to turn off autoneg in /etc/system (or
via appropriate ndd commands) and select the speed/duplex by hand, much like
you have to do to run 100Mbps full duplex.
I made a support call to Sun and they came to the same conclusion.
This is standard 10Mbps half-duplex ethernet. Nothing fancy. It's pretty sad
that a $50 PC card can always get it right when these $500 Sun cards blow it
20% of the time. :^P
But enough sour grapes. My sincerest thanks to all those who helped!
----- Begin Original Question -----
[Some portions deleted for brevity...]
Hi Managers -- We've recently put an Ultra Enterprise 450 (2x250MHz CPUs,
512MB RAM, Solaris 2.6) into service and are having an intermittent
problem. About 1 out of every 5-6 times we reboot the machine (using 'init
6', for example), the ethernet interfaces don't come up. We see the
following messages (trimmed for brevity):
SunOS Release 5.6 Version Generic [UNIX(R) System V Release 4.0]
Copyright (c) 1983-1997, Sun Microsystems, Inc.
mem = 524288K (0x20000000)
avail mem = 516227072
Ethernet address = 8:0:20:xx:yy:zz
root nexus = Sun Ultra Enterprise 450 (2 X UltraSPARC-II 248MHz)
[...lots of trimming here...]
SUNW,hme0: CheerIO 2.0 (Rev Id = c1) Found
PCI-device: network@1,1, hme #0
SUNW,hme0 is /pci@1f,4000/network@1,1
SUNW,hme1: Local Ethernet address = 8:0:20:aa:bb:cc
PCI-device: SUNW,hme@4,1, hme #1
SUNW,hme1 is /pci@1f,4000/SUNW,hme@4,1
SUNW,hme0: Link Down - cable problem?
essage repeated 2 times
SUNW,hme0: Link Down - cable problem?
pseudo-device: pm0
pm0 is /pseudo/pm@0
pseudo-device: tod0
tod0 is /pseudo/tod@0
pseudo-device: vol0
vol0 is /pseudo/vol@0
SUNW,hme0: Link Down - cable problem?
fdthree0 at ebus0: offset 14,3023f0
fdthree0 is /pci@1f,4000/ebus@1/fdthree@14,3023f0
SUNW,hme0: Link Down - cable problem?
SUNW,hme1: Link Down - cable problem?
[...and so on: messages repeat every 4 seconds, ad infinitum]
hme0 is the builtin interface and hme1 is an X1033A SunSwift Fast Ethernet
adapter (basically the same thing as the builtin, but on a PCI card). We are
only running standard 10Mbps half-duplex right now using the RJ-45
connector. And, indeed, there is no link - the link light on the hub is off
(I know this doesn't mean too much, but thought I'd mention it).
When this happens, _both_ interfaces are always dead - this leads me to
suspect a driver problem initializing the card. Normally, the cards will say
something like:
SUNW,hme0: Using Internal Transceiver
SUNW,hme0: 10 Mbps half-duplex Link Up
Which seems to imply that auto-negotiation is failing. But, like I mentioned
earlier, only about 20% of the time. Anybody seen anthing like this? Thanks
for your comments!
----- End Original Question -----
JLB
-- Jody L. Baze Blue Sky Tours, Inc. Software Development 10832 Prospect Avenue N.E. System Administration Albuquerque, NM 87112 jody@BlueSkyTours.COM (505) 293-9462
This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:12:12 CDT