From: Geert Devos (
Date: Thu May 02 1996 - 17:22:48 CDT

Hi all,

You guys just stump me. The turn-around on this kind of question (pretty
new stuff like the SunSwift card I mean) is like nothing I have ever seen

I have to apologize to everybody, if not in person then as a cenacle of
wisdom. I (unvoluntarilly) lead all of you, bar one, on a wild goose chase.

I didn't mention this is a system that leads a life as unpayed whore
(excuse me for the _very_ blunt expression, but it's true) since it is used
for testing purposes of _any_ kind (how do I load a system on I/O so that
it crashes, what is the live throughput on any kind of network or SCSI
interface from any vendor, how do I get a backup restored when somebody
trashed the Legato Database, ...; the works).

Six weeks ago I co-installed a SunSwift adapter for testing purposes (I
didn't do the complete installation, but that should be no excuse), and
completely forgot about it.
Back then we installed the SunSwift adapter in another slot (all slots are
open on the back, to make installation of a card easier) in the SS20. After
removing the SunSwift and "boot -r" everything worked just fine, so "us
two" assumed the card had been removed from the proper config files
alltogether. Little did we know ....

What should have happened on "boot -r" but clearly didn't, was that the
previous instance of the SunSwift adapter card should have been completely
removed in /etc/path_to_inst.

Upon "boot -r", the Sun saw a new instance of a SunSwift adapter, and
decided to place the Fast Ethernet interface on "hme1", because "hme0
already existed". Why this didn't happen with the F&W controler still has
me stumped ; it should have behaved the same way if you ask me.

Today I was in contact with SunService, and they asked me to :
"cat /etc/path_to_inst|grep hme".

This showed me two clear instances of "hme".

This was a giveaway, and I just tried "hme1" fpr a change. When this worked
fine I edited /etc/path_to_inst to reflect the current config.

After rebooting the system, everything was just fine, apart from a small
netmask problem that took just a couple of minutes to rectify and reboot.

I didn't try the solution proposed by Andi Paton <>
because I didn't retrieve my mail untill later today, but I will try this
tomorrow. I'll let you guru's know tomorrow what the result is (I'll just
move the d... card to another slot for that and go the same route again).

Once again, and with a very deep red blush on my cheeks, I have to admit I
managed to put the proverbial foot where everybody knows it lands from time
to time..... I'm in the process of kicking the habbit, BTW.

Thanks to everybody who responded.


start original messages
Hi all,

Take one SPARCstation 20/50, running Sol 2.5, necessary patches installed,
one network interface (onboard), a couple of hard disks attached to it. RUn
of the mill system.

Add the drivers that packages that come with the card, and install the
patches that come on the same cdrom. Halt the system.

Add one SUNSWIFT adapter card (F&W SCSI + 100Base-T), attach a F&W RAID
cluster to it, hook the 100Base-T to a standard 10Base-T hub (the 100Base-T
hub arrived later on the day), find the card at ....hme@0... at the ok
prompt, and then "boot -r" the system.

The RAID cluster is recognized with no problem, and I get no grumbles
concerning the extra network interface. Formatting the RAID and putting a
working file system on it creates no problems.

But when I try to do something with the "hme" interface it just remains
invisible, non-existant, nada, no-go, ....

I _did_ create the /etc/hostname.hme0, and added the interface with a valid
IP address in the /etc/hosts file. And I rebooted the system afterwards (-r
of course).

The only thing my system says during bootup is that there's no such
interface as hme0 on my system. And of course it doesn't turn on routing
between both interfaces, which is the initial idea.

Fishy detail : I installed an identical card on a Ultra1/140 last month in
the same way, and that worked from the start.

Anybody got any idea about this problem?

Will summarize, of course.


start included messages with >>new remarks<<

From: "Todd A. Radermacher" <>

You may have to manually "plumb" the interface....check the man page.


From: anthony baxter <>

FWIW, when I put a Sun 100baseT card in a Solaris 2.3 SS10, the device
it was visible as depended what SBUS slot it was in. Try creating
/etc/hostname.hme{0,1,2,3,4,5}, and see if that helps.

>>Obvious. Worked just fine when I tried it.<<


From: Casper Dik <casper@holland.Sun.COM>

Do you need to install the drivers? They come standard with Solaris 2.5,
the drivers are probably for Solaris 2.4 only.

>>Right, Casper, you don't need to install them on 2.5<<
>>But the included RTFM told me to look for some "SUNWhmdl" and "SUNWhmdlu"<<
>>packages. It turns out they weren't included in the final distribution...<<

Have you tried "ifconfig hme0 plumb"?

Any sign of the hme card in /var/adm/messages?

>>Yep, from start it read "...hme0:no such interface...", but no "no carrier"<<
>>or anything of the sort. The interface just wasn't there after the reboot.<<



From: (Alan Pickett)


have you tried a show-devs at the boot prom? Look for a
line with hme and you know that the hardware is seen by the

>>Yep, told me "...hme@0....", setting off my short fuse<<

Next try ifconfig hme0 plumb
This should make the interface available for all other ifconfig

Good look!



From: Jon Maiman <>

        Just a guess, but do you need to isssue an ifconfig plumb command
on the interface? I.e.: ifconfig hme0 plumb. For standard 10base-T
lance ethernet interfaces you have to issue this command to initialize the
streams drivers. Hope this helps.


From: Andi Paton <>


rem_drv hme
add_drv -i "SUNW,hme" hme
ifconfig hme0 plumb

>>This looks like you've had this kind of problematic system before.<<
>>I'll try this one out on May 3rd, and add to the summary.<<


From: (John Malick)

When you installed Solaris 2.5 did you install the clusters necessary or after
installation install the clusters:

1. SUNWhmd
2. SUNWhmdu

end included messages

- Geert Devos -
- Fax @ work +32 2 451 13 91 - Voice @ work +32 2 451 12 69 -
- system administrator/system integrator/chief numskull -
- and about to leave Graphidec for good (May 6th, to be precise) -
-A committee is a life form with six legs or more and no brain (R. A. Heinlein)-


This archive was generated by hypermail 2.1.2 : Fri Sep 28 2001 - 23:10:59 CDT