Minor_init failed for module... & unable to configure nics

Question

Hello,
following the instructions of: "Sun Two-Node Cluster How to Guide"
I successfully installed the "Sun Cluster 3.1U4" on my sun ultra 5 workstations (on the two nodes and on the administrative console), which are all running solaris 10 1/06.
During the installation process I choose the option "configure later" as suggested by the guide.
Before configuring the nodes and proceeding with the steps, I wanted to check the correct installation of the nics. I have three nics plugged in the two nodes and two in the administrative console. One per computer is integrated, the others are three pci 10/100 Ark and two Intel nics.
The fact is that the output of dladm show-link is just hme0 for all the computers, none of the other nics seems to be recognized.
in the output of prtconf I see
pci instance 1 ethernet (drive not attached) ethernet (driver not attached)
instead with prtconf -D
pci instance 1 ethernet ethernet (driver simba)
this is the output of one of my nodes that has two Ark nics.
I tried with devfsadm -v (as I read in a post) and the output is
minor_init failed for module /usr/lib/devfsadm/linkmod/SUNW_scmd_link.so
this string also appears at every boot (also when I boot -r)
This last thing I found treated as a problem in "sun java enterprise administration guide":
this is like my situation, but of course I am not running solaris on an x86 system, and I also do not receive "NOTICE NO PCI PROB"
x86 machines running Solaris 10 fail to come up in cluster mode due to changes made for the Solaris boot architecture project. The following error messages are displayed when the machine boots up:
NOTICE: Can�t open /etc/cluster/nodeid
NOTICE: BOOTING IN NON CLUSTER MODE
NOTICE: NO PCI PROP
NOTICE: NO PCI PROP
Configuring devices.
Hostname: pvyom1
devfsadm: minor_init failed for module /usr/lib/devfsadm/linkmod/SUNW_scmd_link.so
Loading smf(5) service descriptions: 24/24
/usr/cluster/bin/scdidadm: Could not load DID instance list.
Cannot open /etc/cluster/ccr/did_instances.
Not booting as part of a cluster
/usr/cluster/bin/scdidadm: Could not load DID instance list.
Cannot open /etc/cluster/ccr/did_instances.
Note: path_to_inst might not be updated. Please �boot -r� as needed to update.
Solution Perform these steps:
1. Add etc/cluster/nodeid to /boot/solaris/filelist.ramdisk.
2. Enter these commands:
# bootadm update-archive
# reboot -- -r
I did not perform the steps indicated as the "solution" for this.
I wait because maybe the scinstall configuration can fix this, doesn't it?
I just want to end my two node cluster config happily and begin to work with it.
Thanks in advantage,
Francesco

Answer 1

The primary problem you seem to have is that your network adapters are not recognized. So this is not a Cluster problem. You talk about "three pco 10/100 Ark and two intel nics". The question is, if those cards have a corresponding solaris driver. Since you seem to indicate that prtdiag does report that there is no driver attached, you may be out of luck for the drivers coming with Soalris 10 01/06 you use. Maybe the vendor of those cards offers drivers that need to get installed?
Otherwise I guess you are out of luck using them, without finding a usabel kernel driver.
For the message
minor_init failed for module /usr/lib/devfsadm/linkmod/SUNW_scmd_link.so
you see, this is harmless and not your problem. You can ignore it.
This error happens when the scmd_link devfsadm link module is attempted to be
loaded before clustering has started. It is a harmless situation.
Greets
Thorsten