SUMMARY: Solaris 8 7/01 and network devices

>>> On 04 Oct 2001 12:14:12 -0500, Johnie Stafford <stafforj@core.afcc.com> said:

>>> On 04 Oct 2001 11:13:37 -0500, Johnie Stafford <stafforj@core.afcc.com> said:

 jms> In every version of Solaris previous to Solaris 8 7/01, network
 jms> devices were numbered sequentially starting at 0 (hme0, hme1...;
 jms> qfe0, qfe1...). But starting with Solaris 8 7/01 the devices are
 jms> not numbered starting at 0. For instance, we just installed an
 jms> E10k domain and the qfe ports are numbered qfe4 through qfe11
 jms> for the two cards installed.

 jms> Is it possible to get them to number from 0 again?


 jms> A clarification to my original message. The systems are jumpstarting
 jms> this way. I've had several people suggest modifying the
 jms> /etc/path_to_inst file to correct the problem. I'm looking for
 jms> something that fixes this problem automaticly during jumpstart.

 jms> The responses have led me to look into the path_to_inst file on one of
 jms> the machines in question. It appears that the hardware discovery
 jms> process in Solaris8 7/01 discovers a phantom PCI bus that does not
 jms> exist, and that is where my missing qfe instances are assigned.


Thanks to everybody that replied. The path_to_inst responses got me
going in the right direction. The problem turned out to be a corrupt
jumpstart image. 

	Johnie