Red Hat Bugzilla – Bug 211286
anaconda tries to configure QLA4xxx HBA as an eth device
Last modified: 2007-11-30 17:07:35 EST
Description of problem:
Version-Release number of selected component (if applicable):
RHEL 5 beta 1 M4 & M5
Steps to Reproduce:
1. Start a fresh install with a QLA4xxx HBA in the server
anaconda displays qla4xxx and NICs for the server. This is only an issue
because the TOE NIC driver is not included in the kernel and the eth numbers
displayed by anaconda don't match the kernel. The two HBAs get eth0 and eth1
and the two NICs were given eth2 and eth3 by anaconda; however, when you you to
configure the NICs they both fail. I found that by selecting eth0 and eth1 I
was able to configure the NICs.
Either the QLA4xxx HBAs should not be listed as ethernet devices and only iSCSI
hardware initiators, the qla3xxx driver is placed in the kernel so the QLA4xxx
HBA can be used as a TOE NIC, or anaconda should not accept the QLA4xxx as
An lspci() will display two devices for a QLA4xxx HBA. One is the qla3xxx which
is a TOE NIC and the other is a qla4xxx which is an iSOE (iSCSI initiator). I
believe QLogic reports the HBA as a network device and this will cause issues.
anaconda needs to recognize the HBA as an iSCSI initiator and load the driver
so it can be used to configure a boot device.
Can you provide the output of lspci, lspci -nv, and kudzu -p?
The qla3xxx networking driver was included in kernel 2.6.18-1.2715.el5 on Thu
Oct 05 2006. Please test this again in Beta 2.
Created attachment 138876 [details]
Attached is the output of the commands Jeremy requested.