Stephan, I will research this after I get Scientific Linux 4 out the door(today and tomorrow) -Connie Sieh On Wed, 20 Apr 2005, Stephan Wiesand wrote: > Hi Yannick, Martin et al, > > thanks for your responses. As a stopgap measure, I put the kickstart file > onto a floppy, and now kickstart installs are humming away. The first one > just finished the postinstallation without problems, and there are no > indications of anything being wrong with the NIC. > > Lspci identifies those cards as 82541GI/PI: Class 0200: 8086:1076 (rev > 05). The servers are Dell 2850s (no, we still prefer opterons ;-). > > A Dell 1850 we set up in January has NICs with exactly the same PCI, > and back then kickstart just worked, three out of three times. > > I still think the problem is with the installer, and that there's > a good chance that applying the patched pump/anaconda-runtime/netconfig > - if not yet done - might save many SL sites some grief. > > Thanks again, > Stephan > > > > On Wed, 20 Apr 2005, Bly, MJ (Martin) wrote: > > > Stephan et al, > > > > Not sure this has any bearing on the situation but we had some problems > > with two particular servers with e1000 NICs that refused to talk > > correctly to some new Nortel switches, whereas their identical brethrin > > worked fine. The two NICs have the 82543GC chip and are tagged at the > > PRO/1000T server adapter, the rest were one of the other (older?) 8254X > > chips. The systems are (still) running RedHat 7.2 so the leap is rather > > large to SL 3. > > > > All our other e1000 NICs work fine with SL3.0.X. > > > > Being in a hurry, we didn't investigate, choosing to swap out the cards > > for known good ones. Ideas at the time were that the autonegotiation > > wasn't working but as I say, we didn't persue the problem. > > > > Martin. > > > >> -----Original Message----- > >> From: [log in to unmask] > >> [mailto:[log in to unmask]] On > >> Behalf Of Stephan Wiesand > >> Sent: 20 April 2005 16:26 > >> To: [log in to unmask] > >> Subject: 3.0.4 kickstart problem w/ e1000 > >> > >> > >> Hi All, > >> > >> I'm having trouble getting kickstart installations going on a > >> couple of > >> servers with intel gigabit NICs. > >> > >> The symptoms match what's described in > >> https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=110036 > >> > >> Initial network setup works, but after the NIC is brought up for the > >> second time, things just fail. Even the link indicator is > >> off, and the > >> switch doesn't think there's anything connected to the other > >> end of the > >> cable any longer (switches are cisco gear, and we tried > >> different models > >> even running different OSs). > >> > >> A non-kickstart NFS installation works fine (at least up to the point > >> where I see the graphical installer - haven't really done it yet). > >> I can also boot a 3.0.4 rescue system and bring the interface > >> up and down > >> and up again many times in a row without problems. > >> > >> Red Hat issued an erratum > >> http://rhn.redhat.com/errata/RHEA-2004-518.html > >> > >> that's supposed to fix these problems. I wonder: did these > >> fixes make it > >> into what's used during a kickstart installation of SL 3.0.4? > >> > >> Any additional hints are much apreciated. > >> > >> > >> Thanks, > >> Stephan > >> > >> -- > >> > >> ---------------------------------------------------- > >> | Stephan Wiesand | | > >> | | | > >> | DESY - DV - | phone +49 33762 7 7370 | > >> | Platanenallee 6 | fax +49 33762 7 7216 | > >> | 15738 Zeuthen | | > >> | Germany | | > >> ---------------------------------------------------- > >> > > > >