-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, I'm trying to kickstart a bunch of boxes with a RH 7.3 Kickstart machine. I've tried to do this on multiple boxes we have with the same hardware. They are generic dual processor boxes with 2 hard drives and a Serial Console card. The motherboards are Tyans with built on Ethernet. (I can supply the whole boot litany if necessary). I boot the boxes from the Disk 1 CDROM (2.4.18-3BOOT kernel, RH 7.3 2.96-110). The console says: <6>USB Mass Storage support registered. <4>3c59x: Donald Becker and others. www.scyld.com/network/vortex.html <6>02:08.0: 3Com PCI 3c905C Tornado at 0x2000. Vers LK1.1.16 <135>Oct 10 22:19:10 loader: PUMP: sending discover <135>Oct 10 22:19:10 loader: breq: opcode: 1<135>Oct 10 22:19:10 loader: breq: h w: 1<135>Oct 10 22:19:10 loader: breq: hwlength: 6<135>Oct 10 22:19:10 loader: b req: hopcount: 0<135>Oct 10 22:19:10 loader: breq: xid: 0x217e8bdd<135>Oct 10 22 :19:10 loader: breq: secs: 0<135>Oct 10 22:19:10 loader: breq: flags: 0x0000<135 >Oct 10 22:19:10 loader: breq: ciaddr: 0.0.0.0<135>Oct 10 22:19:10 >loader: breq: yiaddr: 0.0.0.0<135>Oct 10 22:19:10 loader: breq: server_ip: 0.0.0.0<135>Oct 10 22:19:10 loader: breq: bootp_gw_ip: 0.0.0.0<135>Oct 10 22:19:10 loader: breq: h waddr: <135>Oct 10 22:19:10 loader: breq: servername: <135>Oct 10 22:19:10 loade r: breq: bootfile: <135>Oct 10 22:19:10 loader: breq: vendor: 0x63 0x53 0x82 0x6 3<135>Oct 10 22:19:10 loader: breq: vendor: 53 1 0x01<135>Oct 10 22:19:10 lo So it seems to find a 3C905C ethernet card. It also says (another console) * finished bus probing * found suggestion of 3c59x * found 3c59x device * found suggestion of usb-ohci * modules to insert 3c59x * inserted /tmp/3c59x.o * load module set done * sending dhcp request through device eth0 * pump told us: No DHCP reply received * No dhcp response received Why is it trying to load 3C59x when it says the card is a 3C905C? I've booted 3 other machines on this subnet (one a kickstart client) and they've all received DHCP addresses just fine. I can also see (and capture) the DHCP conversation with tcpdump where the client asks for an address (and gets a response) 5 times! When I manually ifconfig the eth0 interface on the kickstart client from a # prompt, I get ridiculous amounts of packet loss when pinging the kickstart server. No other machines on the subnet have that problem (none are 3C905Cs either). I've changed ports, cables, etc. to no avail. The switch is set to autoneg and the link looks like 100 FDX. I read in some archives about this card having problems with old versions of PUMP or Linux kernels (and dhcp) but the posts are from more than a year ago. Any suggestions? Ideas? Comments? Any help would be greatly appreciated. Thanks. Dave Mangot CDN Systems/Infrastructure Architect Exodus, A Cable & Wireless Service -----BEGIN PGP SIGNATURE----- Version: PGPfreeware 6.5.8 for non-commercial use <http://www.pgp.com> iQA/AwUBPTy7CgyKhoWak/IMEQKAfQCfZJwwx1ZupvM44Fvlsua3tfk6pGMAn2q3 gZ1AkfsRu82YXYQdfQhq4zyj =e2Ai -----END PGP SIGNATURE-----
