JHBrewer Posted June 26, 2004 Report Share Posted June 26, 2004 I've been happily running a LAN under Mandrake 9.2 with a cable connection to eth0 (a DEC Tulip board) and a hub connected to eth1 (a 3COM board). So I decided to upgrade to Mandrake 10. First I tried Upgrading. When it finished, eth0 was assigned to the 3COM board and eth1 did not exist, notwithstanding the fact that HardDrake could see the Tulip just fine. It simply didn't assign it to eth0 or eth1. So I figured it was "another one of those Upgrade conflicts" and Installed clean instead. Except for the extra work, this had no effect. Still no eth0. So I switched the cables in order to be able to use the newly assigned eth0 for my cable connection -- which works, but of course now I have a new IP. However, none of the GUI config tools can recognize the Tulip board, presumably since it is not assigned to eth0 and there IS NO eth1 any more. So my LAN is dead until I either find some command-line way to assign eth1 or switch back to Mandrake 9.2 (or, more likely, bail to another distro). This doesn't seem like much of an improvement to me. I hope someone can explain (a) what people were thinking when they made these changes; and (B) what, if anything, I can do to recover under Mandrake 10. Quote Link to comment Share on other sites More sharing options...
streeter Posted June 27, 2004 Report Share Posted June 27, 2004 (edited) First, we'll see if the module has actually been loaded - type "lsmod | grep tulip" at the command line. (assuming it is the tulip driver you need - post exact card type) If you are rewarded with a line starting with tulip, then the card has indeed been recognised and the driver loaded. If you get no output, type modprobe tulip as root You can then set up the card by typing ifconfig eth1 xxx.xxx.xxx.xxx up where xxx.xxx.xxx.xxx is the required IP address. Test with ping xxx.xxx.xxx.xxx If this all works, you should be able to force it by putting alias eth0 tulip alias eth1 <your 3com driver name> in /etc/modprobe.conf - you may already have the 3com entry there. Change eth0 and eth1 to suit. Then check out the howtos referred to in the sticky posts at the top of the networking forum for more info. If you have problems, post output from above commands, and also from the ifconfig command. Chris Edited June 27, 2004 by streeter Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted June 28, 2004 Author Report Share Posted June 28, 2004 "lsmod | grep tulip" yields "tulip 45248 0" so I guess it has the driver loaded OK, although "modprobe tulip" (I thought I'd check) yields nothing. But then "ifconfig eth1 192.168.0.1 up" yeilds SIOCSIFADDR: No such device eth1: unknown interface: No such device eth1: unknown interface: No such device which confirms that there is no eth1 defined. I've been trying to discover exactly where such things as eth0, eth1 etc. are defined, but all the docs I've seen just assume they are always there -- which has always been the case before Mdk 10. I added "alias eth1 tulip" to modprobe.conf but that doesn't seem to have any effect either. Thanks for the help, though. I'm not really helpless, but I'm at a loss as to where I should go to read up on this stuff. Any suggestions? Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted June 30, 2004 Author Report Share Posted June 30, 2004 A little more perhaps-relevant detail: Somewhere I read that I should check the contents of /proc/net/dev so I had a look. It contains a line for eth0, one for lo and one for sit0 whereas on Mandrake 9.2 it had one for lo, one for eth0 and one for eth1. Evidently sit0 is a new invention. Under 9.2 there was also a file /proc/net/dev_mcast with two entries for eth0 and one for eth1. Now that file has 3 entries for eth0. Under 9.2 that was all there was in /proc/net/ but now there is an additional directory called /proc/net/dev_snmp6/ which contains a file called eth0 and another called lo, each of which has a bunch of lines with Icmp6... and a number on each line. Except for the numbers they look identical. Obviously there have been some changes in the way devices are assigned, at least for network devices. This seems to be "for experts only" which usually means it should work without any user savvy. (Like it used to.) But it doesn't. Either some guru takes an interest in this and bails me out or I'll have to switch to some other distro; if so, I won't be back. I haven't got time to be debugging _released_ system tools! Quote Link to comment Share on other sites More sharing options...
streeter Posted June 30, 2004 Report Share Posted June 30, 2004 I'm no guru, but please post output of lsmod , lspci , ifconfig and contents of /etc/modprobe.conf. Also any mention of the NICs in the modprobe.* files. Chris Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted July 6, 2004 Author Report Share Posted July 6, 2004 lsmod: . . . tulip 45248 0 3c59x 39144 0 . . . lspci: 00:00.0 Host bridge: Intel Corp. 82845 845 (Brookdale) Chipset Host Bridge (rev 03) 00:01.0 PCI bridge: Intel Corp. 82845 845 (Brookdale) Chipset AGP Bridge (rev 03) 00:1e.0 PCI bridge: Intel Corp. 82801BA/CA/DB/EB PCI Bridge (rev 12) 00:1f.0 ISA bridge: Intel Corp. 82801BA ISA Bridge (LPC) (rev 12) 00:1f.1 IDE interface: Intel Corp. 82801BA IDE U100 (rev 12) 00:1f.2 USB Controller: Intel Corp. 82801BA/BAM USB (Hub #1) (rev 12) 00:1f.3 SMBus: Intel Corp. 82801BA/BAM SMBus (rev 12) 00:1f.4 USB Controller: Intel Corp. 82801BA/BAM USB (Hub #2) (rev 12) 01:00.0 VGA compatible controller: nVidia Corporation NV11 [GeForce2 MX/MX 400] (rev b2) 02:0b.0 SCSI storage controller: Adaptec AHA-7850 (rev 03) 02:0c.0 Ethernet controller: Digital Equipment Corporation DECchip 21041 [Tulip Pass 3] (rev 21) 02:0d.0 Multimedia audio controller: Creative Labs SB Live! EMU10k1 (rev 07) 02:0d.1 Input device controller: Creative Labs SB Live! MIDI/Game Port (rev 07) 02:0e.0 Ethernet controller: 3Com Corporation 3c905C-TX/TX-M [Tornado] (rev 78) ifconfig: eth0 Link encap:Ethernet HWaddr 00:04:75:71:42:78 inet addr:24.87.89.54 Bcast:255.255.255.255 Mask:255.255.252.0 inet6 addr: fe80::204:75ff:fe71:4278/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:213047 errors:0 dropped:0 overruns:0 frame:0 TX packets:210338 errors:0 dropped:0 overruns:0 carrier:0 collisions:260 txqueuelen:1000 RX bytes:66985845 (63.8 Mb) TX bytes:182820644 (174.3 Mb) Interrupt:10 Base address:0xb400 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:5616756 errors:0 dropped:0 overruns:0 frame:0 TX packets:5616756 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:728685126 (694.9 Mb) TX bytes:728685126 (694.9 Mb) modprobe.conf (edited - I added the line about eth1!) alias autofs autofs4 alias eth0 3c59x alias eth1 tulip alias sound-slot-0 snd-emu10k1 install scsi_hostadapter /sbin/modprobe aic7xxx; /bin/true install snd-emu10k1 /sbin/modprobe --first-time --ignore-install snd-emu10k1 && { /sbin/modprobe snd-pcm-oss; /bin/true; } install usb-interface /sbin/modprobe usb-uhci; /bin/true remove snd-emu10k1 { /sbin/modprobe -r snd-pcm-oss; } ; /sbin/modprobe -r --first-time --ignore-remove snd-emu10k1 last few lines of dmesg: . . . EXT3 FS on hda8, internal journal EXT3-fs: mounted filesystem with ordered data mode. PCI: Found IRQ 10 for device 0000:02:0e.0 3c59x: Donald Becker and others. www.scyld.com/network/vortex.html 0000:02:0e.0: 3Com PCI 3c905C Tornado at 0xb400. Vers LK1.1.19 Linux Tulip driver version 1.1.13 (May 11, 2002) It ssems to know the Tulip is there but does not get an address or IRQ for it. ??? Quote Link to comment Share on other sites More sharing options...
streeter Posted July 6, 2004 Report Share Posted July 6, 2004 Pity you reinstalled - the answer may have been in your old /etc/modules file - I wonder if you need to specify the IRQ/something else when the module is loaded. Might help to have a look here:http://www.scyld.com/tulip.html Only other thoughts I have at the moment are - is the tulip 'card' an onboard one - if so, check your BIOS. If both cards are 'real' PCI cards, try swapping the slots - sometimes works wonders. Have you tried booting say, knoppix ? Does it work here (you can often get useful info from knoppix output) Or get a new (cheap) NIC - got to be easier than installing another distro. :) Chris Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted July 7, 2004 Author Report Share Posted July 7, 2004 Well, this is interesting. I ran harddrake2 (kudzu no longer exists, I guess) and checked the detected Ethernat boards. It sees both boards, but for the Tulip it lists Module: de2104x rather than tulip. So I edited my /etc/modprobe.conf file and where it said alias eth1 tulip I changed it to alias eth1 de2104x and then did modprobe de2104x after which ifup eth1 worked! Now ifconfig shows eth0 Link encap:Ethernet HWaddr 00:04:75:71:42:78 inet addr:24.87.89.54 Bcast:255.255.255.255 Mask:255.255.252.0 inet6 addr: fe80::204:75ff:fe71:4278/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:7363 errors:0 dropped:0 overruns:0 frame:0 TX packets:5467 errors:0 dropped:0 overruns:0 carrier:0 collisions:6 txqueuelen:1000 RX bytes:5326622 (5.0 Mb) TX bytes:1128543 (1.0 Mb) Interrupt:10 Base address:0xb400 eth1 Link encap:Ethernet HWaddr 00:E0:29:10:4A:84 inet addr:192.168.0.1 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::2e0:29ff:fe10:4a84/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:6 errors:0 dropped:0 overruns:0 frame:0 TX packets:13 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:1787 (1.7 Kb) TX bytes:714 (714.0 B) Interrupt:9 lo Link encap:Local Loopback inet addr:127.0.0.1 Mask:255.0.0.0 inet6 addr: ::1/128 Scope:Host UP LOOPBACK RUNNING MTU:16436 Metric:1 RX packets:423 errors:0 dropped:0 overruns:0 frame:0 TX packets:423 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:0 RX bytes:49675 (48.5 Kb) TX bytes:49675 (48.5 Kb) and I am now setting out to get the LAN back up. BTW, I backed up _everything_ from 9.2 before I started, so I was able to check the old version... and there WAS NO /etc/modprobe.conf under 9.2 as far as I can tell. Huh?! Wish me luck. It may only _appear_ to be fixed. Quote Link to comment Share on other sites More sharing options...
streeter Posted July 7, 2004 Report Share Posted July 7, 2004 AHA!! That would explain a lot :) I should have looked the card up... >>.. and there WAS NO /etc/modprobe.conf For the record, the equivalent files were called modules and modules.conf Chris Quote Link to comment Share on other sites More sharing options...
Gowator Posted July 7, 2004 Report Share Posted July 7, 2004 im pasting your solution onto the sticky :D this should help everyone with 'tulip' drivers for the Tulip it lists Module: de2104x rather than tulip. So I edited my /etc/modprobe.conf file and where it said alias eth1 tulip I changed it to alias eth1 de2104x and then did modprobe de2104x after which ifup eth1 worked! Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted July 8, 2004 Author Report Share Posted July 8, 2004 Yes, well... Under Mandrake 9.2 the tulip driver was working fine, no mention of any de2104x -- tulip is specified in both modules.conf and modules. Now it insists on using de2104x which _appears_ to be accepted when I ifup eth1 and ifconfig reports eth1 running, HOWEVER... On boot, it still says FAILED when it tries to start up eth1, and in fact eth1 is NOT working despite its appearance in ifconfig. I'd assume that harddrake simply has it wrong, except that I have talked to people who have only a single Tulip board and upgraded to Mandrake 10 from 9.2 without any problem. So at this stage it _appears_ to be a problem specific to the use of two different Ethernet boards at the same time, one of which is a Tulip. This is probably well beyond my ability to fix, so the "other distro" solution is looking pretty attractive. I have a love/hate experience with Mandrake generally (I love the nice GUIs for system configuration, but by the same token I hate it when they don't work; my learning-curve time would have been better spent becoming more proficient with the arcane command-line tools of Linux)-- but, as in the Boy Scouts, "there's never enough time to do it right in the first place, but there's always time to do it over!" Maybe I'll go all the way over to Free BSD. :unsure: :sad: Quote Link to comment Share on other sites More sharing options...
Gowator Posted July 8, 2004 Report Share Posted July 8, 2004 LOL FreeBSD is really scary !!! although I here dragonfly is getting more user friendly for installing... I have an old tulip card somewhere I remember the driver... I think you sum up with wizards quite well. See the sticky... they can work well but when they fail they tend to make a real mess. Once place they tend to make a real mess is when you have two NICs they presume you have one. I honestly think your best bet is install webmin and delete eth1 completely then recreate it.... you can do this by hand with the arcane command to but it doesnt sound like you fancy that right now...and I cant say I blame you. If its getting an IP from DHCP it seems to be working if its from you setting it with ifconfig then I guess it doesnt prove anything. Quote Link to comment Share on other sites More sharing options...
JHBrewer Posted July 16, 2004 Author Report Share Posted July 16, 2004 It's working now. I had forgotten to chmod +x /etc/sysconfig/network-scripts/ifcfg-eth1 after I had manually created that file (!) by copying the ifcfg-eth0 file created by the installation and editing it to match the hardware driver for the other card. Sheer luck I stumbled across this oversight and corrected it. This changes nothing in disappointment with Mandrake 10. The installer apparently assumes that one Ethernet board is enough for anyone, and diaables the second one (or never ENables it), whereas under 8 and 9 it assumed correctly that if you had two boards it meant you were planning to masquerade a LAN. I can unly conclude that Mandrake is following the same "dumb it down for Windoze users" policy that is ruining Red Hat, alas. I would actually be reassured to hear that this was just a screwup, rather than a new "feature". Grrrr.... Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.