I have a testbed machine on which I've installed various OSs at various times, including FC6 from the same set of CDs repeatedly. The last couple of installs have refused, adamantly, to connect to the Net -- so adamantly that I had begun to suspect hardware trouble, in the form of some sudden failure of a cable or my ethernet card, but couldn't find it.
So I booted up a live CD of Puppy Linux. Once I figured out that I needed to ignore all its dialup stuff and just start with the setup icon, it connected right up. No sweat at all.
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
I've started from the icon, new to me, that looks like a yellow star on a blue balloon, with what might be a mouse at its lower left. Properties for that give a command "/usr/bin/system-config-network"
I can get to a box that lets me tell it to obtain IP address settings automatically with dhcp, told it yes, and tried with and without the address of my router. No Joy. The MAC address it has is the one it has always had; but probing claims that doesn't exist.
I've tried going to the hosts tab and the edit button, putting in or taking out the IP of the router. No joy.
What do I have to do to this thing?? I've never had to do any of this with any of the other machines I've installed FC6 on with these same CDs -- nor 5, or 4, or 3, ... -- nor the first time with this same machine.
Beartooth wrote:
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
Does the install actually see the physical nic? Could it possibly be loading the wrong driver for it? Before you go into software hunting mode, I always check hardware drivers first, make sure they're there and are correct. I'd check /var/log/dmesg see what it says regarding the card.
And not to burst your bubble, but I just installed FC6 on a relatively old machine that has 3 Intel nics in it (one dual and one single.) Just a breeze...configured all three, rebooted and there they were. I'm running updates right now. By golly, I'll be going home early today.
On Mon, 29 Jan 2007 13:59:47 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
Does the install actually see the physical nic? Could it possiblybe loading the wrong driver for it? Before you go into software hunting mode, I always check hardware drivers first, make sure they're there and are correct. I'd check /var/log/dmesg see what it says regarding the card.
Dunno how to tell most of that. I did "cat /var/log/dmesg | more" and looked through it, but 99 44/100% is over my head. I also tried "cat /var/log/dmesg|grep eth0" and got nothing.
What should I be looking for in that log??
And not to burst your bubble, but I just installed FC6 on arelatively old machine that has 3 Intel nics in it (one dual and one single.) Just a breeze...configured all three, rebooted and there they were. I'm running updates right now. By golly, I'll be going home early today.
I'm not even sure what a nic is; but I've owned this machine since it came out the door of the retailer (in 11/98), and nothing has been done about any ethernet card(s) in years. It has the same old 10/100 it's had since who goosed the moose, and nothing at all has been done to any of its hardware in many moons. I know it should work, and did work -- and has always figured itself out on boot.
Beartooth wrote:
I also tried "cat /var/log/dmesg|grep eth0" and got nothing.
Try grepping for 'eth' in /var/log/messages. See if that produces anything. I have some machines whose drivers get logged there as opposed to dmesg.
What should I be looking for in that log??
Something akin to:
Jan 29 06:59:35 trinity kernel: e100: eth0: e100_probe: addr 0xf4020000, irq 3, MAC addr 00:02:B3:BE:02:87
I'm not even sure what a nic is; but I've owned this machine since it came out the door of the retailer (in 11/98), and nothing has been done about any ethernet card(s) in years. It has the same old 10/100 it's had since who goosed the moose, and nothing at all has been done to any of its hardware in many moons. I know it should work, and did work -- and has always figured itself out on boot.
The fact that it's so old might not be such a bad thing. Sometimes because a cadr has been around for so long, there's a good chance its driver has too. On the other hand, it's also possible that the driver has been ripped out of the current source code and you'd have to do things manually to get it working again.
Like I said, let's start by checking to make sure the system knows the hardware is actually present. Grep /var/log/messages and let me know.
On Mon, 2007-01-29 at 22:01 +0000, Beartooth wrote:
I'm not even sure what a nic is
Network Interface Card. Though, these days, with ethernet interfaces being built onto the motherboard, rather than being a separate plug-in card, the term is a bit of a misnomer.
Beartooth wrote:
I have a testbed machine on which I've installed various OSs at various times, including FC6 from the same set of CDs repeatedly. The last couple of installs have refused, adamantly, to connect to the Net -- so adamantly that I had begun to suspect hardware trouble, in the form of some sudden failure of a cable or my ethernet card, but couldn't find it.
So I booted up a live CD of Puppy Linux. Once I figured out that I needed to ignore all its dialup stuff and just start with the setup icon, it connected right up. No sweat at all.
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
I've started from the icon, new to me, that looks like a yellow star on a blue balloon, with what might be a mouse at its lower left. Properties for that give a command "/usr/bin/system-config-network"
I can get to a box that lets me tell it to obtain IP address settings automatically with dhcp, told it yes, and tried with and without the address of my router. No Joy. The MAC address it has is the one it has always had; but probing claims that doesn't exist.
I've tried going to the hosts tab and the edit button, putting in or taking out the IP of the router. No joy.
What do I have to do to this thing?? I've never had to do any of this with any of the other machines I've installed FC6 on with these same CDs -- nor 5, or 4, or 3, ... -- nor the first time with this same machine.
The most obvious question would be what kind of network card is it and are the modules loading for it?
Now THAT would be something you have to include in your post when you ask a question like this.
Regards, Ed.
On Tue, 30 Jan 2007 12:13:10 +0900, Edward Dekkers wrote:
The most obvious question would be what kind of network card is it and are the modules loading for it?
I haven't the faintest idea. I don't speak hardware -- much less try to handle it with trifocal fingers and arthritic eyeballs. I've never so much as gotten a case off a computer without breaking something trying to get it back on.
Now THAT would be something you have to include in your post when you ask a question like this.
OK, fine -- how do I find out? The hardware browser I see in FC6, or what looks to me to be one, is very different from what used to be called that in FC 1 - 5; but I do have a way to open something that I think will tell me -- once I know where and how to look. So how do I know a network card when I see one? I bet it isn't called that.
And what command a/o log will tell me whether modules (modules??) are loading? I'll be glad to do the work and report, once I know how.
Beartooth wrote:
OK, fine -- how do I find out?
Have you tried my previous suggestion? Grep for eth in /var/log/messages (instead of dmesg):
grep eth /var/log/messages
What's the result?
As for modules, if any, you can type 'lsmod' at the command line to see what's (currently) loaded. I'd also look at the contents of /etc/modprobe.conf to see if it's doing anything with eth0 at all (assuming it sees it): cat /etc/modprobe.conf
On Mon, 29 Jan 2007 20:53:20 +0000, Beartooth wrote: [...]
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
I've started from the icon, new to me, that looks like a yellow star on a blue balloon, with what might be a mouse at its lower left. Properties for that give a command "/usr/bin/system-config-network"
I can get to a box that lets me tell it to obtain IP address settings automatically with dhcp, told it yes, and tried with and without the address of my router. No Joy. The MAC address it has is the one it has always had; but probing claims that doesn't exist.
I seem to have forgotten to mention that at one point eth0 on some screen was set to use only a static IP -- and the whole other half of that screen, on which I'd've tried to tell it to get an IP from dhcp on the router, was greyed out.
I later found a different screen that did seem to let me tell it to use the dynamic IP from the router -- and I've tried restarting the network, logging out and back in, and even rebooting right after that -- with no success. So I'm not sure that the command to go get an IP is actually registering.
Any time I first reboot and/or log in, doing /sbin/ifconfig gets me only what the network connection icons do : lo only, with no indication of any eth0 nor way to add one.
I'd've thought I could become root and do "nano -w /sbin/ifconfig" -- if I knew what to add -- but what I get is four hundred-odd lines of gibberish, with a note at the bottom saying it's converted from Mac format, whatever that is.
la /etc|grep .conf gets me umpteen bazillion config files, and no doubt grepping .cfg and the like will get me bazillions more. Is there one I can open that's a text file, and figure out what to add into it??
Beartooth wrote:
I'd've thought I could become root and do "nano -w /sbin/ifconfig" -- if I knew what to add -- but what I get is four hundred-odd lines of gibberish, with a note at the bottom saying it's converted from Mac format, whatever that is.
That's the wrong approach (and wrong file to be editing.) All network interface (cards) configurations are in /etc/sysconfig/network-scripts/ and gateway information is set in /etc/sysconfig/network - leave ifconfig alone. :) Try my previous suggestion please.
On Mon, 29 Jan 2007 15:45:46 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
I also tried "cat /var/log/dmesg|grep eth0" and got nothing.
Try grepping for 'eth' in /var/log/messages. See if that producesanything. I have some machines whose drivers get logged there as opposed to dmesg.
OK, that got me two screensful
What should I be looking for in that log??
Something akin to:Jan 29 06:59:35 trinity kernel: e100: eth0: e100_probe: addr 0xf4020000, irq 3, MAC addr 00:02:B3:BE:02:87
I see several lines containing things like "eth0 alias e100" or "eth1 alias 3c501" -- and also a whole bunch containing expresions like "//etc/sysconfig/networking/devices/ifcfg-eth0" or the same with "-eth1" -- but none with addr nor MAC addr. [...]
The fact that it's so old might not be such a bad thing. Sometimesbecause a cadr has been around for so long, there's a good chance its driver has too. On the other hand, it's also possible that the driver has been ripped out of the current source code and you'd have to do things manually to get it working again.
Like I said, let's start by checking to make sure the system knowsthe hardware is actually present. Grep /var/log/messages and let me know.
Beartooth wrote:
I see several lines containing things like "eth0 alias e100" or "eth1 alias 3c501" -- and also a whole bunch containing expresions like "//etc/sysconfig/networking/devices/ifcfg-eth0" or the same with "-eth1" -- but none with addr nor MAC addr.
Sounds like you may have more then one nic, or it THINKS you have more than one nic. I realize you got a bunch of lines, but could yuo go ahead and capture that and send it to me (OFF-LIST)?
grep eth /var/log/messages > eth-messages.txt
Then send that eth-messages.txt to me. Hopefully that might shed some light. I have a meeting coming up shortly but I'll be around afterwards and later this evening as well.
On Tue, 30 Jan 2007 12:37:08 -0700, Ashley M. Kirchner wrote:
That's the wrong approach (and wrong file to be editing.) Allnetwork interface (cards) configurations are in /etc/sysconfig/network-scripts/ and gateway information is set in /etc/sysconfig/network - leave ifconfig alone. :) Try my previous suggestion please.
OK, I just posted a reply to that.
Btw, I just looked (with the GUI, which seems to use gedit) at /etc/sysconfig/network-scripts. There *are* a bunch in there with names like ifcfg-eth0, and the one I'm looking at resembles things I've had to be walked through editing in FC3 or 4 -- with entries like DEVICE=, ONBOOT=yes, and so on. And the HWADDR= looks to be what it has always been.
Beartooth wrote:
Btw, I just looked (with the GUI, which seems to use gedit) at /etc/sysconfig/network-scripts. There *are* a bunch in there with names like ifcfg-eth0, and the one I'm looking at resembles things I've had to be walked through editing in FC3 or 4 -- with entries like DEVICE=, ONBOOT=yes, and so on. And the HWADDR= looks to be what it has always been.
Theoretically you should have two ifcfg- files, one for eth0 and one for lo:
ls -l /etc/sysconfig/network-scripts/ifcfg* /etc/sysconfig/network-scripts/ifcfg-eth0 /etc/sysconfig/network-scripts/ifcfg-lo
This is assuming your machine only has ONE nic in it. If it has more then one, you'd see an ifcfg-eth1 as well, basically one ifcfg- file for each interface.
Meeting...
On Tue, 30 Jan 2007 12:48:11 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
I see several lines containing things like "eth0 alias e100" or "eth1 alias 3c501" -- and also a whole bunch containing expresions like "//etc/sysconfig/networking/devices/ifcfg-eth0" or the same with "-eth1" -- but none with addr nor MAC addr.
Sounds like you may have more then one nic, or it THINKS you havemore than one nic. I realize you got a bunch of lines, but could yuo go ahead and capture that and send it to me (OFF-LIST)?
Hmmm -- That machine doesn't even see the router, afaict. So I can't use ssh nor scp, much less send myself an email with that machine and read/forward it with this one -- let alone use my newsreader with it.
I can try to hand-copy, reading one monitor and typing into another. (The problem machine has its own peripherals; it's not behind the KVM switch with my others.) But it's apt to be less than accurate. I've been trying to put the eth-messages.txt onto a floppy or a cd, and read that with this machine; but I can't seem to follow my antiquated nose to a place hwere I can do either.
grep eth /var/log/messages > eth-messages.txt Then send that eth-messages.txt to me. Hopefully that might shedsome light. I have a meeting coming up shortly but I'll be around afterwards and later this evening as well.
On Tue, 30 Jan 2007 12:27:09 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
OK, fine -- how do I find out?
Have you tried my previous suggestion? Grep for eth in/var/log/messages (instead of dmesg):
grep eth /var/log/messages What's the result?
I get about 2 1/2 or 3 screens of stuff. I also did grep eth /var/log/messages > eth-messages.txt, which worked. What's more, so did cp eth-messages.txt /home/btth and then (as root in /home/btth) chown -v btth eth-messages.txt, which told me it had changed ownership to btth.
But I can't seem to get K3B to see that my CD drive (an HP CD-Writer) on that machine can in fact burn CDs; nor can I dig up from ancient memory how to copy it onto a floppy.
Therefore, since the old P2 also can't see the router (I checked the cables!), and therefore isn't likely to be able to do scp, I can't seem to get the file into anywhere I can email or post it from. Catch-22 ...
As for modules, if any, you can type 'lsmod' at the command line tosee what's (currently) loaded. I'd also look at the contents of /etc/modprobe.conf to see if it's doing anything with eth0 at all (assuming it sees it): cat /etc/modprobe.conf
lsmod from a root prompt gets me "bash: lsmod: command not found"
cat /etc/modprobe.conf from the same root prompt gets :
alias eth0 e100 alias eth1 3c501
On Tue, 30 Jan 2007 12:57:38 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
Btw, I just looked (with the GUI, which seems to use gedit) at /etc/sysconfig/network-scripts. There *are* a bunch in there with names like ifcfg-eth0, and the one I'm looking at resembles things I've had to be walked through editing in FC3 or 4 -- with entries like DEVICE=, ONBOOT=yes, and so on. And the HWADDR= looks to be what it has always been.
Theoretically you should have two ifcfg- files, one for eth0 and onefor lo:
ls -l /etc/sysconfig/network-scripts/ifcfg*
That gets me four lines, one each for ifcfg-eth0, ifcfg-eth1, ifcfg-lo, and ifcfg-netgear (netgear being my nickname for the interface between me and my wireless broadband provider, which is a Netgear MBR814.
/etc/sysconfig/network-scripts/ifcfg-eth0 /etc/sysconfig/network-scripts/ifcfg-lo
Btw, I'm pretty sure that at least eth1 and netgear are things I've added through the GUI this afternoon. Even eth0 may be one. But the GUI app that shows eth0 and eth1 won[t let me activate either; and the other GUI shows only lo, with no way to add anything.
This is assuming your machine only has ONE nic in it. If it hasmore then one, you'd see an ifcfg-eth1 as well, basically one ifcfg- file for each interface.
Beartooth wrote:
On Mon, 29 Jan 2007 20:53:20 +0000, Beartooth wrote: [...]
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
I've started from the icon, new to me, that looks like a yellow star on a blue balloon, with what might be a mouse at its lower left. Properties for that give a command "/usr/bin/system-config-network"
I can get to a box that lets me tell it to obtain IP address settings automatically with dhcp, told it yes, and tried with and without the address of my router. No Joy. The MAC address it has is the one it has always had; but probing claims that doesn't exist.
I seem to have forgotten to mention that at one point eth0 on some screen was set to use only a static IP -- and the whole other half of that screen, on which I'd've tried to tell it to get an IP from dhcp on the router, was greyed out.
I later found a different screen that did seem to let me tell it to use the dynamic IP from the router -- and I've tried restarting the network, logging out and back in, and even rebooting right after that -- with no success. So I'm not sure that the command to go get an IP is actually registering.
Any time I first reboot and/or log in, doing /sbin/ifconfig gets me only what the network connection icons do : lo only, with no indication of any eth0 nor way to add one.
I'd've thought I could become root and do "nano -w /sbin/ifconfig" -- if I knew what to add -- but what I get is four hundred-odd lines of gibberish, with a note at the bottom saying it's converted from Mac format, whatever that is.
la /etc|grep .conf gets me umpteen bazillion config files, and no doubt grepping .cfg and the like will get me bazillions more. Is there one I can open that's a text file, and figure out what to add into it??
You have to enable it to boot at startup or enable it through the interface. I believe either typing neat or system-config-network or getting to it through the system/administration/network menu would allow you to adjust the settings.
It should also tell you under the hardware tab what type of network card it thinks you have on the computer. In the DNS tab should be your hostname, primary and secondary DNS and the search path for the DNS.
My logic is that it must have configured some type of card since you said that you could choose static or DHCP.
If you highlight the device and choose edit, you should see a checkmark next to the "Activate device when computer starts"
For the other choice, it should be set to "automatically obtain IP address settings with dhcp"
For finding out the type of NIC is using from the command terninal, lspci should show all the info for your PCI devices.
To find out the driver loaded for the card, lsmod should show you which module driver was loaded for the card.
Jim
Beartooth wrote:
cat /etc/modprobe.conf from the same root prompt gets :
alias eth0 e100 alias eth1 3c501
This tells me the machine thinks it has 2 network cards in it. Is that not the case?
On Tue, 2007-01-30 at 19:33 +0000, Beartooth wrote:
Any time I first reboot and/or log in, doing /sbin/ifconfig gets me only what the network connection icons do : lo only, with no indication of any eth0 nor way to add one.
I've had that on one box where ACPI and Linux didn't get along. I had to tweak BIOS options.
I'm wondering whether you'd better off to run the "neat" tool (a GUI tool for network configuration), remove all the current hardware devices configured, and try adding them again.
Beartooth wrote:
On Tue, 30 Jan 2007 12:13:10 +0900, Edward Dekkers wrote:
The most obvious question would be what kind of network card is it and are the modules loading for it?
I haven't the faintest idea. I don't speak hardware -- much less try to handle it with trifocal fingers and arthritic eyeballs. I've never so much as gotten a case off a computer without breaking something trying to get it back on.
Now THAT would be something you have to include in your post when you ask a question like this.
OK, fine -- how do I find out? The hardware browser I see in FC6, or what looks to me to be one, is very different from what used to be called that in FC 1 - 5; but I do have a way to open something that I think will tell me -- once I know where and how to look. So how do I know a network card when I see one? I bet it isn't called that.
And what command a/o log will tell me whether modules (modules??) are loading? I'll be glad to do the work and report, once I know how.
Jim's already answered this one but in short:
lspci for listing PCI devices in your system and lsmod for listing loaded modules.
These need to be run from a CLI.
Regards, Ed.
P.S. I've been using Linux since Redhat 5.2 and I STILL can not do everything I want graphically. I'm afraid for fixing problems like this the CLI is still your best bet.
P.P.S. Oh, sorry - CLI = Command Line Interface. Basically just open a terminal window from your menu - that is the CLI.
Edward Dekkers wrote:
P.S. I've been using Linux since Redhat 5.2 and I STILL can not do everything I want graphically. I'm afraid for fixing problems like this the CLI is still your best bet.
Using RHL from the Realplayer, Netscape 4.7.x series and Wordperfect preview edition. I cannot recall if this was 5.1 or 5.2
Anyway, using setup from the cli was my early way of configuring the system. Now using the GUI tools more often, it always is a challenge when they change the GUI tools from one version to a completely different acting tool. Network configuration and X configuration come to mind.
Jim
On Tue, 30 Jan 2007 16:17:05 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
cat /etc/modprobe.conf from the same root prompt gets :
alias eth0 e100 alias eth1 3c501
This tells me the machine thinks it has 2 network cards in it. Isthat not the case?
News to me if it does. How do I check?
Beartooth wrote:
News to me if it does. How do I check?
This may seem rather stupid, but how many ethernet plugin holes do you see on the back of the machine?
On Wed, 2007-01-31 at 19:24 +0000, Beartooth wrote:
On Tue, 30 Jan 2007 16:17:05 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
cat /etc/modprobe.conf from the same root prompt gets :
alias eth0 e100 alias eth1 3c501
This tells me the machine thinks it has 2 network cards in it. Isthat not the case?
News to me if it does. How do I check?
Just run "lspci | grep -i net" and it'll show you what's in the PCI bus (and motherboard), e.g. (from my laptop):
[root@golem2 ~]# lspci | grep -i net 03:02.0 Network controller: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller (rev 02) 03:06.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - If at first you don't succeed, quit. No sense being a damned fool! - ----------------------------------------------------------------------
On Tue, 30 Jan 2007 21:12:42 +0000, Anne Wilson wrote:
On Tuesday 30 January 2007 20:52, Beartooth wrote:
lsmod from a root prompt gets me "bash: lsmod: command not found"
^^^^^^^^^^^^^^^^^
That's typical of what you get when you try it as user. Have you tried it as root?
Actually, I keep two tabs on my terminal (with background color contrasting to the default!) permanently signed on to root, one at / and one at /home/btth for things I do in that directory.
I didn't c&p from there only because the problem machine, not seeing even the router, gives me no way I've yet found to use c&p to another machine that does access the Net. (Its K3B doesn't recognize my CD drive as a burner, and I haven't actually used a floppy for so long I don't remember how.)
But I did mean the phrase "from a root prompt" to imply that that machine shows me exactly what, as a matter of fact, I also get on the biggest newest one, which I'm using now :
[root@localhost btth]# lspci bash: lspci: command not found [root@localhost btth]# lsmod bash: lsmod: command not found [root@localhost btth]#
On Wed, 31 Jan 2007 11:15:23 +0900, Edward Dekkers wrote:
Beartooth wrote:
On Tue, 30 Jan 2007 12:13:10 +0900, Edward Dekkers wrote:
The most obvious question would be what kind of network card is it and are the modules loading for it?
[...]
OK, fine -- how do I find out? The hardware browser I see in FC6, or what looks to me to be one, is very different from what used to be called that in FC 1 - 5; but I do have a way to open something that I think will tell me -- once I know where and how to look. So how do I know a network card when I see one? I bet it isn't called that.
And what command a/o log will tell me whether modules (modules??) are loading? I'll be glad to do the work and report, once I know how.
Jim's already answered this one but in short:
lspci for listing PCI devices in your system and lsmod for listing loaded modules.
I get the same result on the non-connecting machine as on this one :
[root@localhost btth]# lspci bash: lspci: command not found [root@localhost btth]# lsmod bash: lsmod: command not found [root@localhost btth]#
These need to be run from a CLI.
Actually, I got into linux in the first place (with RH 7.2) in order to be able to run Pine, which I had had at work over a telnet link to an AIX machine in the IT division -- and what I liked and still do about Pine is that it is command-line driven; but it's still the one app that my fingers know, without requiring conscious direction.
What's left of my memory seems to have woefully buttery linux fingers; there are a few things I now do with the CLI for preference -- namely of course the ones I do most often. But when in doubt, I always try the GUI first, because my visual memory works better -- and me a linguist <sigh...>
P.S. I've been using Linux since Redhat 5.2 and I STILL can not do everything I want graphically. I'm afraid for fixing problems like this the CLI is still your best bet.
Last time I had any of the sort, with FC2 or 3 iirc, I did in fact work entirely by editing .conf files -- with a helpful soul walking me through -- but I disremember which .conf files they were ...<whimper, sniff ...>
SO : it's possible that someone put a second card in there for me at some time -- but I don't recall it. Therefore, given that lspci and lsmod aren't getting anything, how do I check??
On 31/01/07, Beartooth Beartooth@swva.net wrote:
On Tue, 30 Jan 2007 21:12:42 +0000, Anne Wilson wrote:
On Tuesday 30 January 2007 20:52, Beartooth wrote:
lsmod from a root prompt gets me "bash: lsmod: command not found"
^^^^^^^^^^^^^^^^^That's typical of what you get when you try it as user. Have you tried it as root?
Actually, I keep two tabs on my terminal (with background color contrasting to the default!) permanently signed on to root, one at / and one at /home/btth for things I do in that directory.
I didn't c&p from there only because the problem machine, not seeing even the router, gives me no way I've yet found to use c&p to another machine that does access the Net. (Its K3B doesn't recognize my CD drive as a burner, and I haven't actually used a floppy for so long I don't remember how.)
But I did mean the phrase "from a root prompt" to imply that that machine shows me exactly what, as a matter of fact, I also get on the biggest newest one, which I'm using now :
[root@localhost btth]# lspci bash: lspci: command not found [root@localhost btth]# lsmod bash: lsmod: command not found [root@localhost btth]#
$ su -
Use the "-" when you switch to root.
Beartooth escribió:
On Wed, 31 Jan 2007 11:15:23 +0900, Edward Dekkers wrote:
Jim's already answered this one but in short:
lspci for listing PCI devices in your system and lsmod for listing loaded modules.
I get the same result on the non-connecting machine as on this one :
[root@localhost btth]# lspci bash: lspci: command not found
Install pciutils.
[root@localhost btth]# lsmod bash: lsmod: command not found
This should be in module-init-tools. Are you sure you have the path configured correctly? Try with /sbin/lspci.
On Tue, 30 Jan 2007 17:48:15 -0500, Jim Cornette wrote:
You have to enable it to boot at startup or enable it through the interface. I believe either typing neat or system-config-network or getting to it through the system/administration/network menu would allow you to adjust the settings.
OK, typing "neat" and <enter> from a root prompt gets the GUI I've been mostly using. Clicking the activate mark, with either eth0 or eth1 highlighted, gets me an error message saying the device seems not to be present.
It should also tell you under the hardware tab what type of network card it thinks you have on the computer.
On that tab, it calls eth0 an Intel EthernetExpress/100 drive -- and eth1 a 3Com 3c501 -- and says both are configured. The IPSec tab, btw, shows a check mark under Profile, Host2Host under Type, 192.168.x.y (the correct number, under which I access the router from any browser on any other machine) under Destination, and "netgear" -- the nickname I gave the router yesterday -- under Nickname.
In the DNS tab should be your hostname, primary and secondary DNS and the search path for the DNS.
Everything there was blank. I simply copied verbatim from what I see on a connecting machine.
It still wouldn't simply activate either eth, but did tell me (as it has so often) that I might want to restart network services or the machine. The command "services network restart" gets me only an error saying it's not recognized. I rebooted. /sbin/ifconfig after the reboot still shows only the loopback running, with no mention of ethX.
My logic is that it must have configured some type of card since you said that you could choose static or DHCP.
If you highlight the device and choose edit, you should see a checkmark next to the "Activate device when computer starts"
Yup. I do.
For the other choice, it should be set to "automatically obtain IP address settings with dhcp"
Yup. It is.
For finding out the type of NIC is using from the command terninal, lspci should show all the info for your PCI devices.
lspci as root still shows me only "bash: lspci: command not found"
To find out the driver loaded for the card, lsmod should show you which module driver was loaded for the card.
again, not found. The first character in each of those *is* lower case L, isn't it, not the number 1??
On Wed, 31 Jan 2007 12:18:14 +1030, Tim wrote:
On Tue, 2007-01-30 at 19:33 +0000, Beartooth wrote:
Any time I first reboot and/or log in, doing /sbin/ifconfig gets me only what the network connection icons do : lo only, with no indication of any eth0 nor way to add one.
I've had that on one box where ACPI and Linux didn't get along. I had to tweak BIOS options.
I'm wondering whether you'd better off to run the "neat" tool (a GUI tool for network configuration), remove all the current hardware devices configured, and try adding them again.
Ouch! Well, I will if I have to; I guess I can copy them back from the post I just wrote, following up to Jim Cornette in this subthread.
Would it be any easier to try tweaking the BIOS first? I might could get it to see my CD-Writer that way, too ...
On Wed, 31 Jan 2007 12:37:11 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
News to me if it does. How do I check?
This may seem rather stupid, but how many ethernet plugin holes doyou see on the back of the machine?
On the contrary! But the one with the ethernet cable already in it is the only one; I tried the cable's other end against the two similar, and it was too big; they must be phone line jacks.
Otoh, that "3c501" is familiar from somewhere, and I'm pretty sure I haven't typed it in.
On Wed, 2007-01-31 at 20:27 +0000, Beartooth wrote:
On Tue, 30 Jan 2007 17:48:15 -0500, Jim Cornette wrote:
You have to enable it to boot at startup or enable it through the interface. I believe either typing neat or system-config-network or getting to it through the system/administration/network menu would allow you to adjust the settings.
OK, typing "neat" and <enter> from a root prompt gets the GUI I've been mostly using. Clicking the activate mark, with either eth0 or eth1 highlighted, gets me an error message saying the device seems not to be present.
It should also tell you under the hardware tab what type of network card it thinks you have on the computer.
On that tab, it calls eth0 an Intel EthernetExpress/100 drive -- and eth1 a 3Com 3c501 -- and says both are configured. The IPSec tab, btw, shows a check mark under Profile, Host2Host under Type, 192.168.x.y (the correct number, under which I access the router from any browser on any other machine) under Destination, and "netgear" -- the nickname I gave the router yesterday -- under Nickname.
In the DNS tab should be your hostname, primary and secondary DNS and the search path for the DNS.
Everything there was blank. I simply copied verbatim from what I see on a connecting machine.
It still wouldn't simply activate either eth, but did tell me (as it has so often) that I might want to restart network services or the machine. The command "services network restart" gets me only an error saying it's not recognized. I rebooted. /sbin/ifconfig after the reboot still shows only the loopback running, with no mention of ethX.
My logic is that it must have configured some type of card since you said that you could choose static or DHCP.
If you highlight the device and choose edit, you should see a checkmark next to the "Activate device when computer starts"
Yup. I do.
For the other choice, it should be set to "automatically obtain IP address settings with dhcp"
Yup. It is.
For finding out the type of NIC is using from the command terninal, lspci should show all the info for your PCI devices.
lspci as root still shows me only "bash: lspci: command not found"
To find out the driver loaded for the card, lsmod should show you which module driver was loaded for the card.
again, not found. The first character in each of those *is* lower case L, isn't it, not the number 1??
Yes, "lima-sierra-papa-charlie-india". Keep in mind that most of these commands are intended to be run as root with root's environment and path (e.g. a normal user does not have "/sbin" in his/her path).
If you (as a normal user) simply did an "su" to become root then you won't have root's environment. You must use "su -" ("su-space-dash") to get root's environment and path as well as becoming root. In any case, you should be able to run the command as "/sbin/lspci".
---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - Squawk! Pieces of Seven! Pieces of Seven! Parity Error! - ----------------------------------------------------------------------
On Wed, 31 Jan 2007 11:38:44 -0800, Rick Stevens wrote:
On Wed, 2007-01-31 at 19:24 +0000, Beartooth wrote:
On Tue, 30 Jan 2007 16:17:05 -0700, Ashley M. Kirchner wrote:
Beartooth wrote:
cat /etc/modprobe.conf from the same root prompt gets :
alias eth0 e100 alias eth1 3c501
This tells me the machine thinks it has 2 network cards in it. Isthat not the case?
News to me if it does. How do I check?
Just run "lspci | grep -i net" and it'll show you what's in the PCI bus (and motherboard), e.g. (from my laptop):
[root@golem2 ~]# lspci | grep -i net 03:02.0 Network controller: Broadcom Corporation BCM4318 [AirForce One 54g] 802.11g Wireless LAN Controller (rev 02) 03:06.0 Ethernet controller: Realtek Semiconductor Co., Ltd. RTL-8139/8139C/8139C+ (rev 10)
Well -- doubting, I tried it. But sure enough, I still get
bash: lspci: command not found
On Wed, 31 Jan 2007 17:12:46 -0300, Martin Marques wrote:
Beartooth escribió:
On Wed, 31 Jan 2007 11:15:23 +0900, Edward Dekkers wrote:
Jim's already answered this one but in short:
lspci for listing PCI devices in your system and lsmod for listing loaded modules.
I get the same result on the non-connecting machine as on this one :
[root@localhost btth]# lspci bash: lspci: command not found
Install pciutils.
Well, I can't exactly tell it "yum install pciutils" without a Net connection. Maybe there's a way to check the install media for it, but I have no idea how ..
[root@localhost btth]# lsmod bash: lsmod: command not found
This should be in module-init-tools. Are you sure you have the path configured correctly? Try with /sbin/lspci.
AHA! I get a whole bunch of stuff -- which I reall wish I could c&p -- about bridges and controllers. And, btw, /sbin/lsmod gets some fourteen lines of stuff, too.
On Wed, 2007-01-31 at 21:30 +0000, Beartooth wrote:
On Wed, 31 Jan 2007 17:12:46 -0300, Martin Marques wrote:
Beartooth escribió:
On Wed, 31 Jan 2007 11:15:23 +0900, Edward Dekkers wrote:
Jim's already answered this one but in short:
lspci for listing PCI devices in your system and lsmod for listing loaded modules.
I get the same result on the non-connecting machine as on this one :
[root@localhost btth]# lspci bash: lspci: command not found
Install pciutils.
Well, I can't exactly tell it "yum install pciutils" without a Net connection. Maybe there's a way to check the install media for it, but I have no idea how ..
[root@localhost btth]# lsmod bash: lsmod: command not found
This should be in module-init-tools. Are you sure you have the path configured correctly? Try with /sbin/lspci.
AHA! I get a whole bunch of stuff -- which I reall wish I could c&p -- about bridges and controllers. And, btw, /sbin/lsmod gets some fourteen lines of stuff, too.
Then try "/sbin/lspci | grep -i net" and it SHOULD display only the lines that contain "net" (in any combination of upper and lower case).
---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - Grabel's Law: 2 is not equal to 3--not even for large values of 2. - ----------------------------------------------------------------------
On Wed, 31 Jan 2007 13:22:04 -0800, Rick Stevens wrote:
[... snipperoo ...]
again, not found. The first character in each of those *is* lower case L, isn't it, not the number 1??
Yes, "lima-sierra-papa-charlie-india". Keep in mind that most of these commands are intended to be run as root with root's environment and path (e.g. a normal user does not have "/sbin" in his/her path).
If you (as a normal user) simply did an "su" to become root then you won't have root's environment. You must use "su -" ("su-space-dash") to get root's environment and path as well as becoming root. In any case, you should be able to run the command as "/sbin/lspci".
AHA! Hadn't ever known that. I exited and gave it the whole nine yards "su - root"
NOW lspci and lsmod work strait from the prompt, sure enough -- and lspci | grep -i net gives me (retyping) "00:09.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 10)" -- and no mention of any other ethernet anything, 3c501 or otherwise.
On Wednesday 31 January 2007 19:59, Beartooth wrote:
On Tue, 30 Jan 2007 21:12:42 +0000, Anne Wilson wrote:
On Tuesday 30 January 2007 20:52, Beartooth wrote:
lsmod from a root prompt gets me "bash: lsmod: command not found"
^^^^^^^^^^^^^^^^^That's typical of what you get when you try it as user. Have you tried it as root?
Actually, I keep two tabs on my terminal (with background color contrasting to the default!) permanently signed on to root, one at / and one at /home/btth for things I do in that directory.
I didn't c&p from there only because the problem machine, not seeing even the router, gives me no way I've yet found to use c&p to another machine that does access the Net. (Its K3B doesn't recognize my CD drive as a burner, and I haven't actually used a floppy for so long I don't remember how.)
But I did mean the phrase "from a root prompt" to imply that that machine shows me exactly what, as a matter of fact, I also get on the biggest newest one, which I'm using now :
[root@localhost btth]# lspci bash: lspci: command not found [root@localhost btth]# lsmod bash: lsmod: command not found [root@localhost btth]#
Sorry - somehow I missed that.
Anne
On Wed, 2007-01-31 at 22:37 +0000, Anne Wilson wrote:
On Wednesday 31 January 2007 19:59, Beartooth wrote:
On Tue, 30 Jan 2007 21:12:42 +0000, Anne Wilson wrote:
On Tuesday 30 January 2007 20:52, Beartooth wrote:
lsmod from a root prompt gets me "bash: lsmod: command not found"
^^^^^^^^^^^^^^^^^That's typical of what you get when you try it as user. Have you tried it as root?
Actually, I keep two tabs on my terminal (with background color contrasting to the default!) permanently signed on to root, one at / and one at /home/btth for things I do in that directory.
I didn't c&p from there only because the problem machine, not seeing even the router, gives me no way I've yet found to use c&p to another machine that does access the Net. (Its K3B doesn't recognize my CD drive as a burner, and I haven't actually used a floppy for so long I don't remember how.)
But I did mean the phrase "from a root prompt" to imply that that machine shows me exactly what, as a matter of fact, I also get on the biggest newest one, which I'm using now :
[root@localhost btth]# lspci bash: lspci: command not found [root@localhost btth]# lsmod bash: lsmod: command not found [root@localhost btth]#
Sorry - somehow I missed that.
---- as user...
echo $PATH /usr/kerberos/bin:/usr/java/jre/bin:/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/NX/bin
as user that performs 'su' /usr/kerberos/sbin:/usr/java/jre/bin:/usr/kerberos/bin:/usr/java/jre/bin:/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/NX/bin:/usr/NX/bin
as user that performs 'su -' # note the dash /usr/kerberos/sbin:/usr/kerberos/bin:/usr/java/jre/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/NX/bin:/root/bin
Note that each is different and it is an important difference...
# which lspci lsmod /sbin/lspci /sbin/lsmod
and /sbin path (and for that matter, /usr/sbin path) is only available to the user who performs the command 'su -'
Craig
Beartooth wrote:
su -
NOW lspci and lsmod work strait from the prompt, sure enough -- and lspci | grep -i net gives me (retyping) "00:09.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 10)" -- and no mention of any other ethernet anything, 3c501 or otherwise.
The 3c501 could be removed. You might want to investigate whether the Intel 82557/8/9 is supported and by what driver.
It sounds like you now have info that one more familiar with your hardware type can help solve.
Jim
On Wed, 2007-01-31 at 18:50 -0500, Jim Cornette wrote:
Beartooth wrote:
su -
NOW lspci and lsmod work strait from the prompt, sure enough -- and lspci | grep -i net gives me (retyping) "00:09.0 Ethernet controller: Intel Corporation 82557/8/9 [Ethernet Pro 100] (rev 10)" -- and no mention of any other ethernet anything, 3c501 or otherwise.
The 3c501 could be removed. You might want to investigate whether the Intel 82557/8/9 is supported and by what driver.
Should be the "e100" driver. According to the source code:
* I. General * * The driver supports Intel(R) 10/100 Mbps PCI Fast Ethernet * controller family, which includes the 82557, 82558, 82559, * 82550, 82551, and 82562 devices. 82558 and greater controllers * integrate the Intel 82555 PHY. The controllers are used in * server and client network interface cards, as well as in * LAN-On-Motherboard (LOM), CardBus, MiniPCI, and ICHx * configurations. 8255x supports a 32-bit linear addressing * mode and operates at 33Mhz PCI clock rate.
So verify that the "e100" driver is loaded via "/sbin/lsmod".
---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - "And on the seventh day, He exited from append mode." - ----------------------------------------------------------------------
On Wed, 2007-01-31 at 21:24 +0000, Beartooth wrote:
Well -- doubting, I tried it. But sure enough, I still get
bash: lspci: command not found
Try again, as the root user...
On Wed, 31 Jan 2007 15:44:34 -0700, Craig White wrote:
as user...
echo $PATH /usr/kerberos/bin:/usr/java/jre/bin:/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/NX/bin
as user that performs 'su' /usr/kerberos/sbin:/usr/java/jre/bin:/usr/kerberos/bin:/usr/java/jre/bin:/usr/local/bin:/usr/bin:/bin:/usr/X11R6/bin:/usr/NX/bin:/usr/NX/bin
as user that performs 'su -' # note the dash /usr/kerberos/sbin:/usr/kerberos/bin:/usr/java/jre/bin:/usr/local/sbin:/usr/local/bin:/sbin:/bin:/usr/sbin:/usr/bin:/usr/NX/bin:/root/bin
Note that each is different and it is an important difference...
# which lspci lsmod /sbin/lspci /sbin/lsmod
and /sbin path (and for that matter, /usr/sbin path) is only available to the user who performs the command 'su -'
OK, Very Dumb Question : is "su - " any different from "su - root"?
It was a great day for those of us with trifocal fingers and arthritic eyeballs -- all praise to linux developers! -- when some kind soul enhanced the gnome terminal so that it keeps a separate bash history for each profile. Nobody goes near my machines but me; I can afford to keep two tabs open to root profile routinely -- and start each time with a stroke or two to the up arrow. So I might as well have the most powerful command on it, and stick to that.
On Wed, 31 Jan 2007 16:02:06 -0800, Rick Stevens wrote:
On Wed, 2007-01-31 at 18:50 -0500, Jim Cornette wrote:
[...]
The 3c501 could be removed. You might want to investigate whether the Intel 82557/8/9 is supported and by what driver.
Should be the "e100" driver. According to the source code:
I. GeneralThe driver supports Intel(R) 10/100 Mbps PCI Fast Ethernetcontroller family, which includes the 82557, 82558, 82559,82550, 82551, and 82562 devices. 82558 and greater controllersintegrate the Intel 82555 PHY. The controllers are used inserver and client network interface cards, as well as inLAN-On-Motherboard (LOM), CardBus, MiniPCI, and ICHxconfigurations. 8255x supports a 32-bit linear addressingmode and operates at 33Mhz PCI clock rate.So verify that the "e100" driver is loaded via "/sbin/lsmod".
All right, making assurance doubly sure, I logged in using "su - root" *and* did cd before /sbin/lsmod.
I see columns labelled Module, Size, and Used by. "e100" is not mentioned in any of them -- and without a Net connection, I have no way to download anything. Can I get it off the CDs, as I must have been doing before?
Or is it there somewhere and has to be activated or enabled somehow?
On Thu, 2007-02-01 at 16:58 +0000, Beartooth wrote:
On Wed, 31 Jan 2007 16:02:06 -0800, Rick Stevens wrote:
On Wed, 2007-01-31 at 18:50 -0500, Jim Cornette wrote:
[...]
The 3c501 could be removed. You might want to investigate whether the Intel 82557/8/9 is supported and by what driver.
Should be the "e100" driver. According to the source code:
I. GeneralThe driver supports Intel(R) 10/100 Mbps PCI Fast Ethernetcontroller family, which includes the 82557, 82558, 82559,82550, 82551, and 82562 devices. 82558 and greater controllersintegrate the Intel 82555 PHY. The controllers are used inserver and client network interface cards, as well as inLAN-On-Motherboard (LOM), CardBus, MiniPCI, and ICHxconfigurations. 8255x supports a 32-bit linear addressingmode and operates at 33Mhz PCI clock rate.So verify that the "e100" driver is loaded via "/sbin/lsmod".
All right, making assurance doubly sure, I logged in using "su - root" *and* did cd before /sbin/lsmod.
I see columns labelled Module, Size, and Used by. "e100" is not mentioned in any of them -- and without a Net connection, I have no way to download anything. Can I get it off the CDs, as I must have been doing before?
It should have been installed by default. Verify that you have the following file on the system:
/lib/modules/`uname -r`/kernel/drivers/net/e100.ko
(yes, those are graves or "backticks"--the other character on the "~" key). If you do have the file, then (as root via the "su -") try this command:
modprobe e100
to force the driver to load. Then run:
ifconfig -a
and see if eth0 doesn't magically show up. If it does, NOW you have a driver loaded and the NIC is seen. Add the line:
alias eth0 e100
to your /etc/modprobe.conf file and you should be good to go. Configure the NIC as you wish and "Bob's your Uncle!".
---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - "Very funny, Scotty. Now beam down my clothes." - ----------------------------------------------------------------------
Beartooth escribió:
and /sbin path (and for that matter, /usr/sbin path) is only available to the user who performs the command 'su -'
OK, Very Dumb Question : is "su - " any different from "su - root"?
No!
$ man su
On Thursday 01 February 2007 16:58, Beartooth wrote:
All right, making assurance doubly sure, I logged in using "su - root" *and* did cd before /sbin/lsmod.
I see columns labelled Module, Size, and Used by. "e100" is not mentioned in any of them -- and without a Net connection, I have no way to download anything. Can I get it off the CDs, as I must have been doing before?
Or is it there somewhere and has to be activated or enabled somehow?
Beartooth, when I had a problem on an old laptop I deleted all the existing wifi setups (from the network gui tool), then ran kudzu. It found the connection correctly, and everything was fine after that.
If it doesn't work for you you're no worse off than now :-)
Anne
On Thu, 01 Feb 2007 09:17:11 -0800, Rick Stevens wrote:
On Thu, 2007-02-01 at 16:58 +0000, Beartooth wrote:
[...]
So verify that the "e100" driver is loaded via "/sbin/lsmod".
All right, making assurance doubly sure, I logged in using "su - root" *and* did cd before /sbin/lsmod.
I see columns labelled Module, Size, and Used by. "e100" is not mentioned in any of them -- and without a Net connection, I have no way to download anything. Can I get it off the CDs, as I must have been doing before?
It should have been installed by default. Verify that you have the following file on the system:
/lib/modules/`uname -r`/kernel/drivers/net/e100.ko
(yes, those are graves or "backticks"--the other character on the "~" key).
I think I must have retyped that correctly, because bash turned the uname part into "2.6.19-1.2895.fc6" and left the rest. But I also think I do not have it, because bash then added ":No such file or directory"
If you do have the file, then (as root via the "su -") try this command:
modprobe e100
to force the driver to load.
I tried that anyway, and sure enough (I think) it responded with a line saying "-bash: modeprobe: command not found"
I did the install, as usual, in custom mode -- and tried to leave out games, chat, and anything else I thought surely I'd never use (but took lots of development stuff, because I do run betas like Pan and Dillo). This has been my usual practice for years -- only, this time, I must have mis-guessed something inessential that I should have taken.
Or this may be the same machine on which recently I got so snarled up in yum dependency hell that I finally commanded "yum remove elfutils" -- and didn't read the list of other things it would remove carefully enough. (I don't think it is. I think I just gave up, wiped that install with DBAN, and installed again.)
Anybody know a straightforward way to get it from the CDs?
Or failing that, if I have to do the whole install over, what in particular (besides elfutils) should I make a particular point of getting?
On Thu, 2007-02-01 at 19:16 +0000, Beartooth wrote:
On Thu, 01 Feb 2007 09:17:11 -0800, Rick Stevens wrote:
On Thu, 2007-02-01 at 16:58 +0000, Beartooth wrote:
[...]
So verify that the "e100" driver is loaded via "/sbin/lsmod".
All right, making assurance doubly sure, I logged in using "su - root" *and* did cd before /sbin/lsmod.
I see columns labelled Module, Size, and Used by. "e100" is not mentioned in any of them -- and without a Net connection, I have no way to download anything. Can I get it off the CDs, as I must have been doing before?
It should have been installed by default. Verify that you have the following file on the system:
/lib/modules/`uname -r`/kernel/drivers/net/e100.ko
(yes, those are graves or "backticks"--the other character on the "~" key).
I think I must have retyped that correctly, because bash turned the uname part into "2.6.19-1.2895.fc6" and left the rest. But I also think I do not have it, because bash then added ":No such file or directory"
Hmmm. Let me try:
[root@prophead ~]# ls /lib/modules/`uname -r`/kernel/drivers/net/e100.ko /lib/modules/2.6.18-1.2869.fc6/kernel/drivers/net/e100.ko
If you do have the file, then (as root via the "su -") try this command:
modprobe e100
to force the driver to load.
I tried that anyway, and sure enough (I think) it responded with a line saying "-bash: modeprobe: command not found"
You must be root ("su -") and it's "modprobe" (only one "e"). If you're not root, then "/sbin/modprobe e100".
I did the install, as usual, in custom mode -- and tried to leave out games, chat, and anything else I thought surely I'd never use (but took lots of development stuff, because I do run betas like Pan and Dillo). This has been my usual practice for years -- only, this time, I must have mis-guessed something inessential that I should have taken.
Or this may be the same machine on which recently I got so snarled up in yum dependency hell that I finally commanded "yum remove elfutils" -- and didn't read the list of other things it would remove carefully enough. (I don't think it is. I think I just gave up, wiped that install with DBAN, and installed again.)
Anybody know a straightforward way to get it from the CDs?
It's part of the kernel RPM. If you've installed the kernel, you've got it.
So, here's what to do:
1. Run "su -" and become root 2. Run "ls /lib/modules/`uname -r`/kernel/drivers/net/e100.ko" and verify you have the file. 3. Delete the "/etc/sysconfig/hwconf" file and run kudzu to rediscover your hardware:
rm /etc/sysconfig/hwconf kudzu
4. Run "modprobe e100" to load the driver 5. Run "ifconfig" and verify you have device eth0. 6. Try to configure eth0 using the command line:
ifconfig eth0 192.168.0.254 netmask 255.255.255.0
7. Try "ifconfig" again. You should see something like:
[root@prophead ~]# ifconfig eth0 Link encap:Ethernet HWaddr 00:0F:FE:02:16:38 inet addr:192.168.0.254 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::20f:feff:fe02:1638/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:33007499 errors:0 dropped:0 overruns:0 frame:0 TX packets:15837216 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:2322911622 (2.1 GiB) TX bytes:616269113 (587.7 MiB) Interrupt:193
The "HWaddr", "inet6 addr:", "Interrupt" and traffic info (RX/TX bytes) will differ from what I show above, but you get the idea.
8. If you see that stuff, great! Edit the /etc/modprobe.conf file and add a line that reads:
alias eth0 e100
9. Run your normal configuration. Delete all existing configs and put in the new stuff you want.
That should handle it. ---------------------------------------------------------------------- - Rick Stevens, Senior Systems Engineer rstevens@vitalstream.com - - VitalStream, Inc. http://www.vitalstream.com - - - - The gene pool could use a little chlorine. - ----------------------------------------------------------------------
On Thu, 01 Feb 2007 17:58:08 +0000, Anne Wilson wrote:
Beartooth, when I had a problem on an old laptop I deleted all the existing wifi setups (from the network gui tool), then ran kudzu. It found the connection correctly, and everything was fine after that.
If it doesn't work for you you're no worse off than now :-)
Well, I don't do wifi on that machine; strictly an ethernet cable connection. But I took you to mean eth0 and eth1; deleted them; ran kudzu, tried to add them back; and got more of the same old same old.
So I deleted them again, and rebooted. Then I tried "yum update yum" (as being a relatively small task). No joy.
Just in case, I did "yum clean all" followed by "rpm --rebuilddb" -- which ran quite a long time, as usual. It still didn't connect.
Since I last got into Gmane, Rick Stevens has suggested another hardware-removal procedure, which I'll be trying while anyone reads this. Stay tuned.
On Thu, 01 Feb 2007 11:52:30 -0800, Rick Stevens wrote:
On Thu, 2007-02-01 at 19:16 +0000, Beartooth wrote:
[...]
I think I must have retyped that correctly, because bash turned the uname part into "2.6.19-1.2895.fc6" and left the rest. But I also think I do not have it, because bash then added ":No such file or directory"
Hmmm. Let me try:
[root@prophead ~]# ls /lib/modules/`uname -r`/kernel/drivers/net/e100.ko /lib/modules/2.6.18-1.2869.fc6/kernel/drivers/net/e100.ko
Odd : you appear to be running an older kernel than I got off my CD ...
Anybody know a straightforward way to get it from the CDs?
It's part of the kernel RPM. If you've installed the kernel, you've got it.
So, here's what to do:
- Run "su -" and become root
- Run "ls /lib/modules/`uname -r`/kernel/drivers/net/e100.ko" and verify you have the file.
Aargh. It says "ls: /lib/modules/`uname -r`/kernel/drivers/net/e100.ko: no such file or directory"
Delete the "/etc/sysconfig/hwconf" file and run kudzu to rediscover your hardware:
rm /etc/sysconfig/hwconf kudzu
I did that slightly differently, just to be sure : "cd /etc/sysconfig" followed by "ls | grep hw" -- which did give me hwconf -- then deleted it and ran kudzu (which gave me my prompt back, with no report).
- Run "modprobe e100" to load the driver 5.
That gave
"FATAL: Could not load /lib/modules/2.6.19-1.2895.fc6/modules.dep: No such file or directory."
So I did cd / (just to be sure), then cd /lib, then ls; sure enough, there's a subdirectory called modules. So I tried cd modules, then ls -- and all it contains is a subdirectory, called *not* 2.6.19-[and so on], as we got from `uname -r` -- but 2.6.18-1.2798.fc6
I looked in that. It contains build, extra, kernel, source, updates, weak-updates, and eleven things of the form modules.*
Here I stop for now -- have we discovered something?
Run "ifconfig" and verify you have device eth0. 6. Try to configure eth0 using the command line:
ifconfig eth0 192.168.0.254 netmask 255.255.255.0
- Try "ifconfig" again. You should see something like:
[root@prophead ~]# ifconfig eth0 Link encap:Ethernet HWaddr 00:0F:FE:02:16:38 inet addr:192.168.0.254 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::20f:feff:fe02:1638/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:33007499 errors:0 dropped:0 overruns:0 frame:0 TX packets:15837216 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:2322911622 (2.1 GiB) TX bytes:616269113 (587.7 MiB) Interrupt:193
The "HWaddr", "inet6 addr:", "Interrupt" and traffic info (RX/TX bytes) will differ from what I show above, but you get the idea.
- If you see that stuff, great! Edit the /etc/modprobe.conf file and
add a line that reads:
alias eth0 e100
- Run your normal configuration. Delete all existing configs and put
in the new stuff you want.
That should handle it.
On Mon, 29 Jan 2007 20:53:20 +0000, Beartooth wrote:
[...]
So I took out the CD and tried again. It still kept telling me device eth0 "seems not to be present." I configured an eth1; trying to activate that got the same result.
[...] I want to thank everyone for all the instruction; I learned a lot of stuff, even though the fool install never did connect.
I finally wiped the whole machine, yet again, with DBAN -- and had another crack at installing CentOS on one hard drive, with FC6 on the other. I thought I had figured out what to tell anaconda when.
CentOS does connect, without a hitch; I installed Pine, Opera, and a couple other things, and did yum update, without a hitch.
But I was wrong yet again about how to set up the dual boot -- it's trying to use rootnoverify in /etc/grub.conf, which doesn't work; and I shudder at trying once more to gnaw my way through the available grub instructions to dope out precisely enough how to make the FC entry look enough like the CentOS entries.
Any hints on a sure way to install either of these OSs after the other? I suspect, it'll be a lot easier to re-do the install -- especially now, before I get any data on that drive -- or even tweak my workspace.
On Mon, 2007-02-05 at 21:41 +0000, Beartooth wrote:
But I was wrong yet again about how to set up the dual boot -- it's trying to use rootnoverify in /etc/grub.conf, which doesn't work; and I shudder at trying once more to gnaw my way through the available grub instructions to dope out precisely enough how to make the FC entry look enough like the CentOS entries.
For information's sake, this is one my FC6 GRUB stanzas:
title Fedora Core (2.6.18-1.2257.fc5) root (hd0,0) kernel /vmlinuz-2.6.18-1.2257.fc5 ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.18-1.2257.fc5.img
The (hd0,0) has to be set for where your /boot partition is, the root= bit for where your system / is, which may be an ordinary partition or a volume group, and hte version numbers have to suit what you've installed.
I avoid volume groups, they're too hard to fix when they go wrong, and aren't of any use when you've got a box that can only have one HDD inserted into them.
Any hints on a sure way to install either of these OSs after the other? I suspect, it'll be a lot easier to re-do the install -- especially now, before I get any data on that drive -- or even tweak my workspace.
You can install Fedora, but *not* install GRUB to the MBR, just into its own partitions. That'll leave prior installs alone, and you could copy and paste its grub file stanzas to the prior installs ones. That's just one way of managing it.
On Tue, 2007-02-06 at 16:35 +1030, Tim wrote:
For information's sake, this is one my FC6 GRUB stanzas:
title Fedora Core (2.6.18-1.2257.fc5) root (hd0,0) kernel /vmlinuz-2.6.18-1.2257.fc5 ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.18-1.2257.fc5.img
That's really my FC5 box, of course. I was cutting and pasting from the wrong box. FC6 is similar, but I can't copy from it at the moment.
On Tue, 06 Feb 2007 16:35:16 +1030, Tim wrote: [...]
Any hints on a sure way to install either of these OSs after the other? I suspect, it'll be a lot easier to re-do the install -- especially now, before I get any data on that drive -- or even tweak my workspace.
You can install Fedora, but *not* install GRUB to the MBR, just into its own partitions. That'll leave prior installs alone, and you could copy and paste its grub file stanzas to the prior installs ones. That's just one way of managing it.
If the other OS is on hda, then I just tell fedora to install on hdb only? Then how do I boot to it to see the grub stanzas to copy?
On Tue, 06 Feb 2007 16:35:16 +1030, Tim wrote:
On Mon, 2007-02-05 at 21:41 +0000, Beartooth wrote:
But I was wrong yet again about how to set up the dual boot -- it's trying to use rootnoverify in /etc/grub.conf, which doesn't work; and I shudder at trying once more to gnaw my way through the available grub instructions to dope out precisely enough how to make the FC entry look enough like the CentOS entries.
For information's sake, this is one my FC6 GRUB stanzas:
title Fedora Core (2.6.18-1.2257.fc5) root (hd0,0) kernel /vmlinuz-2.6.18-1.2257.fc5 ro root=LABEL=/ rhgb quiet initrd /initrd-2.6.18-1.2257.fc5.img
The (hd0,0) has to be set for where your /boot partition is, the root= bit for where your system / is, which may be an ordinary partition or a volume group, and hte version numbers have to suit what you've installed.
[...] All right; since CentOS does connect, I can ssh to it; doing that, I get the following on the p2 machine :
[root@localhost ~]# df Filesystem 1K-blocks Used Available Use% Mounted on /dev/mapper/VolGroup00-LogVol00 18803436 3530836 14317416 20% / /dev/hda1 101086 12455 83412 13% /boot none 192944 0 192944 0% /dev/shm [root@localhost ~]#
If I'm not totally fouled up (yet), that tells me I do have a boot partition on hda, and it's at hda1. Right so far?
Then I can edit /etc/grub.conf on the p2, go to the bottom of the file, delete the rootnoverify and chainload lines, copy & paste your stanza above onto it (with fc5 replaced by fc6) -- or maybe an fc6 stanza from /etc/grub.conf on one of my own other machines, which would also get an fc6 kernel version -- and then what?? Do I just leave what shows as hd(0,0) still hd(0,0)??
Fwiw, here're a couple fedora stanzas from /etc/grub.conf on the machine I'm using at the moment :
title Fedora Core (2.6.19-1.2895.fc6xen) root (hd0,0) kernel /xen.gz-2.6.19-1.2895.fc6 module /vmlinuz-2.6.19-1.2895.fc6xen ro root=/dev/VolGroup00/LogVol 00 module /initrd-2.6.19-1.2895.fc6xen.img
title Fedora Core (2.6.19-1.2895.fc6) root (hd0,0) kernel /vmlinuz-2.6.19-1.2895.fc6 ro root=/dev/VolGroup00/LogVol00 initrd /initrd-2.6.19-1.2895.fc6.img
title Fedora Core (2.6.18-1.2869.fc6) root (hd0,0) kernel /vmlinuz-2.6.18-1.2869.fc6 ro root=/dev/VolGroup00/LogVol00 initrd /initrd-2.6.18-1.2869.fc6.img
Will it work if I just put that (without the xen stanza) verbatim onto the bottom of the /etc/grub.conf that's in the p2 now? Will it find hdb by itself?