I have recently decided to try to switch entirely from Windows XP to Redhat 8.0 but have been met with an initial problem that makes it almost impossible.

I installed redhat initially and my network card was not detected.

I realized I had to turn off plug and play OS in my BIOS, did so, and reinistalled with network card successfully detected and installed.

Internet and network works fine, everything is perfect.

After a few days, internet stops working entirely. Server not found type messages for any network activity.

Reinstalled redhat using same procedure as before, when it worked.

Network card found and setup, but no internet or network.

This is basically where I am now, and I am completely stumped. I have reinstalled 3 times using the same procedure as the time when it worked, with no success.

Here is what I get with ifconfig

eth0 Link encap: Ethernet Hwaddr 00:A0:CC:E6:76:69
inet addr: 209.246.18.170 BCast: 209.246.19.255 Mask 255.255.255.128
UP BROADCAST RUNNING MULTICAST MTU: 1500 Metric: 1
RX packets: 7 errors: 0 dropped: 0 overruns: 0 frame: 0
TX packets: 3 errors: 0 dropped: 0 overruns: 0 carrier: 0
collisions: 0 txqueuelen: 100
RX bytes: 1018 (1018.0 b) TX bytes: 726 (726.0 b)
Interrupt: 5 Base address: 0x1000

lo Link encap: Local Loopback
inet addr: 127.0.0.1 Mask: 255.0.0.0
UP LOOPBACK RUNNING MTU: 16436 Metric: 1
RX packets: 10 errors: 0 dropped: 0 overruns: 0 frame: 0
TX packets: 10 errors: 0 dropped: 0 overruns: 0 carrier: 0
collisions: 0 txqueuelen: 0
RX bytes: 700 (700.0 b) TX bytes: 700 (700.0 b)

netstat -r gives:

Kernal IP routing table

Destination: 209.246.18.128 Gateway: * Genmask: 255.255.255.128 Flags: U MSS: 40 Window: 0 irtt: 0 Iface: eth0

Destination: 127.0.0.0 Gateway: * Genmask: 255.0.0.0 Flags: U MSS: 40 Window: 0 irtt: 0 Iface: lo

Running ping or nslookup just results in "server not found" or similar message.

This is all the information I can give you at the moment. My NIC is a Linksys LNE100TX and I am using COVAD DSL on a LAN with a linksys hub and a couple of other windows and mac OS 10.2 machines.

Thanks very much for any advice, I would really appreciate it.

I should add that, looking at ifconfig, clearly some packets have been going through so its definitely not a hardware or hardware configuration problem...

The ip address that you are getting when doing an ifconfig... is that valid? Is that what you expect to see? What about name resolution? Do you have the proper entries in your /etc/resolv.conf?

What happens if you try to ping or traceroute to your DNS server?



-K.

Actually now that you mention it, the IP address that linux seems to be getting assigned is different to what I would almost always normally get on this computer in windows (it hasnt changed in years)

I'll check those other things.

Thanks

Dear Bungo,

I sympathize with your quest. I have a similar case. I 'upgraded' from a nicely working RedHat 7.1 system to a nonworking 8.0 system!

I have but a single ethernet card and my org. has a fixed IP adress for me. When I boot the bott messages say that eth0 succeeds, but the devise is 'inactive.' Activating it (through a little utility interface) has no effect. So I'm without a network.

All the RedHat people managed to ask me is which network card I have. I told them a few days ago, but no reaction yet.

What is happening here?

Redhat has been secretely bought up by Microsoft and this is an act of sabotage against newbies.

I bet Microsoft wishes they could "buy" Linux, but...

Anyway, both of your problems... What type of internet connectivity do you have? Cable Modem, DSL, Local LAN, modem or router, etc... Some details on your network setup please.

That 209.x address you say is different than what you'd normally expect. How different? Did you key it in or was it DHCP'ed (I assume DHCP but don't hurt to ask).

What brand/model (and/or chipset) is your ethernet card? Does something that looks like it show up in an "lsmod"?

What is the contents of /etc/resolv.conf?

Can you route IP via number and not name? If so, look at DNS resolution (/etc/resolv.con). If not, let's look at more basic things.

Can you ping the 209.x local address? What about localhost?

A "netstat -rn" for me pointing to my Linksys router and cable modem:

Kernel IP routing table
Destination Gateway Genmask Flags MSS Window irtt Iface
192.168.1.0 0.0.0.0 255.255.255.0 U 40 0 0 eth0
127.0.0.0 0.0.0.0 255.0.0.0 U 40 0 0 lo
0.0.0.0 192.168.1.1 0.0.0.0 UG 40 0 0 eth0

Yours wraps and is hard to tell but there doesn't appear to be a default gateway. A destination of 0.0.0.0 with a gateway of your local router or gateway device.

Let's see what can be done with the answers to a few of those questions.

I am using DSL, with a normal ethernet modem (not USB or anything) and a hub that generally works perfectly well on any OS, on any computer, without any configuration other than chosing DHCP.

The IP address given to me in Linux differs from that given to me in windows only in the last 3 digits.

I have a couple of other computers on this network, and they all have their own IP addresess that differ by those same last 3 digits, and yes, are assigned by DHCP, and that is what I use in both windows and Linux to get an IP.

My network card is a Linksys LNE100TX (version 2.0) - not sure of the chipset, but heres the product link: http://www.linksys.com/Products/prod...rid=31&grid=26

Don't know what significance this has but I remember something to do with 'Tulip' drivers in Linux for that type of card chipset . . .

Here is the contents of my /etc/resolv.conf:

; generated by /sbin/dhclient-script
search ionix.net
nameserver 209.0.104.150
nameserver 209.0.104.140
nameserver 209.0.104.150

That's all I can tell you at the moment, in the middle of writing a History essay and cant boot to linux at the moment. I will report back with the other information you reccomended I find in coming hours.

Thanks a lot for your help... this problem has really screwed up my (our) potential linux conversion!

Dear helpers

My eth0 problem got solved! It's a bit embarrassing, and then again, it's not.

It appears that when my system comes up, it starts the 7.1 kernel rather than the 8.0 kernel. I have a big nasty window on my boot screen that lists LINUX at the top followed by an entry like "2 14 18".

By hittin the space bar when that window came up it shifted from the LINUX entry to the "2 14 18" entry. Next thing I know, the system boots differently, my network is up and I can even see my jaz drive again.

What I cannot believe, however, is that an 8.0 upgrade boots with the 7.1 kernel as default! Why do they do that? And how on earth was I ever going to figure this out?

My experience has been that a RedHat "upgrade" overwrites previous RedHat kernel versions with the one on the update CD version. If you compile a "custom" kernel usually a RedHat upgrade will not overwrite that. That's about the only time I've seen one do that.

If you chose "lilo" as your boot manager look in /etc/lilo.conf

If you chose "grub" as your boot manager look in /etc/grub.conf

You should be able to see what kernels are known and which is set to be default, etc... What is defined in these files is what comes up as it's booting when you are asked which kernel to boot.

If you make changes to lilo.conf (be careful!) you have to run /sbin/lilo afterwards to re-write this info to the boot sector or MBR or wherever it is. Dont' reboot if this doens't run without any complaints! Fix any complaints it produces and re-run /sbin/lilo before rebooting.

For /etc/grub.conf you don't have to run anything, just edit and make changes to it and reboot. An advantage to grub is that if you install a new RedHat kernel RPM file it adds entries to grub.conf for that kernel automatically...

I believe you have hijacked my post, the problem is NOT solved. My redhat install was a CLEAN install and there is no problem with kernals. Thanks.

Nobody is intending to "hijack" anything Try this on for size.. RedHat is very good about something, far better than Windows in this one small thing... If you define IP settings for a certain brand/model ethernet card and later remove that card and put in a new card here is what will happen when kudzu runs on the next boot. It will inform you the old card is gone and ask if you want to remove it's configuration. Agree with it.. It will find the card and ask if you want it configured. Agree.. It will then ask if you want the configuartion of the old card migrated to the new card. Agree and you're up and running with all the same IP information. Very nice...

So, barring any more complicated fixes, you could remove the card, boot up, let kudzu remove the configuration and then shutdown and let it be found on bootup the next time. I think the "card" is fine though, it's a DHCP/routing thing.

As to your answers to my questions... Does your DSL provider give out IP addresses via DHCP rather than PPPoE or PPPoA? I deal with about 6 different DSL providers fairly regularly and none of them do DHCP (yours might though, I'm just asking for confirmation). Also, do you pay for multiple IPs so that your more than a few computers all get IPs from them? That's also not common but not unheard of. You don't have a Linksys/Netgear/Dlink or similar Cable/DSL sharing router do you? I generally use a router like this and hard-coded IPs and hopefully this info isn't outdated. Check out "pump" which I assume is what is getting a DHCP address for you from your ISP. "pump --help" for usage. A -R parameter renews for instance. The last 3 digits being different than what was leased when you had Windows booted is not a big deal I'd say...

Anyway, it looks like you are successfully getting a DHCP lease but the lack of a default route is what concerns me. Release and renew the lease if it's not working and check a "netstat -rn". If there is no route to 0.0.0.0 as "default gateway" you won't be able to communicate across the internet very effectively.

Let us know what you find...

Actually all of the information has been answered above...

but again yes I am using DHCP

again no I have no router, just a linksys hub:

DSL line - DSL modem
DSL modem - linksys hub
linksys hub - all the computers

yes we pay for multiple IPs, a couple of static ones (which are configured manually)

there is no problem with card configuration, this has been established

the card works (I am using it now in windows), and it used to work in redhat (on my *second* install, which would later go bad), I am now on my third or fourth install and it does not work.

I have never had to configure IPs manually for any OS using DHCP on this network and I don't intend to, I am thinking of just not using redhat, is this the best solution?

pump is an invalid command in redhat 8, so yes that is out of date. kudzo has no effect.

thanks again