FC4 and Intel 82559 NIC?

Harold Hallikainen harold at hallikainen.com
Mon Jan 30 02:07:13 UTC 2006


> On Sat, 2006-01-21 at 14:23 -0800, Harold Hallikainen wrote:
>> >
>> >> On Tue, 2006-01-10 at 20:46 -0800, Harold Hallikainen wrote:
>> >>> > On Sun, 2006-01-08 at 19:06 -0800, Harold Hallikainen wrote:
>> >>> >> >
>> >>> >> > This weekend I loaded FC4 onto my server that had been running
>> >>> RH8.
>> >>> Everything seems to have gone smoothly EXCEPT the LAN does not work.
>> >>> >> It
>> >>> >> > seems to find the two ethernet interfaces, but does not talk to
>> >>> the
>> >>> remainder of the network. I've tried both static IP and DHCP.
>> Neither
>> seems to work. Do I need a special driver for the NIC? If so, where do
>> >>> >> I
>> >>> >> > find it, and how do I install it?
>> >>> >> >
>> >>> >> > THANKS!
>> >>> >> >
>> >>> >> > Harold
>> >>> >> >
>> >>> >> Following up... I found the e100 driver at
>> >>> >> http://www.intel.com/support/network/adapter/1000/linux/e100.htm
>> and
>> >>> followed the instructions there. Stuff is still not working. Looking
>> at
>> >>> /var/log/messages, I see
>> >>> >> e100: eth0: eth0: e100_probe: addr 0xd6202000, irq 11, MAC addr
>> >>> 00:E0:81:03:AF:3C
>> >>> >> I see a similar message for eth1.
>> >>> >> A while later is
>> >>> >> e100: eth0: e100_wathcdog: link up, 10Mbps, half-duplex
>> >>> >> A while later it cannot find ntp.org
>> >>> >> Then NETDEV WATCHDOG eth0; transmit timed out
>> >>> >> then another watchdog message.
>> >>> >> Any ideas?
>> >>> >
>> >>> > First, as the root user, do "ifconfig" and verify that there's
>> >>> actually
>> >>> IP addresses assigned to your NIC.  Next (also as root), do "netstat
>> -rn"
>> >>> and verify that the default gateway is pointing at your router.
>> >>> >
>> >>> > If that's all good, then you may be blocked by either iptables or
>> >>> SELinux.  Try turning off iptables temporarily by "service iptables
>> off"
>> >>> (as root again).  If that doesn't work, try (as root) "setenforce 0"
>> to
>> >>> temporarily disable SELinux.
>> >>> >
>> >>> Thanks for the responses! It's really a pain to not have network
>> connectivity. I can't just copy and paste stuff into an email.
>> Instead,
>> >>> I'm copying off one screen and typing on another (my laptop). So
>> excuse
>> >>> the typos! It's interesting that I'm having network problems with
>> this
>> one
>> >>> machine. I've installed FC4 on another server and my laptop with no
>> problem. Oh well...
>> >>> I tried the stuff above to get the network back and was about to
>> type all
>> >>> the results when I thought I'd try something else (anything to avoid
>> typing all that stuff). This is a dual processor machine that had
>> always
>> >>> worked fine using SMP on RH8. I tried telling it to NOT run smp on a
>> reboot, and the network works great! No other changes! I'll mess with
>> it
>> >>> some more in the next couple days. Is anyone else running smp on
>> FC4?
>> >> I do, a lot.  Two things to check...first, make bloody sure the
>> motherboard is running the latest BIOS from the vendor.  If this is an
>> AMD-based SMP machine, try booting with the "noapic" option set (add "
>> noapic" to the end of the kernel line at the grub boot screen). If you
>> don't know how to do that:
>> >> 1. Boot the machine.
>> >> 2. At the screen that displays the various kernels, use the arrow
>> keys
>> to select the SMP kernel and hit "E".
>> >> 3. Use the arrow keys to scroll down to the line that starts with
>> "kernel".
>> >> 4. Hit "E" again.  Use the arrow keys to go to the end of the line.
>> 5.
>> Put in " noapic" (don't forget the space) at the end of the line. 6.
>> Hit "ENTER" to exit the edit mode, then "B" to boot the machine. 7. The
>> machine should now boot with the "noapic" option.
>> >> NOTE: This is temporary and will only survive this boot.  If you need
>> the "noapic" option permanently, you must edit the /boot/grub/grub.conf
>> file.
>> >
>> >
>> > Thanks! I'll give it a try! The BIOS is the original one that came
>> with
>> the board (maybe 5 years ago). I haven't done any BIOS upgrades in a
>> while, but back when I did, they seemed to assume you were running DOS.
>> I
>> > just looked at the Tyan website and see that BIOS updates are done by
>> booting from floppy with DOS. I'll try upgrading the BIOS as soon as I
>> can
>> > find the time (within a week, I hope). Ideally that will ALSO fix the
>> problem I had when I tried a pair of 80G drives (install went fine, but
>> then during reboot the drives could not be found by FC4... the BIOS says
>> they're there, though.).
>> >
>> > THANKS!
>> >
>> > Harold
>> >
>>
>> Following up on this, the bios update DID get the NIC working! I'm
>> really
>> surprised the bios had anything to do with the NIC. I tend to think of
>> the
>> old days where bios did more basic things... Anyway, it works! The
>> update
>> still doesn't make my new Hitachi 80G drives work, but the old 40G
>> drives
>> are working fine, so I guess I'll stick with them for a while. During
>> the
>> FC4 install, pretty early in the install, I get a complaint that hda
>> could
>> not be found. The system then continues with the install. When
>> installing
>> the file system, it properly identifies both drives. The installation
>> then
>> continues fine. On reboot, the system starts and loads some stuff and
>> then
>> gives some sort of drive not found error.
>>
>> Error messages on reboot include:
>> Couldn't find device with uuid '.....' (skipping the long numbers)
>> Couldn'tfind all physical volumes for volume group VolGroup01.
>> Couldn't find device with uuid '.....' (same id as last time)
>> Volume group "VolGroup01" not found
>> ERROR: /bin/lvm exited abnormally with value 5 !(pid 396).
>>
>> ... bunch more error messages, then kernel panic.
>>
>> So, I guess the bios perhaps doesn't like these drives? But it (and FC4)
>> seem to identify them properly.
>
> It appears that there's a spinup issue with one of the drives.  You set
> up LVM and one of the drives that makes up the volume group (a "physical
> volume" or "PV") didn't answer the bell.  So, the volume group ("VG")
> couldn't be created.  Without the VG, the logical volume ("LV") couldn't
> be brought up and since you probably have "/" on it, you don't have
> the root of the filesystem.
>
> See if there's a setting on your BIOS that causes the drives to spin up
> at power up rather than when the bus is scanned.  This will give the
> drives a bit longer to spin up and the laggard drive may come up in time
> for the PV scan to see it.
>


THANKS for the help! No such setting in the BIOS. I did see a jumper
setting on the drives marked "autospin disable." From the description, it
looks like setting autospin disable would keep the drive from spinning up
on power up. Instead, it would wait for an ATA command. I don't have that
jumper in, so it looks like the drive should spin up on power up. So, the
problem remains.

I've got my old 40G drives in the machine and will probably just leave it
like that until I run out of room. This is a standby server anyway, and
the motherboard is getting old. By the time I run out of room, I'll
probably be ready for another new machine.

THANKS!

Harold

-- 
FCC Rules Updated Daily at http://www.hallikainen.com




More information about the Redhat-install-list mailing list