MD boot issue in 810

[lemming86_au]

I haven't tried that, I will try it when I get home, thanks.
So which should they use nv or nvidia?

[pigdog]

Hi,

The driver should be nvidia for your video cards but if you have an older version (96 legacy) loaded you would run nv.

If the 180.11 is running in the NVIDIA X server settings you might need to try installing the 185 variant of the drivers.

[dothedog]

lemming86_au,
Did the video driver make a difference? I have had 3 different motherboards not start with the same symptoms (kt3 ultra/e1000/gf8500, via epia, atom330/rtl8111c/i945). I have actually gotten it past the starting chrond but now it is hanging at starting NIS services. Please post if you get past this.

DoTheDog

[WhateverFits]

I had what appears to be the same problem. If you hit Control-F12 (IIRC) then you get a list of recent messages from the system. The very last one I get is about the network card changing state. Network cards changing their connection state on a network boot is not a good thing. I got this when I updated my media director and it pulled some update from Ubuntu that broke this for me. I'm still trying to figure out what package.

[dothedog]

Whateverfits,
Did you ever get it figured out? I actually got an atom330 board to run. But when I tried to get the second one running I end up getting this:

Code: [Select]

Oct  1 21:55:01 moon61 NetworkManager: <info>  (eth0): carrier now ON (device state 1)
Oct  1 21:55:01 moon61 nm-system-settings:    SCPlugin-Ifupdown: init!
Oct  1 21:55:01 moon61 anacron[6353]: Jobs will be executed sequentially
Oct  1 21:55:01 moon61 kernel: [    0.905038]  domain 0: span 1,3 level SIBLING
Oct  1 21:55:01 moon61 nm-system-settings:    SCPlugin-Ifupdown: update_system_hostname
Oct  1 21:55:01 moon61 kernel: [    0.905043]   groups: 3 1
Oct  1 21:55:01 moon61 kernel: [    0.905050]   domain 1: span 0-3 level CPU
Oct  1 21:55:02 moon61 nm-system-settings:    SCPluginIfupdown: guessed connection type (eth0) = 802-3-ethernet
Oct  1 21:55:02 moon61 kernel: [    0.905055]    groups: 1,3 0,2
Oct  1 21:55:02 moon61 kernel: [    0.905284] net_namespace: 840 bytes
Oct  1 21:55:02 moon61 kernel: [    0.905284] Booting paravirtualized kernel on bare hardware
Oct  1 21:55:02 moon61 kernel: [    0.905284] Time: 21:52:58  Date: 10/01/09
Oct  1 21:55:02 moon61 nm-system-settings:    SCPlugin-Ifupdown: update_connection_setting_from_if_block: name:eth0, type:802-3-ethernet, autoconnect:0, id:Ifupdown (eth0), uuid: 681b428f-0000-0000-0000-000039000000
Oct  1 21:55:02 moon61 kernel: [    0.905284] NET: Registered protocol family 16
Oct  1 21:55:02 moon61 nm-system-settings:    SCPlugin-Ifupdown: addresses count: 1
Oct  1 21:55:02 moon61 kernel: [    0.908120] EISA bus registered
Oct  1 21:55:02 moon61 nm-system-settings:    SCPlugin-Ifupdown: No dns-nameserver configured in /etc/network/interfaces
Oct  1 21:55:02 moon61 kernel: [    0.908123] ACPI: bus type pci registered
Oct  1 21:55:02 moon61 nm-system-settings:    SCPlugin-Ifupdown: autoconnect
Oct  1 21:55:02 moon61 kernel: [    0.908203] PCI: Found Intel Corporation 945G/GZ/P/PL Express Memory Controller Hub without MMCONFIG support.
Oct  1 21:55:02 moon61 nm-system-settings:    SCPluginIfupdown: management mode: unmanaged
Oct  1 21:55:03 moon61 kernel: [    0.910017] PCI: Using configuration type 1 for base access
Oct  1 21:55:03 moon61 kernel: [    0.916547] ACPI: EC: Look up EC in DSDT
Oct  1 21:55:03 moon61 nm-system-settings:    SCPlugin-Ifupdown: device added (udi: /org/freedesktop/Hal/devices/net_00_1c_c0_c9_2b_e3, iface: eth0): well known
Oct  1 21:55:03 moon61 /USR/SBIN/CRON[6582]: (root) CMD (/usr/sbin/ntpdate dcerouter &>/dev/null)
Oct  1 21:55:03 moon61 kernel: [    0.923568] ACPI: Interpreter enabled
Oct  1 21:55:03 moon61 nm-system-settings:    Ifupdown: get unmanaged devices count: 1
Oct  1 21:55:03 moon61 kernel: [    0.923623] ACPI: (supports S0 S1 S3 S4 S5)
Oct  1 21:55:03 moon61 nm-system-settings:    SCPlugin-Ifupdown: (153420008) ... get_connections.
Oct  1 21:55:03 moon61 NetworkManager: <info>  (eth0): device state change: 1 -> 2
Oct  1 21:55:03 moon61 kernel: [    0.923817] ACPI: Using IOAPIC for interrupt routing

I believe that is the same "state change" you mention in your previous post? I really wish I knew what the voodoo magic to get this to boot is. The funny thing is, the one that actually boots (exact same atom 330 board) won't boot consistently, it usually boots up on the second or third try...

I have tried compiling the latest r8168 driver from realtek (it uses it for boot), added Driver "intel" to my xorg.conf file (its an intel i945), added dns-nameserver 192.168.80.1 to /etc/network/interfaces (because it complained about that), changed architecture from amd64 to i386, and none of it seems to help. If you have had any progress please let me know.

Thanks,
DoTheDog

[WhateverFits]

<info>  (eth0): device state change: 1 -> 2

Yeah, that looks familiar. No, I never figured it out since I got it working, updated just the pluto/linuxmce stuff, ignored the core Ubuntu updates, and all is working then I started back to college and have had no time to monkey with it. I'll probably have LOTS of time in December between semesters to test this out and get more detail. My current belief is it was one of the packages (from Ubuntu, not linuxmce) that has to do with the networking stack, not the drivers or kernel itself. I upgraded all my packages except the kernel once and still got the problem.

I have a large list to go through and I have to reboot every time to test so it takes a while. I plan on doing a binary search for it. Upgrade half the list. Reboot. Fail? Rebuild and cut that list in half again. Of course, I don't know what do do once I find the package!

[dothedog]

Whateverfits, So how did you "ignore the core Ubuntu updates". Did you do this on the Core or on the MD? Did you just comment out the ubuntu sources in the your apt sources? Let me know. I would love to get these things working.

Dothedog

[dothedog]

Ok Guys, I think I solved this:
since you can't boot into a tty when you get this, I did a chroot

Code: [Select]

sudo chroot /usr/pluto/diskless/<moon#> /bin/bash


Code: [Select]

update-rc.d -f NetworkManager remove
reboot MD

Apparently there is a bug in NetworkManager in Intrepid that has a problem with static IP addresses. You don't need NetworkManager if you are only using static ips.

Please try this and let me know if it works. 

Dothedog

[lemming86_au]

Dothedog: You legend!

I was getting the (eth0): device state change: 1 -> 2 as well, but removing NetworkManager as you suggested has fix the problem!

Regards,
Josh

[dothedog]

lemming86_au,
I'm glad it worked for you. I just marked this one solved I am surprised more people don't have this issue. I had the same thing happen on 4 different ethernet cards, so I figured it wasn't a hardware driver issue. Took a lot of googling but I'm glad we got it figured out.

Dothedog

[WhateverFits]

dothedog, I ignored the other updates by manually running apt-get update packagenamehere on my MD.

But I guess that doesn't matter any longer since disabling NetworkManager worked perfectly! Thanks!

[levster]

Ok Guys, I think I solved this:
since you can't boot into a tty when you get this, I did a chroot

Code: [Select]

sudo chroot /usr/pluto/diskless/<moon#> /bin/bash


Code: [Select]

update-rc.d -f NetworkManager remove
reboot MD

Apparently there is a bug in NetworkManager in Intrepid that has a problem with static IP addresses. You don't need NetworkManager if you are only using static ips.

Please try this and let me know if it works. 

Dothedog

dothedog,

The process you describe here, is it to enable MD to PXE? I have a problem with my two Zotac boards (330 and 230), with both reporting the same error while trying to go through the initial steps of network booting for the first time. Both PCs reboot after about 10 to 15 seconds, occasionally reporting that they "cannot connect to router". Do I have to change something on the Core?

Sorry for the novice question,

Lev

[lemming86_au]

Levster,

The "cannot connect to router" is a different issue than what is being discussed here, but I have noticed some other topics in the forum discussing it.

I believe there is a temp fix in this topic - http://forum.linuxmce.org/index.php?topic=8939.0

Regards,
Josh

[levster]

Josh,

Thanks.

I was able to push myself a bit further with "/usr/pluto/bin/startup-script.sh", but am now stuck on "diskless setup failed" message....  

[samsung]

Josh,

Thanks.

I was able to push myself a bit further with "/usr/pluto/bin/startup-script.sh", but am now stuck on "diskless setup failed" message....  

yup me 3.