Set-up of a tftp server for PXE installation gives TIMEOUT error

colucix · 06-25-2009, 04:37 AM

Hi all,

I'm trying to set up a tftp + DHCP server for a PXE installation on a machine running Fedora 9. My requirement is to setup a method for OS installation on an old notebook, without CD-ROM, without floppy and without USB boot capability. A nightmare!

I have followed the guide lines in the howto http://www.stanford.edu/~alfw/PXE-Ki...Kickstart.html. In summary I have modified the file /etc/dhcpd.conf file (on an already working DHCP server)

Code:

$ cat /etc/dhcpd.conf
# DHCP configuration generated by Firestarter
not authoritative;
ddns-update-style interim;
ignore client-updates;

allow bootp;
allow booting;

subnet 192.168.0.0 netmask 255.255.255.0 {
        option routers 192.168.0.1;
        option subnet-mask 255.255.255.0;
        option domain-name-servers XXX.XXX.XXX.XX;
        option ip-forwarding off;
        range dynamic-bootp 192.168.0.16 192.168.0.24;
        default-lease-time 21600;
        max-lease-time 43200;
}

group {
  next-server ocean-1;
  filename "/tftpboot/pxelinux.0";
}

then I restarted the dhcp daemon using /etc/init.d/dhcpd restart.

After that I installed the tftp-hpa server version 5.0 downloaded from http://www.kernel.org/pub/software/network/tftp/. I created a /tftpboot directory with user nobody, group nobody and world-readable permissions. I open port 69 on the local firewall. Then I started the tftp server by

Code:

# /sbin/chkconfig xinetd on
# /sbin/chkconfig tftp on

I checked that the server is actually listening on port 69 using netstat. Now the actual problem. When I try to connect from a tftp client, I get a transfer timed out error. Here is the transcript of a session:

Code:

$ tftp 192.168.0.1
tftp> status
Connected to 192.168.0.1.
Mode: netascii Verbose: off Tracing: off Literal: off
Rexmt-interval: 5 seconds, Max-timeout: 25 seconds
tftp> get testfile
Transfer timed out.

tftp> status
Connected to 192.168.0.1.
Mode: netascii Verbose: off Tracing: off Literal: off
Rexmt-interval: 5 seconds, Max-timeout: 25 seconds
tftp> quit

I really don't know what to try next. I googled around but no real solution has been given to similar problems. Thank you.

colucix · 06-25-2009, 05:43 AM

Update: I did some progress!

Googling around I found a post from someone that suggested to try with CentOS instead of Fedora. It was right! I put the tftp server on a CentOS machine in the same LAN and I simply used the Fedora server as DHCP. To do this I had to edit the line

Code:

next-server 192.168.0.4;

in /etc/dhcpd.conf to point to the tftp installation. Now it works, but I encountered another problem: after loading initrd and the kernel image, I get a kernel panic - not syncing error.

I have to mention that the system I'm trying to install is Windows XP (forgive me... this is a request from my boss) and maybe the kernel image is not compatible with the pxelinux boot. I will further investigate this issue. Eventually I will try to install Linux to see if it works. Any suggestion is really appreciated. Thank you.

colucix · 06-25-2009, 08:56 AM

Update 2: some step ahead!

Now I'm trying to boot CentOS 5.3 via PXE installation. The kernel and initrd.img are loaded, but the installation hangs at some point. Here is a transcript of the last lines I see on the netbook screen:

Code:

<omitted>
SElinux: Initializing.
selinux_register_security: Registering secondary module capability
Capability LSM initialized as secondary
Mount-cache hash table entries: 512
CPU: L1 I cache: 32k, L1 D cache: 32k
CPU: L2 cache:1024k
Intel machine check architecture supported.
Intel machine check reporting enabled on CPU#0.
Checking 'hlt' instruction... OK.
SMP alternatives: switching to UP code
Freeing SMP alternatives: 14k freed
CPU0: Intel(R) Pentium(R) M processor 1000Mhz stepping 05
SMP motherboard not detected.
Local APIC not detected. Using dummy APIC emulation.
Brought up 1 CPUs

At this point it hangs indefinitely. I don't hear any disk activity both on the netbook and on the tftp server. They just rest in silence.

Here is my pxelinux.cfg/default:

Code:

prompt 1
default linux
timeout 100

label linux
kernel vmlinuz
append initrd=initrd.img ramdisk_size=9216 noapic acpi=off

What's now? Thank you.

colucix · 06-26-2009, 09:14 AM

Update 3: I give up!

After some other attempt I decided to run memtest86 via pxe boot. The memtest86 started successfully, but it gave a lot of errors and after one minute it ended up with Unexpected Interrupt - Halting CPU0. Then follows the image of the stack. I suspect that the previous error was due to the same problem causing the computer to hang and halt the CPU.

At this point I have no other clue and I will tell to my boss to throw this old notebook away. Or at least to try to change the RAM. Any advice?