Centos 7 PXE Boot failing

[Matti]

Hello everyone,

I'm trying to setup a PXE boot server with TFTP. (PXE + TFTP on 1 server whilst another server is providing the DHCP-IP)
I'm able to receive my IP from the boot server but then I'm resulting in a tftp connection timeout.
https://i.imgur.com/Xaw551I.png

Does anyone have experience in this kind of problem?

Thanks in advance,
Matti

default cfg tftp:
DEFAULT menu.c32
PROMPT 0
TIMEOUT 50

MENU title ########## PXE Boot Menu ##########
LABEL 1
MENU DEFAULT
MENU LABEL ^1) Fedora 26
KERNEL Fedora26/vmlinuz
APPEND initrd=Fedora26/initrd.img root=nfs:172.16.128.9:/srv/tftp/pxelinux/Fedora26/rootfs
LABEL 2
MENU LABEL ^2) Centos 7 x68_64
KERNEL centos7_x86_64/vmlinuz
APPEND initrd=centos7_x86_64/initrd.img root=nfs:172.16.128.9:/srv/tftp/pxelinux/centos7_x86_64/rootfs

LABEL 3
MENU LABEL ^3) Boot from local drive
LOCALBOOT

dhcp config:
# ISC DHCPD configuration -- don't edit manually!
#
# Ansible managed

#
# Global options
#
next-server 172.16.128.9;
filename "pxelinux.0";
default-lease-time 28800;
max-lease-time 43200;
option subnet-mask 255.255.0.0;
option routers 172.16.128.1;
option domain-name "green.local";
option domain-name-servers 172.16.128.133, 172.16.128.134;

#
# Includes
#

#
# Subnet declarations
#
subnet 172.16.0.0 netmask 255.255.0.0 {
range 172.16.0.100 172.16.127.254;
filename "pxelinux.0";
allow bootp;
allow booting;
}
host client2 {
hardware ethernet 00:11:22:33:44:02;
fixed-address 172.16.0.2;
}
host client3 {
hardware ethernet 00:11:22:33:44:03;
fixed-address 172.16.0.3;
}
host client4 {
hardware ethernet 00:11:22:33:44:04;
fixed-address 172.16.0.4;
}
host client5 {
hardware ethernet 00:11:22:33:44:05;
fixed-address 172.16.0.5;
}
host client6 {
hardware ethernet 00:11:22:33:44:06;
fixed-address 172.16.0.6;
}

[TrevorH]

Have you opened the correct tftp ports in the firewall? UDP 69 iirc.

[Matti]

Since it's produced in a local environment I disabled firewall

[TrevorH]

Well the symptoms you posted are either it's blocked by firewall, it's not got a running tftp daemon to talk to or you've pointed it to a wrong ip address.

[Matti]

I've been troubleshooting it for the last couple of days, do you mind checking it out through teamviewer or something else?

Sincerely,
Matti

[TrevorH]

Sorry, I don't do that sort of thing and I would definitely not recommend opening things up to random people you find on the internet.

[Matti]

Kinda out of options tho

[TrevorH]

Use tcpdump on your tftp server to see if packets arrive from the client. See if responses go back. Crank up the loglevel on the tftp daemon and read the logs.

[gerrito]

Hi Matti,

I've recently put a lot of hours in a multi PXE boot server ( and still not finished for all of my distributions ).
One of the things i learned ( and still dont know ) why there isnt one way to start everything via a ftp, nfs, or iso service.
Maybe there is......but i am not experience enough to tell u.

i only can share my experience with centos7. Centos is one of the distribution that worked pretty well on my PXE server with a kick start as well.
I had the same problem with nfs share, instead i used ftp share and it worked well.

default menu.c32
prompt 0
timeout 30
MENU TITLE PXE Menu
LABEL centos7_x64
MENU LABEL CentOS 7_X64
KERNEL /networkboot/vmlinuz
APPEND initrd=/networkboot/initrd.img inst.repo=ftp://172.168.x.xx/pub ks=ftp://172.168.x.xx/pub/centos7.cfg

If it is a Virtual Machine, maybe this would help :

- Base memory of the vm ( make it +1500MB )
- Select the boot order ( network )

Hope it will help