Take the 2-minute tour ×
Server Fault is a question and answer site for system and network administrators. It's 100% free, no registration required.

I am handling 34 Debian servers. Two of these are diskful machines (which boot from their HDs), the remaining 32 are PXE clients without disks.

The two machines with disks are called uni01 and uni02. uni01 sends a single system image via NFS to the nodes that do network booting (it also runs DHCP + TFTP).

The network is a class A subnet (mask 255.255.252.0). Problem is, one specific node (uni19) cannot contact uni02. They're in the same subnet, the other hosts can connect just fine, except this one.

root@uni19:~# traceroute uni02
traceroute to uni02 (10.248.212.216), 30 hops max, 60 byte packets
1  uni19 (10.248.212.255)  2997.791 ms !H  2997.789 ms !H  2997.784 ms !H

However arping works for uni02:

root@uni02:/home/einar# arping uni19
ARPING 10.248.212.255
60 bytes from 10:1f:74:2c:78:aa (10.248.212.255): index=0 time=172.946 usec
60 bytes from 10:1f:74:2c:78:aa (10.248.212.255): index=1 time=208.391 usec

It doesn't for uni19:

root@uni19:~# arping uni02
ARPING 10.248.212.216
^C
--- 10.248.212.216 statistics ---
4 packets transmitted, 0 packets received, 100% unanswered (0 extra)

Routing tables:

root@uni19:~# route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         uni01           0.0.0.0         UG    0      0        0 eth0
10.248.212.0    *               255.255.252.0   U     0      0        0 eth0


 root@uni02:~# route
 Kernel IP routing table
 Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
 default         uni01           0.0.0.0         UG    0      0        0 eth0.682
 10.248.212.0    *               255.255.252.0   U     0      0        0 eth0.682
 10.248.216.0    10.248.212.1    255.255.255.0   UG    0      0        0 eth0.682
 10.248.217.0    10.248.212.1    255.255.255.0   UG    0      0        0 eth0.682

uni01 and uni02 use specific interfaces for a VLAN, but all the other hosts have been switched to the same VLAN to be able to communicate (in fact, also uni01 is on the same VLAN, but all serers are able to PXE boot).

What can I do to troubleshoot this? I already confirmed there are no firewalls up.

share|improve this question
    
have you already tried tcpdump? –  dawud Jul 11 '13 at 12:28
    
That's what I'm doing right now. Will edit the post (or add an answer) once I'm done. –  Einar Jul 11 '13 at 12:33

1 Answer 1

up vote 0 down vote accepted

PEBKAC. That's the only way I could describe this.

The reason why this happened: a typo in /etc/network/interfaces assigned the wrong broadcast address to the interface in uni02, address that by chance was the one used by uni19. So I replaced the broadcast address and all is fine.

share|improve this answer

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.