Home / ubuntu / Questions / 548826
Accepted
Jos
Asked: 2014-11-13 08:30:36 +0800 CST

Two hosts can't ping one another

  • 772

I have two hosts in my network. Both are connected to my router. From another host, I can ping and ssh into both of them. Both have an excellent connection to all other hosts. However, they cannot ping or see one another. One (192.168.124.101, called Hallway) runs Ubuntu 14.10, the other (192.168.124.104, called Rapunzel) is a Raspberry Pi with Raspbian. Both addresses are static. The hosts are connected to the electrical network through Powerline adapters.

Some output:

jos@hallway:~$ ifconfig eth1
eth1      Link encap:Ethernet  HWaddr 00:14:85:82:2a:2f
          inet addr:192.168.124.101  Bcast:192.168.124.255  Mask:255.255.255.0
          inet6 addr: fe80::214:85ff:fe82:2a2f/64 Scope:Link
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:2959658 errors:0 dropped:91 overruns:0 frame:0
          TX packets:2366877 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:3406445368 (3.4 GB)  TX bytes:964531766 (964.5 MB)
          Interrupt:21
jos@hallway:~$ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         Comtrend.Home   0.0.0.0         UG    0      0        0 eth1
192.168.124.0   *               255.255.255.0   U     0      0        0 eth1
jos@hallway:~$ arp -a 192.168.124.104
rapunzel (192.168.124.104) at b8:27:eb:a4:84:a3 [ether] PERM on eth1

and vice versa:

pi@Rapunzel ~ $ ifconfig eth0
eth0      Link encap:Ethernet  HWaddr b8:27:eb:a4:84:a3
          inet addr:192.168.124.104  Bcast:192.168.124.255  Mask:255.255.255.0
          UP BROADCAST RUNNING MULTICAST  MTU:1500  Metric:1
          RX packets:60981 errors:0 dropped:0 overruns:0 frame:0
          TX packets:40710 errors:0 dropped:0 overruns:0 carrier:0
          collisions:0 txqueuelen:1000
          RX bytes:19883333 (18.9 MiB)  TX bytes:2562029 (2.4 MiB)
pi@Rapunzel ~ $ route
Kernel IP routing table
Destination     Gateway         Genmask         Flags Metric Ref    Use Iface
default         Comtrend.Home   0.0.0.0         UG    0      0        0 eth0
192.168.124.0   *               255.255.255.0   U     0      0        0 eth0
pi@Rapunzel ~ $ arp -a 192.168.124.101
hallway (192.168.124.101) at 00:14:85:82:2a:2f [ether] on eth0

I have tried adding and deleting routes and arp entries. The current arp entries were added automatically, however (so there is some sort of contact). At first, a ping on either side would mention that the other side was "Destination host unreachable". The arp entries would give "<incomplete>" as the MAC address. After the arp entries were added, the pings simply time out.

It is probably not a firewall problem, for I have never used iptables on either of these machines (sudo iptables -L gives the default output on both hosts). I have rebooted both machines and the router, but so far nothing helps.
Things worked until yesterday, when I disconnected and reconnected one of the hosts. What am I overlooking?



Edit: I ran a tcpdump arp from host 101, and sure enough, I got numerous requests like the following:

23:12:23.460191 ARP, Request who-has hallway tell rapunzel, length 46
23:12:23.460214 ARP, Reply hallway is-at 00:14:85:82:2a:2f (oui Unknown), length 28

So why does no other traffic get through?

I ran an arping probe from both sides:

jos@hallway:~arping -b -D -w 2000 -I eth1 -s 192.168.124.101 192.168.124.104
ARPING 192.168.124.104 from 192.168.124.101 eth1
^CSent 38 probes (38 broadcast(s))
Received 0 response(s)

and something similar at the other end (only formatted differently, this is Raspbian after all).

networking

2 Answers

  1. Best Answer

    Probably a dumb question: Are these both directly connected into the router or are they going through a switch? If through a switch, turn off both hosts, turn off switch, turn off router. Turn on router, turn on switch, turn on hosts.

    (If you can give me more info on the exact brand of the switch, I might be able to tell you how to reset just the 3 ports of the switch)

    Leaving the above in as not everyone is using powerline adapters, but un-plug/replug the powerline adapters and/or update their firmware to "unstick" them… >:)

    • 4

  2. This turned out to be not as interesting as it first appeared. As Fabby mentioned, the powerline adapter must have got stuck. I connected one of the hosts to a different adapter and it worked straight away. Then I reconnected it to the first adapter, and again no traffic got through. While it would appear that the adapter has a hardware failure, keep in mind that it is only blocking traffic to and from one particular host. Thank you Fabby and Alaa Ali for your suggestions.

    • 2