ping is not working in a small testing network

[f.mayer]

Hej!

I have set up a small network for testing with 3 computers:
PC1 (WinXP) <---> server (RedHat) <---> PC2 (WinXP)

On the server runs RedHat and I am using there nistnet for network
analyses. None of the computers are connected to the internet and thus
I turned off firewalls on all computers.

When I ping from PC1 to the server and vice versa everything is ok. But
when I ping from PC2 to the server it loses around 50% of the packets
(even though I have shut down nistnet). When I ping from PC1 to PC2
it's the same (packet loss of ~50% and more sometimes).

I double checked all settings, but it should work (and in fact is has
worked once).

Any idea, where I could / should search for the error? Why drops the
ping command packets - any reason for that?

Thanks in advance,
Franz

 

[Dan]

Have you considered that it could be a hardware issue on your PC2 (NIC
card or cable)?

 

[DanS]

(e-mail address removed) wrote in @i39g2000cwa.googlegroups.com:

Hej!

I have set up a small network for testing with 3 computers:
PC1 (WinXP) <---> server (RedHat) <---> PC2 (WinXP)

On the server runs RedHat and I am using there nistnet for network
analyses. None of the computers are connected to the internet and thus
I turned off firewalls on all computers.

When I ping from PC1 to the server and vice versa everything is ok. But
when I ping from PC2 to the server it loses around 50% of the packets
(even though I have shut down nistnet). When I ping from PC1 to PC2
it's the same (packet loss of ~50% and more sometimes).

I double checked all settings, but it should work (and in fact is has
worked once).

Any idea, where I could / should search for the error? Why drops the
ping command packets - any reason for that?

Thanks in advance,
Franz

The diagram shows a milti-homed RedHat server, 1 NICS, each PC -> RH
connection should be a separate subnet.

PC1 192.168.1.2 -> RH1 192.168.1.1
PC2 192.168.100.2 -> RH2 192.168.100.1

Each PC should also have its default gateway set to the repsective RH IP
address.

Double check all the routing tables as well, looking for double/multiple
routes.

DanS

 

[f.mayer]

Dan, thanks for the hint! Apparently one NIC card is not working
properly; now the pinging between each PC{1,2} and server is working,
but not pinging from PC1 to PC2.

But I guess that's a Linux setting problem - unfortunately I've no clue
why it is still not working; so I asked in the Linux group
(comp.os.linux.networking)

Regards, Franz

 

[DanS]

(e-mail address removed) wrote in @i39g2000cwa.googlegroups.com:

Dan, thanks for the hint! Apparently one NIC card is not working
properly; now the pinging between each PC{1,2} and server is working,
but not pinging from PC1 to PC2.

But I guess that's a Linux setting problem - unfortunately I've no clue
why it is still not working; so I asked in the Linux group
(comp.os.linux.networking)

Regards, Franz

I unfortunately have as much experience with Linux as I do with being a
millionaire.

See, the default gateway setting is used when the PC doesn't know where
the other subnet is. It is NOT used when the PC knows where the
destination is. It would know where it is if 1) it is in the same subnet,
or 2) there is a static route manually entered into the PC's routing
table to tell it where to send it.

PC1 tries to ping PC2. It doesn't know where PC2 is because PC2 is in a
different subnet, so it forwards the request to the default GW, the Linux
box. The linux box SHOULD know where PC2 is, because it is a member of
each of their respective subnets, one on each NIC.

PC2 should return the ping to the Linux box because it doesn't know where
PC1 is, and the Linux box should send it back to PC1.

Did you try a tracert to see where it stops ?

DanS