can ping/ssh into machine, but machine can't reach out.

can ping/ssh into machine, but machine can't reach out.

Post by To » Wed, 07 Nov 2001 09:01:10



The machine was working fine until all of a sudden. I can ping and ssh
the RH7 Linux machine, but this RH7 machine can't reach out. So
basically, all incoming connections work, but no outgoing connections
work.

The problematic machine uses dhcp, which seems more or less correct. I
first try pinging a hostname, but it can't find the hostname. So I
figure I have nameserver problems. But I can't ping the hostname's IP
either. I boot into windows on the same machine and it performs
properly. I copy down all the network info and compare it to the linux
side. Everything is correct (default gateway, nameservers, etc.) I try
to specify everything as static IP, but it still can't reach out.

ifconfig seems more or less correct, except for the line that reads
NOTRAILERS RUNNING on eth0. On an exact functioning twin of the
machine, it is RUNNING MULTICAST. I can't figure out what that means.

I have disabled the firewall but it still doesn't help. This is a very
important machine to have internet access, so it's frustrating that
the network portion is half working.

Any suggestions would be helpful, even shots in the dark. I'm
desparate.

 
 
 

can ping/ssh into machine, but machine can't reach out.

Post by Dean Thompso » Wed, 07 Nov 2001 21:24:25


Hi!,

Quote:> The machine was working fine until all of a sudden. I can ping and ssh
> the RH7 Linux machine, but this RH7 machine can't reach out. So
> basically, all incoming connections work, but no outgoing connections
> work.

> The problematic machine uses dhcp, which seems more or less correct. I
> first try pinging a hostname, but it can't find the hostname. So I
> figure I have nameserver problems. But I can't ping the hostname's IP
> either. I boot into windows on the same machine and it performs
> properly. I copy down all the network info and compare it to the linux
> side. Everything is correct (default gateway, nameservers, etc.) I try
> to specify everything as static IP, but it still can't reach out.

> ifconfig seems more or less correct, except for the line that reads
> NOTRAILERS RUNNING on eth0. On an exact functioning twin of the
> machine, it is RUNNING MULTICAST. I can't figure out what that means.

Okay, the first thing to check is that it isn't a DHCP renew/lease problem.
The system isn't getting a new IP address after a certain period of time ?.
If you do a /sbin/route -n, what is the output, does it look like you have a
default route to get off your network.  Finally, when you do the ifconfig
eth0, are you seeing your network card listed there with a correct IP address,
subnet mask and broadcast address ?

See ya

Dean Thompson

--
+____________________________+____________________________________________+

| Bach. Computing (Hons)     | ICQ     - 45191180                         |
| PhD Student                | Office  - <Off-Campus>                     |
| School Comp.Sci & Soft.Eng | Phone   - +61 3 9903 2787 (Gen. Office)    |
| MONASH (Caulfield Campus)  | Fax     - +61 3 9903 1077                  |
| Melbourne, Australia       |                                            |
+----------------------------+--------------------------------------------+

 
 
 

can ping/ssh into machine, but machine can't reach out.

Post by To » Sat, 10 Nov 2001 03:57:42



> Okay, the first thing to check is that it isn't a DHCP renew/lease problem.
> The system isn't getting a new IP address after a certain period of time ?.
> If you do a /sbin/route -n, what is the output, does it look like you have a
> default route to get off your network.  Finally, when you do the ifconfig
> eth0, are you seeing your network card listed there with a correct IP address,
> subnet mask and broadcast address ?

Thanks for replying,

I'm not in front of the machine now, but this is what I know:
The system has no problems getting it's IP when it boots up (set to
dhcp). But when I run ifdown/ifup, it complains about Operation
failed. However, eth0 is visible when I run ifconfig and it does have
the appropriate IP address, subnet mask, broadcast. Only difference is
"NOTRAILERS RUNNING" which I haven't been able to figure out...

All information was left unchanged since its working state, so default
route, etc, etc...is all correct. I even booted to windows, got all
the info, and assigned everything statically. Still, nothing.

I'll see if anything has changed since the problem...once again, any
off-the-wall guesses at the problem will help.

Tom

 
 
 

can ping/ssh into machine, but machine can't reach out.

Post by Karl Heye » Sat, 10 Nov 2001 04:43:01





>> Okay, the first thing to check is that it isn't a DHCP renew/lease
>> problem. The system isn't getting a new IP address after a certain
>> period of time ?. If you do a /sbin/route -n, what is the output,
>> does it look like you have a default route to get off your network.
>>  Finally, when you do the ifconfig eth0, are you seeing your
>> network card listed there with a correct IP address, subnet mask
>> and broadcast address ?

> Thanks for replying,

> I'm not in front of the machine now, but this is what I know: The
> system has no problems getting it's IP when it boots up (set to
> dhcp). But when I run ifdown/ifup, it complains about Operation
> failed. However, eth0 is visible when I run ifconfig and it does
> have the appropriate IP address, subnet mask, broadcast. Only
> difference is "NOTRAILERS RUNNING" which I haven't been able to
> figure out...

If you did static then thats not the problem. What is the ifconfig
counts like. does tcpdump show traffic being sent out.

karl.

 
 
 

1. Canned "ping"

    Within a ksh script, I would like to first determine whether or not
a remote
host is alive before I attempt other network access to it, (file reads
or writes).
The ping command seems to be the logical solution, but if I have to wait
for
the ping command to fail due to network time-out to indicate that the
remote
host is not available I've gained nothing over just letting my other
network
access commands time-out.

    Is there a way to execute ping wrapped in a timer - such as in a
subprocess
that gets killed by it's parent if it doesn't end within a practical
time interval?
It seems to me that that should work, but what I'm really looking for is

someone who would share code that has already been written.

    Much appreciated.

2. A very special illustration for all the peoples of COLA

3. It's not bad canned meat...

4. Linux & 128 M RAM

5. Different login message for different users

6. Help: machines on subnet can't ping the machines outside subnet

7. 1.3.88 kernel, FDC_S82078B, and Buslogic 445C

8. Help - can't ping one of my machines (but it can ping me).

9. Ping: I can't ping machine names!

10. Canned Proxy URLs to Filter

11. CGI: apache canned response