Hint?

Hint?

Post by rapska » Sat, 21 Sep 2002 12:19:54



Whenever I try to do a traceroute to any external address, all of the probes
timout from the first.

I can ping external addresses no problem, but no external traceroutes can
get through.  Traceroutes to internal addresses are fine though, of course
they are only one or two hops :-)

I am pretty sure that traceroute uses UDP packets by default, but NTPDate
can get through just fine and that also is set to use UDP packets.

Does this originate with my router or with the firewall settings of my
system itself?  Do I need to open a certain port on my router for UDP
traffic to allow traceroute probes to come back?  Or perhaps modify ipchains
on this system?

Anyone have any clues or ideas as to what the issue is here?

TIA!

[followups set to c.o.l.a.]

--
rapskat -  11:10pm  up 2 days, 23:04,  4 users,  load average: 0.14, 0.25, 0.26
111 processes: 109 sleeping, 1 running, 1 zombie, 0 stopped
CPU states:  1.0% user,  0.7% system,  0.0% nice,  0.8% idle

In a museum in Havana, there are two skulls of Christopher Columbus,
"one when he was a boy and one when he was a man."
                -- Mark Twain

 
 
 

Hint?

Post by Kenneth Stephe » Sat, 21 Sep 2002 13:11:06



> Whenever I try to do a traceroute to any external address, all of the probes
> timout from the first.

> I can ping external addresses no problem, but no external traceroutes can
> get through.  Traceroutes to internal addresses are fine though, of course
> they are only one or two hops :-)

> I am pretty sure that traceroute uses UDP packets by default, but NTPDate
> can get through just fine and that also is set to use UDP packets.

> Does this originate with my router or with the firewall settings of my
> system itself?  Do I need to open a certain port on my router for UDP
> traffic to allow traceroute probes to come back?  Or perhaps modify ipchains
> on this system?

rapskat,

        traceroute does use UDP (the return packets are ICMP though). If the
timeouts are originating from the first hop, then that machine is very
likely restricting UDP ports. In order for traceroute to work, the UDP
packets are sent to a very high port number that typically is not used
by other applications. If the first hop is a firewall which doesnt allow
traffic for those ports to go through, traceroute wont work even though
other UDP applications (which may be explicitly permitted by that
machine) work.

Regards,
Kenneth

 
 
 

Hint?

Post by Ian Northeas » Sun, 22 Sep 2002 19:04:13



> Whenever I try to do a traceroute to any external address, all of the probes
> timout from the first.

> I can ping external addresses no problem, but no external traceroutes can
> get through.  Traceroutes to internal addresses are fine though, of course
> they are only one or two hops :-)

> I am pretty sure that traceroute uses UDP packets by default, but NTPDate
> can get through just fine and that also is set to use UDP packets.

> Does this originate with my router or with the firewall settings of my
> system itself?  Do I need to open a certain port on my router for UDP
> traffic to allow traceroute probes to come back?  Or perhaps modify ipchains
> on this system?

> Anyone have any clues or ideas as to what the issue is here?

Missed the strange follow-up, sorry, my reply went astray.

Traceroute uses UDP packets. By default it starts with port 33434 and
increments it by one each probe. As the default maximum hop count is 30
and the default number of probes per hop is three, you probably want to
open 90 ports from 33434 outbound. Note that the man page is misleading
on this point. You also need to ensure that the ICMP port unreachable
responses can get back in.

As you can traceroute to internal addresses but not external, I suspect
that the block is on your router, or you have a firewall rule which
allows ICMP port unreachables from your LAN but not the Internet.

Regards, Ian

 
 
 

Hint?

Post by rapska » Mon, 23 Sep 2002 00:04:17


Whenever I try to do a traceroute to any external address, all of the probes
timout from the first.

I can ping external addresses no problem, but no external traceroutes can
get through.  Traceroutes to internal addresses are fine though, of course
they are only one or two hops :-)

I am pretty sure that traceroute uses UDP packets by default, but NTPDate
can get through just fine and that also is set to use UDP packets.

Does this originate with my router or with the firewall settings of my
system itself?  Do I need to open a certain port on my router for UDP
traffic to allow traceroute probes to come back?  Or perhaps modify ipchains
on this system?

Anyone have any clues or ideas as to what the issue is here?

TIA!

[followups set to c.o.l.a.]

--
rapskat -  11:10pm  up 2 days, 23:04,  4 users,  load average: 0.14, 0.25, 0.26
111 processes: 109 sleeping, 1 running, 1 zombie, 0 stopped
CPU states:  1.0% user,  0.7% system,  0.0% nice,  0.8% idle

In a museum in Havana, there are two skulls of Christopher Columbus,
"one when he was a boy and one when he was a man."
                -- Mark Twain

 
 
 

Hint?

Post by Paul Lutu » Mon, 23 Sep 2002 00:49:35



> Whenever I try to do a traceroute to any external address, all of the
> probes timout from the first.

> I can ping external addresses no problem, but no external traceroutes can
> get through.  Traceroutes to internal addresses are fine though, of course
> they are only one or two hops :-)

> I am pretty sure that traceroute uses UDP packets by default, but NTPDate
> can get through just fine and that also is set to use UDP packets.

> Does this originate with my router or with the firewall settings of my
> system itself?

Easy way to find out -- disable the firewall temporarily.

Quote:> Do I need to open a certain port on my router for UDP
> traffic to allow traceroute probes to come back?

Try "ping"-ing the site of interest. That should tell you if traceroute has
a chance to succeed.

A generic approach is to set your system up in a very normal,
non-specialized way, confirm that traceroute works, than add the more
restrictive elements of your present configuration one by one until the
problem arises.

--
Paul Lutus
www.arachnoid.com

 
 
 

Hint?

Post by Linonu » Mon, 23 Sep 2002 01:24:19


After takin' a swig o' grog, rapskat belched out this bit o' wisdom:

Quote:> Whenever I try to do a traceroute to any external address, all of the probes
> timout from the first.

> I can ping external addresses no problem, but no external traceroutes can
> get through.  Traceroutes to internal addresses are fine though, of course
> they are only one or two hops :-)

> I am pretty sure that traceroute uses UDP packets by default, but NTPDate
> can get through just fine and that also is set to use UDP packets.

What happens with traceroute -I ?
 
 
 

Hint?

Post by Tokr » Mon, 23 Sep 2002 01:25:02



> Whenever I try to do a traceroute to any external address, all of the probes
> timout from the first.
> I can ping external addresses no problem, but no external traceroutes can
> get through.  Traceroutes to internal addresses are fine though, of course
> they are only one or two hops :-)
> I am pretty sure that traceroute uses UDP packets by default, but NTPDate
> can get through just fine and that also is set to use UDP packets.
> Does this originate with my router or with the firewall settings of my
> system itself?  Do I need to open a certain port on my router for UDP
> traffic to allow traceroute probes to come back?  Or perhaps modify ipchains
> on this system?
> Anyone have any clues or ideas as to what the issue is here?

There are documented problems with some Cable/DSL routers and
doing traceroutes, some of which are solved by updating firmware.
No matter what firmware version I installed in my Linksys BEFSR41
router, I have not been able to do any UDP traceroutes (outside
the LAN) even if the machine doing them is in a DMZ.

To do traceroutes, I now use "MTR", which combines the
functionality of ping and traceroute into a single program.

MTR -- <http://www.bitwizard.nl/mtr>

--
Tokra

 
 
 

Hint?

Post by rapska » Mon, 23 Sep 2002 01:39:04


Error Log for Thu, 19 Sep 2002 23:19:35 +0000: segfault in module "rapskat"
- dump details are as follows...

Quote:> Whenever I try to do a traceroute to any external address, all of the
> probes timout from the first.

<snip>

Hey, how'd this get reposted?

Please ignore.  Situation resolved.

--
rapskat -  12:38pm  up 4 days, 12:32,  3 users,  load average: 6.83, 5.35, 2.96
128 processes: 116 sleeping, 12 running, 0 zombie, 0 stopped
CPU states:  0.1% user,  0.9% system,  0.0% nice,  1.0% idle

hubub, hubub, HUBUB, hubub, hubub, hubub, HUBUB, hubub, hubub, hubub.

 
 
 

Hint?

Post by rapska » Mon, 23 Sep 2002 02:05:00


Error Log for Sat, 21 Sep 2002 16:25:02 +0000: segfault in module "Tokra" -
dump details are as follows...

Quote:> There are documented problems with some Cable/DSL routers and doing
> traceroutes, some of which are solved by updating firmware. No matter what
> firmware version I installed in my Linksys BEFSR41 router, I have not been
> able to do any UDP traceroutes (outside the LAN) even if the machine doing
> them is in a DMZ.

Hmmm....that's the same model that I have.  That could be it.

Quote:> To do traceroutes, I now use "MTR", which combines the functionality of
> ping and traceroute into a single program.

> MTR -- <http://www.bitwizard.nl/mtr>

Good tip, and nice proggie.  Thanks!

--
rapskat -   1:04pm  up 4 days, 12:58,  3 users,  load average: 0.64, 0.54, 1.21
113 processes: 110 sleeping, 3 running, 0 zombie, 0 stopped
CPU states:  0.2% user,  1.0% system,  0.0% nice,  0.2% idle

The more things change, the more they stay insane.

 
 
 

1. ipppd hangs my machine. Any hints?

Hello everybody!

I've been working with two Alphas with great pleasure until these last days.
While trying to connect to my ISP with an ISDN card, my computer hangs
completely, and I have to reboot it.

Hardware:
Alphaserver 4/233
32 Mb RAM
4 Gb SCSCI disk
13 Gb IDE disk
3com 905 NIC
TELES PCI (detected as NetJet)

Software:
Debian (woody)
kernel 2.2.17
ipppd version i2.2 patchlevel 12

And that's what syslog is telling:

Dec 11 06:45:47 Alphaserver ipppd[198]: ipppd i2.2.12 (isdn4linux version of
pppd by MH) started
.....
(ipppd goes on ok)
.....
Dec 11 06:45:51 Alphaserver ipppd[198]: local  IP address 193.153.209.23
Dec 11 06:45:51 Alphaserver ipppd[198]: remote IP address 193.152.55.161
Dec 11 06:46:30 Alphaserver kernel: APECS machine check:
Dec 11 06:46:30 Alphaserver kernel:   vector=0x660 la_ptr=0xfffffc00002084c0
Dec 11 06:46:30 Alphaserver kernel:   pc=0xfffffc00003103e0 size=0x1a0
procoffset=0x110 sysoffset 0x1a0
Dec 11 06:46:30 Alphaserver kernel:   expected 0 DCSR 0x0 PEAR 0x0
Dec 11 06:46:30 Alphaserver kernel:  +0 1a0 1a000000110
Dec 11 06:46:30 Alphaserver kernel:  +10 100000204 fffffc01c01101c0
Dec 11 06:46:30 Alphaserver kernel:  +20 6f800000000 fffffc00000aa850
......
(lots and lots of numbers)
......
Dec 11 06:46:30 Alphaserver kernel:  +180 0 204830
Dec 11 06:46:30 Alphaserver kernel:  +190 fffffc0000201628 cca0000332902850

And the machine dies ...

Can anybody help me, please? I have no idea of what that means.

Many thanks

Pablo Diaz

2. Headless multia boot problem

3. on Hint on why "...connection was refused..."?

4. standard group file?

5. BETA-RedHat 5.9/Starbuck: Install Hints

6. where I can find doc. for COMPOSE key?

7. What's the device.hints file that v5.0 is looking for?

8. A question for you all!

9. porting from SysVr3 to SysVr4 - hints and tips needed

10. Looking for a UDP server. Any hints?

11. Need hints on setup for dockable notebook

12. Any hints for Radeon 32mb ddr setup

13. PPP problems, please give me a hint