timed out waiting for a valid DHCP server response

Hi,

Booting up RH w/ 2.4.18-10smp kernel for the first time in a long
time, in a new and different network than the one to which this box
used to be connected - and in which DHCP worked just fine.
Specifically, I'm on a LAN that's connected to a ZyXEL Prestige 600
Series DSL Modem.

/var/log/syslog reveals: "dhcpcp[pid]: timed out waiting for a valid
DHCP server response"

DHCP works in Windoze, so the hardware's fine.

What could be wrong? and how can I try to fix it?...

NB - I have found two people who've described having this problem - or
a problem or problems very much like it - and then its apparently
resolving itself without any change to their configuration.*

I get the basic idea of DHCP, but "the basic idea's" not quite enough
to help me understand what's going on for these people - and what may
well be going on for me - much less does it suggest anything I can
*do* about all this(!)

Finally, I'm no conspiracy theorist, but I just ran across the
following thread - http://www.knoppix.net/forum/viewtopic.php?t=2164,
under the heading "Evidence WinXP 'security updates' may break Linux"
- which, as convoluted and inconclusive as it may be, prompts me to
wonder whether Windoze may have done something (to the NIC's eeprom?)
to cause the problem.

Any thoughts, suggestions, requests for more info. will be most
welcome!

Tks,
Matthew

* C.f. http://www.linuxquestions.org/questi...002/05/4/21462,
and http://groups.google.com/groups?hl=e...com%26rnum%3D1
(if that URL works - it's awful long.)

On 26 Sep 2003 10:14:07 -0700, Matthew Weymar wrote:
> Hi,
>
> Booting up RH w/ 2.4.18-10smp kernel for the first time in a long
> time, in a new and different network than the one to which this box
> used to be connected - and in which DHCP worked just fine.

maybe
/etc/sysconfig/network
/etc/sysconfig/network-scripts/ifcfg-eth0
need changes or your dhcp client is using a stale file which needs to
be deleted.

Matthew Weymar wrote:

> Hi,
>
> Booting up RH w/ 2.4.18-10smp kernel for the first time in a long
> time, in a new and different network than the one to which this box
> used to be connected - and in which DHCP worked just fine.
> Specifically, I'm on a LAN that's connected to a ZyXEL Prestige 600
> Series DSL Modem.
>
> /var/log/syslog reveals: "dhcpcp[pid]: timed out waiting for a valid
> DHCP server response"
>
> DHCP works in Windoze, so the hardware's fine.

Inconclusive, at best...

> What could be wrong?

Log files ? Run dhcpcd manually and use the -d flag.

> and how can I try to fix it?...

Run "pump" - does that get you an IP ?

Try increasing the timeout in dhcpcd.conf - since that's the error, you
might have started there.
(Though I doubt it - the dhcp timeout is pretty long as it is - up to
downright annoying in Windoze...)

> NB - I have found two people who've described having this problem - or
> a problem or problems very much like it - and then its apparently
> resolving itself without any change to their configuration.*

Which does not help any, does it ?

> I get the basic idea of DHCP, but "the basic idea's" not quite enough
> to help me understand what's going on for these people - and what may
> well be going on for me - much less does it suggest anything I can
> *do* about all this(!)

Read the documentation for dhcpcd, and the docs for the modem/dchp server.

Does it say, specifically, how it expects the clients to behave ?

Windows apparently working tells you less than nothing - 50% of the
protocols as implemented by Micro$haft are not RFC-compliant or just plain
broken.

<snipped way-over-72-char-line>

> (if that URL works - it's awful long.)

Yes - the only part needed to get to the article is the "threadm=" ID.
Leave off the rest.

--
Jeroen Geilman

All your bits are belong to us.

Bit Twister <(E-Mail Removed)> wrote in message news:<(E-Mail Removed)>...
> On 26 Sep 2003 10:14:07 -0700, Matthew Weymar wrote:
>
>>Hi,
>>
>>Booting up RH w/ 2.4.18-10smp kernel for the first time in a long
>>time, in a new and different network than the one to which this box
>>used to be connected - and in which DHCP worked just fine.
>
>
> maybe
> /etc/sysconfig/network

NETWORKING=yes
HOSTNAME=localhost.localdomain

> /etc/sysconfig/network-scripts/ifcfg-eth0

DEVICE=eth0
BOOTPROTO=dhcp
ONBOOT=yes

These seem pretty innocuous, do they not?...

> need changes or your dhcp client is using a stale file which needs to
> be deleted.

Per 'man dhcpcd', I add "local0.* /var/log/dhcpcd.log" to
/etc/syslog.conf, reboot (same problem) and check the dhcpcd.log file,
where the only entry refers to the timing out issue. I then run
'dhcpcd
-d". I get:

dhcpcd: MAC address = xx:xx:xx:xx:xx:xx

and then eventually a prompt. Sadly, eth0 is still without an IP
address.

Checking /var/log/dhcpcd.log, however, reveals:

localhost dhcpcd[pid]: broadcasting DHCP_REQUEST for 192.168.x.x
localhost dhcpcd[pid]: timed out waiting for DHCP_ACK response
localhost dhcpcd[pid]: broadcasting DHCP_DISCOVER
localhost dhcpcd[pid]: timed out waiting for a valid DHCP server
response

Interestingly my router's IP is: 192.168.y.1(!), so it seems that
you're
right that there's something "stale" in there. "But what? where?" I am
asking myself....

After *much* poking around, I have come across a post by Brian Cluff
on
another list in which he mentions deleting the
/etc/dhcpc/dhcpcd-eth0.info file, and "getting an IP without a hitch."
Opening this file up and seeing my "bad" 192.168.x.x IP filled me with
optimism - that deleting this file would put me in business,
obviously.
Alas, 'tis not the case.

After deleting /etc/dhcpc/dhcpcd-eth0.info, /var/logdhcpcd.log
(following 'dhcpcd -d') indicates that I'm still "broadcasting
DHCP_REQUEST for 192.168.x.x."

I figure dhcpcd is picking this up from /etc/dhcpc/dhcpcd-eth0.cache,
and so I delete that, run 'dhcpcd -d' again and find that
/var/logdhcpcd.log now includes only the latter two of the four lines
above:

localhost dhcpcd[pid]: broadcasting DHCP_DISCOVER
localhost dhcpcd[pid]: timed out waiting for a valid DHCP server
response

I try creating a new /etc/dhcpc/dhcpcd-eth0.info, substituting
192.168.y.x for 192.168.x.x, but I get these same two lines (above),
and
only these two lines again.

Now I search the net on "new /etc/dhcpc/dhcpcd-eth0.info" and see that
perhaps I should have run "ifdown eth0" and then "ifup eth0". I delete
/etc/dhcpc/dhcpcd-eth0.info again, and run ifdown & ifup. Both
operations fail.

I find a LinuxPlanet.net tutorial on "xDSL and Linux: Go Speed Racer!"
Being a fan, I try the script they suggest (see below). This does
nothing for me. Still no IP address.

Pausing to take a breath... hoping for more suggestions....

Addendum: In light of Jeroen Geilman's reply to my original post, I
try 'pump -i eth0', and find that it does not work.

I think that one of my problems is that since DHCP fails on startup,
NIC
is not properly recognized (because it doesn't have an IP?...) I don't
know.

Tks,
Matthew

From http://linuxplanet.com/linuxplanet/tutorials/1454/4/:

#!/bin/sh
#
# Name: renet
# Author: Nicholas Donovan <(E-Mail Removed))
# Function: resets the IP by bumping a request from DHCPCD
#
# Revisions:

# remove the old information
rm /etc/dhcpc/dhcpcd-eth0.cache -f
rm /etc/dhcpc/dhcpcd-eth0.info -f

# kill the runtime PID reference
rm /var/run/dhcpcd-eth0.pid -f

# request a new IP from server...
echo "making new DHCP request..."
# I added the '-d' argument - unnecessarily??? -- MDW
/sbin/dhcpcd -d
sleep 3

# if changed properly tell it then pipe to log file...
echo "(dhcpcd) IP address changed to $1" | logger

# Display the new IP
clear
echo "Your new DHCP address is:"
/sbin/ifconfig eth0

# End of Script

Simply chmod 744 this script and place it into the /sbin directory.