Bud wrote:
> michales wrote:
> >
> > Ok, I have been difficulties with dial-up but it was my fault. I was without DHCP (prefer manually configuration) If memory serves right: my problem was the resolv.conf file I have mentioned above.
> >
> > Modem connects to Internet, right?
>
> Oh yes, I can go to my ISP's web site but no where else on the WWW.
>

Get the ip address of another website from your other machine with the
host command, for example,

host google.com

Then do:

wget <ip address>

If this sucessfully downloads an index.html file to your current
directory then you have a DNS problem.

Put the following lines in your /etc/resolv.conf

nameserver 4.2.2.1
nameserver 4.2.2.2
nameserver 4.2.2.3

And you should be OK. These are open DNS servers.

If wget doesn't download an index.html file, you have a routing
problem. You're on the same subnet as your ISP because they gave you an
IP address but you can't route to the Internet.

Do a 'route -n' to see if you have a default gateway.

You can manually add one with,

route add default gw <ip address>

Get the correct gateway from the machine that works with 'route -n'.

Old Man wrote:
>
> Are the slack system and the mandrake system actually different machines?
> Are you getting the same ip address, broadcast address and gateway on Slack
> and Mandrake? Do you get the same addresses every time you connect with
> Slackware? On Slackware, this information is in /etc/dhcpc/*.info.
> Mandrake same place?

I'm only using one box, just different HDs. Every time I connect via
dialup I get a different local and remote address.

Thanks for answering.

--

Nietzsche wrote:
>
> Get the ip address of another website from your other machine with the
> host command, for example,
>
> host google.com
>
> Then do:
>
> wget <ip address>
>
> If this sucessfully downloads an index.html file to your current
> directory then you have a DNS problem.

<snip>

As I said to Old Man, I have one box, but two HDs and even have Ubuntu
installed as well, but just to try it out and will give the disk space to
Slack, I hope, when I find the problem, with lots of help here I'm hoping
someone has had a similar WTF in the past.
--

Bud wrote:
> michales wrote:
>> Ok, I have been difficulties with dial-up but it was my fault. I was without DHCP (prefer manually configuration) If memory serves right: my problem was the resolv.conf file I have mentioned above.
>>
>> Modem connects to Internet, right?
>
> Oh yes, I can go to my ISP's web site but no where else on the WWW.
>

So... you have tried to edit the /etc/resolv.conf file?

check your ISP's nameserver with mandrake and write this on paper, after that add this in /etc/resolv.conf

search YourISP.name
nameserver The.Ip.You.Have.Find.From.Mandrake

useally the nameserver are two, in case that one fails, so you can write two lines.

michales wrote:
>
> So... you have tried to edit the /etc/resolv.conf file?
>
> check your ISP's nameserver with mandrake and write this on paper, after
>that add this in /etc/resolv.conf
>
> search YourISP.name
> nameserver The.Ip.You.Have.Find.From.Mandrake
>
> useally the nameserver are two, in case that one fails, so you can write two lines.

Michales, it changes with each dialup, it is a temp file, it's the ones that
my ISP gives me each time I dial in--hmmm modem just droped out. I looked as
you suggested and then looked again and they were the same numbers with
a different heading than what is in my resolve.conf in Slack. You may have
told me something I should look into Michales. I'll let you know how it
goes.
--

On 08/03/06 13:07, Bud wrote:
> And no joy, again!

I'm sorry.

> Yes, modem works fine, connects fine, used both Kppp and ppp-go with stable
> connections with an IP address. With Slack I can access the ISP website but
> no others.

Really? This tells me that the PPP link *IS* working and that there is
something else going on. In a word (or phrase) "Path MTU". I'd be willing
to bet that something about the PPP link under Slack. is a bit different
than it is under the other distro(s). Have you compiled a different kernel?

> No firewall, no windoze for a year or so.

Boo to the no firewall and a pat on the back for no windows as they let
light in.

> Thanks Grant, was hoping to get time to do some exploring of the scrips
> today. Makes me want to know a lot more than I do but . . . ignorance is not
> always bliss.

If you want some help, email me off news group.



Grant. . . .

On Fri, 04 Aug 2006 20:43:15 -0500, "Taylor, Grant" <gtaylor@riverviewtech.net> wrote:

>On 08/03/06 13:07, Bud wrote:
....
>> Yes, modem works fine, connects fine, used both Kppp and ppp-go with stable
>> connections with an IP address. With Slack I can access the ISP website but
>> no others.
>
>Really? This tells me that the PPP link *IS* working and that there is
>something else going on. In a word (or phrase) "Path MTU". I'd be willing
>to bet that something about the PPP link under Slack. is a bit different
>than it is under the other distro(s). Have you compiled a different kernel?

I use PPPoE which is similar to PPP dialup.

Excerpts from my rc.firewall, I do MSS clamping in firewall rather
than the PPP .conf:
....
# Check you have rp-pppoe.conf 'CLAMPMSS=no'

# Leave MAX_MSS empty to perform path mtu discovery, this doesn't work with
# ISPs that drop ICMP traffic -- refer `man iptables`, also make sure you
# have 'mtu path discovery' turned on below if you choose this option.

#MAX_MSS="1452" # maximum MSS (ref: cisco)
#MAX_MSS="1360" # minimum MSS (ref: cisco)
#MAX_MSS="1412" # default MSS (ref: rp-pppoe default)
MAX_MSS="1380"
#MAX_MSS="" # use PMTU discovery
....
# if ISP drops ICMPs, cannot perform mtu discovery
echo 1 > /proc/sys/net/ipv4/ip_no_pmtu_disc
....
# clamp MTU for new TCP connections to world
if [ -n "$MAX_MSS" ]; then # use preset
iptables -A OUTPUT -p tcp --tcp-flags SYN,RST SYN \
-o $X_WORLD -j TCPMSS --set-mss $MAX_MSS
else # use path discovery
iptables -A OUTPUT -p tcp --tcp-flags SYN,RST SYN \
-o $X_WORLD -j TCPMSS --clamp-mss-to-pmtu
fi

I hard code a fixed value for MSS 'cos dodo ISP improperly filters ICMPs,
path discovery unreliable (MTU is MSS + 40).

Other possibility is nameservers, my ip-up for reference:
#!/bin/bash
#
# /etc/ppp/ip-up
#
# Copyright (C) 2004,2005 Grant Coady
#
# environment vars:
# DEVICE
# IFNAME
# IPLOCAL
# IPREMOTE
# PEERNAME
# SPEED
# ORIG_UID
# PPPLOGNAME
# CONNECT_TIME
# BYTES_SENT
# BYTES_RCVD
# LINKNAME
# DNS1
# DNS2

log="/var/log/rp-pppoe"

echo -e "$(date "+%F_%T\t%s")\t\c" >> $log
echo -e "up\t$IFNAME\t$IPLOCAL\t$IPREMOTE" >> $log

# bring up firewall
/etc/rc.d/rc.firewall restart $IFNAME $IPLOCAL > /var/log/ip-up
#
# kick ntp
/usr/sbin/ntpd
# end

You could make sure $DNS1 and $DNS2 are being written to the resolver
file, by writing them to a log file from /etc/ppp/ip-up, then checking
resolver setup in one of two places:

Normal location, but I'm using dnsmasq on this box:
~$ cat /etc/resolv.conf
domain mire.mine.nu
nameserver 127.0.0.1

Alternate location, written on successful PPPoE connection:
~$ cat /etc/ppp/resolv.conf
nameserver 203.194.56.150
nameserver 203.194.27.57

I think 'normal' behaviour is to update /etc/ppp/resolv.conf and symlink
/etc/resolv.conf to it on successful connection, YMMV

Grant.
--
<http://bugsplatter.mine.nu/>

michales wrote:
>
> So... you have tried to edit the /etc/resolv.conf file?
>
> check your ISP's nameserver with mandrake and write this on paper,
>after that add this in /etc/resolv.conf
>
> search YourISP.name
> nameserver The.Ip.You.Have.Find.From.Mandrake
>
> useally the nameserver are two, in case that one fails, so you can
>write two lines.

I cleaned out my /etc/resolve.conf. Now back on Slack, sheesh, how did it
happen in the first place? I now see:
domain icehouse.net #kppp temp entry
nameserver 216.229.160.10 #kppp temp entry
nameserver 216.229.168.10 #kppp temp entry

Working fine, access the WWW etc.

Guess I owe you one Michaels.
--
Expect the worst, it's the best thing you can do.