Petre,
Sorry, I haven't been paying much attention to the DIALD List Lately....
I have had DIALD99.4 Working on both RH6.1 and RH6.2, but I have
dropped it since getting my Cable Modem. But, lets get to your problems....
I see you are using the RH6.1 Distro, and thus firstly you need to
install some of the RH6.1 Bug Fixes. In particular the INITSCRIPTS
and PPPD Updates. There are a number of Network Related
Fixes (in particular a DEFAULT ROUTE Problem in PPPD) that
have been included.
With RH6.1, and above you will have to add the "noauth"
option to your /etc/ppp/options file. This is required due
to changes with PPPD in that the default configuration is
"auth". This setting has to do with the requirement of
External Systems having to authorize themself with your
System.
Additionally, if your ISP uses CHAP or PAP for
authentication, you will have to modify your
/etc/ppp/chap-secrets and /etc/ppp/pap-secrets
files. Here is the entry format:
#client server secret IP address
username * password *
Also, if you are using PAP or CHAP you will need
a "pppd-options" directive in your diald config
file.
pppd-options name username
With respect to IP Configuration Settings, you will need
to enable IP Forwarding and IP Dynamic Address Hacking
in the Kernel. This is done by setting to Run-Time Flags in
the Kernel, "ip_forward" and "ip_dynaddr". The first flag
enables IP Forwarding, while the second enables correct
Dynamic IP Address Replacement on the first PACKET
sent (the first Packet has the IP Address of "local" before
you receive the Dynamic IP from your ISP). You set the
flags by:
echo "1" > /proc/sys/net/ipv4/ip_forward
echo "1" > /proc/sys/net/ipv4/ip_dynaddr
You can have these settings activated at BOOT Time via adding
these two lines to the "/etc/rc.d/rc.local" Boot Script.
Also, you will need to setup some IPCHAINS rules for Forwarding
and Masquerading your Internal Network, as well as, you will
need to load the various Masquerading Modules (FTP, etc...).
You will need to setup DNS Name Resolution to point to DNS
Name Servers on the Internet. You will have to modify the
/etc/resolv.conf file. You can do this manually, or via using
LINUXCONF.
With RH6.x, the default System install does not always create the
"/dev/modem" symbolic link, which points to your selected
COM Port. You can set the symbolic link via LINUXCONF, or
via use the "ln -s /dev/ttySx /dev/modem" command string (replace
"x" with the appropiate COM port #, i.e. 0 = COM1).
I see you are trying to utilize some special PPP and DIALD Scripts
for Activating and Controlling the Link. I would recommend you start
out simple, and merely use the Chat Script generated by LINUXCONF,
when configuring and PPP Device, for DIALD Connection.
I have to head to work, but I can help you later this evening...
Jeff
----- Original Message -----
From: Petre Rodan <[EMAIL PROTECTED]>
To: <[EMAIL PROTECTED]>; <[EMAIL PROTECTED]>
Sent: Wednesday, August 09, 2000 3:11 AM
Subject: it works !!! but there's a Big Bad Bug in diald.
>
> Ok folks. This quest is almost over. I've lost 4 days to find out that
> there was a tiny bug inside.
>
> Don't ask me why, but diald is unable to make both tap0 and then ppp0
> the default gateway.
>
> At compile i've changed config.h to read:
> #undef PATH_ROUTE "/sbin/route"
> and that's exactly where 'route' is.
>
> I've also attached the config which has the 'defaultroute' enabled.
>
> I'm just too lazy to see the diald code, but if there is someone else
> with this problem this is how I resolved it:
>
> look inside my diald starting script and notice the 'route ...' command
> inside
> it is possible that on slow machines the 'sleep' command should be
> increased ...
>
> You must also add these lines in a script that gets loaded after
> ppp-up.
>
> (the 'extip=...' is in a single line! ;)
>
> extip="`/sbin/ifconfig ppp0 | grep 'inet addr' | awk '{print $2}' | sed
> -e 's/.*://'`"
> route add default gw $extip
>
> This line should be added to some script after ip-down.
> route add default gw 192.168.1.10
>
> In every example 192.168.1.10 repesents the (fake) remote address from
> diald.conf
>
> Any comments?
>
> Bye
> Peter
----------------------------------------------------------------------------
----
> #!/bin/sh
> #
> # (C) Copyright 1995-1999 Mike Jagdis <[EMAIL PROTECTED]>
> #
> #
> # Configure the managed interfaces.
> #
> # This script would normally be installed as /etc/rc.ddiface,
> # /etc/init.d/ddiface, /etc/rc[1-5].d/SXXddiface or similar
> # depending on your system religion.
>
>
> # CONFIGURABLE STUFF
>
> export PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/bin
>
> config_dir=/etc/diald
>
> # END OF CONFIGURABLE STUFF
>
>
> action="$1"
> shift
> interfaces="$*"
>
>
> if [ -n "$BASH_VERSION" ]; then
> enable echo
> enable kill
> use_escapes='-e'
> else
> use_escapes=''
> fi
>
>
> case "${action}" in
> start)
> # Try and load compression modules before doing
> # anything that might load SLIP/PPP modules.
> modprobe bsd_comp ppp_deflate > /dev/null 2>&1
>
> echo $use_escapes "Configuring managed interfaces:\c"
>
> cd "$config_dir"
> for link in ${interfaces:-*}
> do
> if [ -d "$link" -a -r "$link/config" ]; then
> echo $use_escapes " $link\c"
> diald -f "$config_dir/$link/config" \
> pidfile "diald.$link"
>
> fi
> done
>
> sleep 1
> /sbin/route add default gw 192.168.1.10
>
> echo
> ;;
>
> stop)
> if [ -z "$interfaces" ]; then
> pids=`cat /var/run/diald.* 2> /dev/null`
> else
> # FIXME: If we take down all the interfaces using
> # a single, generic ipppd we ought to take that
> # down too. But how do we know?
> pids=`for link in $interfaces; do
> cat "/var/run/diald.$link"
> done 2> /dev/null`
> fi
>
> if [ -z "$pids" ]; then
> echo "No managed interfaces to stop"
> exit 0
> fi
>
> kill $pids > /dev/null 2>&1
> sleep 1
> for retry in 1 2 3; do
> if kill $pids > /dev/null 2>&1; then
> : some still exist
> sleep 2
> else
> echo "Managed interfaces stopped"
> break
> fi
> done
> ;;
>
> *) echo "usage: diald start|stop [link ...]"
> exit 1
> ;;
> esac
>
>
>
>
----------------------------------------------------------------------------
----
> debug 24
>
> linkname "fx"
> linkdesc "FX Internet"
>
> # Lock in use devices
> lock
>
> #device /dev/ttyS0
> device /dev/ttyS1
>
> modem
> crtscts
> speed 115200
>
> mode ppp
>
> authsimple "/etc/diald/fx/auth"
>
> accounting-log "/var/log/diald.fx"
> pidfile "/var/run/diald.fx"
> fifo "/var/adm/diald/fx"
> tcpport 10000
>
> # Use fake addresses initially. Once FX has told us our real
> # address we know it is static.
> local 192.168.1.5
> remote 192.168.1.10
>
> # Subnet mask for the wan link
> #netmask 255.255.255.0
> #netmask 0.0.0.0
>
> # The IP addresses will be asigned when connection starts.
> dynamic
>
> # If link goes down by remote end, start it again only if there is
> # outgoing packets.
> two-way
>
> # When link is up, route directly to the real ppp interface, not the proxy
> # interface. Not to do this is a performance lost of about 20 per cent.
> # There are old kernels that do not support reroute. See diald manual for
> # more information
> reroute
>
> #sticky
>
> #pppd-options noipdefault
>
> # We can reach the world via FX...
> defaultroute
>
> # More or less standard stuff.
> redial-timeout 30
> died-retry-count 0
> connect-timeout 600
> #retry-count 3
> #dial-fail-limit 10
> start-pppd-timeout 120
>
> pppd-options noauth user [EMAIL PROTECTED]
>
> connect "/usr/share/diald/connect"
> include "/usr/share/diald/filter/standard"
>
>
>
>
-
To unsubscribe from this list: send the line "unsubscribe linux-diald" in
the body of a message to [EMAIL PROTECTED]
