Re: Fwd: Re: Installed kppp despite VoIP and played with

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Ok Philippe I won't argue.
I only considered the typical case DSL + analog or broadband modem.
In the case of what you and Wolfgang have, just the same as I have except that now the modem/router is a Netgear at 10.0.0.138, where I used DSL on my PC and in addition intermittently (at hospital) a broadband cellular modem with wvdial, I had nothing to do except setting the router to answer DHCP requests and start the router as a DHCP client using a WAN dynamic DNS. The broadband worked just the same (wvdial, ppp0). I used the etwork manager to set ON the interface to be used and OFF the other. On the same router I added a second PC (DSL and analog modem) programmed just the same) without any surprise. I "talked" ethto eth and eth to ppp0 in all combinations between the two PC's without changing any setting besides the ON/OFF network manager button to select the active interface. I did NOT route any eth or ppp to ppp or eth in any machine ro in the pair.
Jacques

On 03/06/2012 04:56 PM, Philippe Vouters wrote:
Dear Jacques,

The 192.168.0.1 nameserver in Wolfgang's configuration should point to his DSL router DNS server/Gateway. I have the exact same configuration using my SFR provider and my SFR provided DSL modem/router box. The ppp issues are totally disconnected from the DSL/Ethernet aspect in Wolfgang's configuration. This is because Wolfgang does not route traffic between his ppp connection and his DSL connection. So when one link is up the other should become useless.

Ludwig has a simple DSL modem (no router) relaying modem information to his computer. Hence Ludwig has configured everything with DHCP whereas Wolfgang may have a fixed or dynamic Ethernet assigned IP address. I am even not sure whether Ludwig still uses a dial-up connection or has switched to DSL. Anyhow he is quite interested in having a correctly configured DSL connection.

Wolfgang's and Ludwig's two problems are only related because they are both running openSUSE and have troubles correctly configuring openSUSE so that www.google.com translates to a useable IP address whichever the connection is used. Ludwig and Wolfgang problems completely differ in the fact they do not use the same openSUSE version. Hence the proposed solutions to each differ.

In the hope this clarifies the issues in your mind.
With my warmest regards to you.
Philippe

Le 06/03/2012 15:18, Jacques Goldberg a écrit :
Gentlemen

I believe that the mess of files has become unmanageable.

Does somebody know what in Wolfgang's SuSE machine sets a DNS of
192.168.0.1 ?

This address pattern 192.168.*.* along with 10.0.*.* are reserved by IP
standards for private local networks which can never been seen on the
world wide Internet. This is of course the reason why you always see one
of these two patterns on the local side of a router, or even a modem
when using just a modem without router.

Now on such a network most routers and unmarried modems are capable of
returning to a loal PC a pseudo-DNS address such as 192.168.0.1 meaning
the router/modem itself (I choosed 192.168 at random , it could have
been 10.0 for the other local network pattern). When the router/modem
connects to the world Internet, in most hardware models it CAN (but is
never forced to),if so programmed, acquire via DHCP a world wide IP
address for itself, and one or more DNS world wide addresses.

This is by far the best and simplest way to set up the local network.
Needless to say this is what I use at home, with four successive faster
and faster modems+routers. I even started more than 10 years ago with an
Alcatel allegedly modem-only where one bit in one byte of ROM had to be
overwritten to make it see the built-in router! At this time I use
Fedora 16 with (sometimes) concurrent Ethernet, WiFi, and cellular
broadband modem, the latter using the setup scheme and the tools, wvdial
included, of an analog modem connection.

I think that Wolfgang now has a crowd of different methods
simultaneously activated with softwares incoherently trying to use them.
I definitely would suggest a major clean up, a careful look at the
capabilities of his modem/router (by slowly reading at least twice the
manual), and a slow, documented at all steps, attempt to follow the
approach which I have reported.

The usual caveat, SuSE (and not only SuSE, the method is due to
Microsoft), is necessary: by definition users are assumed to be
illiterate idiots,we manufacturers know better than them what they need,
and we manufacturers consider of no value to tell them what we do and how.

Jacques

On 03/06/2012 01:35 PM, Philippe Vouters wrote:



-------- Message original --------
Sujet: Re: Installed kppp despite VoIP and played with
Date : Tue, 06 Mar 2012 11:48:21 +0100
De : Philippe Vouters <philippe.vouters@xxxxxxxxxxx>
Pour : Wolfgang Junker <junker@xxxxxxxxxxxxxxxxxxx>
Copie à : Jacques Goldberg <Jacques.Goldberg@xxxxxxx>, Ludwig Gebauer
<ludwiggebauer@xxxxxxx>, Antonio Olivares <olivares14031@xxxxxxxxx>

Dear Wolfgang,

Got the part of the DNS issue analyzing your kppp.log. Here is the
interesting sequence:

10362 open("/etc/resolv.conf.n1pCw8",
O_WRONLY|O_CREAT|O_APPEND|O_LARGEFILE, 0666) = 5
10362 fcntl64(1, F_GETFD) = 0
10362 fcntl64(1, F_DUPFD, 10) = 10
10362 fcntl64(1, F_GETFD) = 0
10362 fcntl64(10, F_SETFD, FD_CLOEXEC) = 0
10362 dup2(5, 1) = 1
10362 close(5) = 0
10362 write(1, "nameserver 195.50.140.248\n", 26) = 26
10362 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
10362 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
10362 write(1, "nameserver 145.253.2.203\n", 25) = 25
10362 rt_sigprocmask(SIG_BLOCK, [CHLD], [], 8) = 0
10362 rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
10362 write(1, "nameserver 192.168.0.1\n", 23) = 23

This means the file which is actually opened is:
/etc/resolv.conf.n1pCw8 which is a temporary file created by
/bin/mkstemp and in which all the ppp negociated DNS IP addresses are
stored as well as including your 192.168.0.1 nameserver.

Then looking after /etc/resolv.conf.n1pCw8 it executes the following
command:
10405 execve("/bin/mv", ["mv", "-f", "/etc/resolv.conf.n1pCw8",
"/etc/resolv.conf.netconfig"], ["PPPD_PID=10340", "PPPLOGNAME=root",
"TERM=raw", "SPEED=57600", "IPREMOTE=145.253.1.100", "IFNAME=ppp0",
"DNS2=145.253.2.203", "DNS1=195.50.140.248", "PWD=/", "SHLVL=3",
"DEVICE=/dev/536ep", "USEPEERDNS=1", "ORIG_UID=0",
"IPLOCAL=212.144.133.153", "_=/bin/mv"]) = 0
So $ /bin/mv -f /etc/resolv.conf.n1pCw8 /etc/resolv.conf.netconfig

Now for your kppp DNS issue have a look to
http://seife.kernalert.de/blog/2008/12/11/using-dialup-with-111-if-networkmanager-does-not-handle-your-device/

This information fully matches the problem and brings in a solution.
With my analysis of your kppp.log above, my Google query has naturally
been "OpenSUSE /etc/resolv.conf.netconfig kppp".

Your truly,
Philippe

Le 05/03/2012 22:32, Wolfgang Junker a écrit :
Philippe,

to put kppprc into /root/.kde4/share/config was really the right point!
Having done this and trying $kppp as su it still crashes, but with the
hint to first set

export $(dbus-launch).

So I did this (whatever it means...), and kppp runs under su *and* under
user! It connects to my internet provider, but I don't get any
throughput, ping says

Network is unreachable

and ssh says

Could not resolve hostname

so this seems to be the old DNS-problem. I did again a
strace -v -e trace=all -f -o kppp.log kppp (this time as su)
and append the kppp.log (it is rather long, sorry for the spam to all
others...).

I can't try efax as a user, because it sits in my root-directory... And
here the rights of my /dev/536ep0:

crw------- 1 root root 240, 1 5. Mär 20:05 536ep0

Wolfgang



On Mon, 5 Mar 2012, Philippe Vouters wrote:

Dear Wolfgang,

Now when executed as root, kppp attempts to open
/root/.kde4/share/config/kppprc and does not find it.
6841 stat64("/root/.kde4/share/config/kppprc", 0x8141d1c) = -1 ENOENT
(No such file or directory)
6841 lstat64("/root/.kde4/share/config/kppprc", 0xbfdfc170) = -1
ENOENT (No such file or directory)

You should really do the the following:

$ sudo mkdir /root/.kde4
$ sudo mkdir /root/.kde4/share
$ sudo mkdir /root/.kde4/share/config
$ sudo cp /home/junker/.kde4/share/config/kppprc
/root/.kde4/share/config/kppprc
Then from a root account:
# kppp
or from a user account:
$ sudo kppp
or identically from a user account:
$ su -c 'kppp'

Yours truly,
Philippe

Le 05/03/2012 20:34, Wolfgang Junker a écrit :
Dear Philippe,

this is right, I am doing kppp as user. For this I have changed the
rights of /usr/bin/kppp to

-rwsr-xr-x 1 root root 554352 22. Feb 2011 kppp

(This is how it worked for SUSE 10.3). If I do kppp as root, the
program
crashes rightaway. The "Permission denied" must somehow be an artefact
of strace, I didn't get these messages before, and, as I said, I did
already get an internet connection with kppp (up to the DNS-issue)
with
precisely the present configuration. E.g. the rights of /dev/modem are
as follows:

lrwxrwxrwx 1 root root 10 5. Mär 20:05 modem -> /dev/536ep
crw-rw-rw- 1 root dialout 240, 1 5. Mär 20:05 536ep

I tried now the $strace -v -e trace=all -f -o kppp.log kppp
as root, but, as I said, the program crashes. I append the
kppp.log-file
which was created.

Wolfgang




On Mon, 5 Mar 2012, Philippe Vouters wrote:

Dear Wolfgang,

As far as it seems you start kppp from your user account and not from
a root account. This explains why you get all these
/etc/ppp/chap-secrets* "Permission denied".

You have this:

[philippe@victor ~]$ grep /etc/ppp/chap-secrets kppp.log
12745 open("/etc/ppp/chap-secrets.new",
O_WRONLY|O_CREAT|O_TRUNC|O_LARGEFILE, 0666) = -1 EACCES (Permission
denied)
12745 unlink("/etc/ppp/chap-secrets.old") = -1 ENOENT (No such
file or
directory)
12745 rename("/etc/ppp/chap-secrets", "/etc/ppp/chap-secrets.old") =
-1 EACCES (Permission denied)
12745 rename("/etc/ppp/chap-secrets.new", "/etc/ppp/chap-secrets") =
-1 ENOENT (No such file or directory)
12745 access("/etc/ppp/chap-secrets.old", F_OK) = -1 ENOENT (No such
file or directory)

Next you have:

12745 open("/dev/modem", O_RDWR|O_NOCTTY|O_NONBLOCK|O_LARGEFILE) = -1
EACCES (Permission denied)
12745 write(2, "error opening modem device !\n", 29) = 29

Looks very strange if efax from a user account is able to open
/dev/modem ! Check for /dev/modem permissions. You might have change
the /dev/modem permissions for hylafax operations.

On my Fedora 16 computer, I have:

[philippe@victor ~]$ ls /usr/bin/kppp
/usr/bin/kppp
[philippe@victor ~]$ ls -l /usr/bin/kppp
lrwxrwxrwx. 1 root root 13 Mar 5 04:29 /usr/bin/kppp -> consolehelper
[philippe@victor ~]$ ls -l /usr/sbin/kppp
-rwxr-xr-x. 1 root root 749004 Dec 3 07:28 /usr/sbin/kppp

If, from my user account, I:
$ strace -v -e trace=all -f -o kppp.vouters.log /usr/sbin/kppp
I see on my screen first that I am not allowed to run pppd unless it
the SUID bit (Capture 19:22:22.png)
Next a dialog to configure kppp (Capture 19:17:21.png).

In conclusion and under Fedora, the usual way is to call
/usr/bin/kppp
which activate a dialog prompting for root password (consolehelper
role), next activating /usr/sbin/kppp from a root context.

Your problem comes from you do not activate kppp from a root context.
Hence all the permission denied.

Kind regards,
Philippe

Le 05/03/2012 18:47, Wolfgang Junker a écrit :
Dear Philippe,

my configuration file for kppp is .kde4/share/config/kppprc and
looks as
follows:

pppdArguments=

[Account0]
AccountingEnabled=0
AccountingFile=
Authentication=4
AutoDNS=1
AutoName=0
BeforeConnect=
BeforeDisconnect=
CallbackPhone=
CallbackType=0
Command=
DNS=145.253.2.11
DefaultRoute=0
DisconnectCommand=
Domain=
ExDNSDisabled=0
Gateway=0.0.0.0
IPAddr=0.0.0.0
Name=Arcor
Password=internet
Phonenumber=01920787
ScriptArguments=
ScriptCommands=
StorePassword=1
SubnetMask=0.0.0.0
Username=arcor
VolumeAccountingEnabled=0
pppdArguments=debug

[General]
DefaultAccount=Arcor
DefaultModem=Modem
NumberOfAccounts=1
NumberOfModems=1
PPPDebug=0
ShowLogWindow=1

[Graph]
Background=255,255,255
Enabled=true
InBytes=0,0,255
OutBytes=255,0,0
Text=0,0,0

[Modem0]
AnswerResponse=CONNECT
AnswerString=ATA
BusyResponse=BUSY
BusyWait=0
ConnectResponse=CONNECT
DLPResponse=DIGITAL LINE DETECTED
Device=/dev/modem
DialString=ATDT
Enter=CR/LF
EscapeGuardTime=50
EscapeResponse=OK
EscapeString=+++
FlowControl=Hardware [CRTSCTS]
HangUpResponse=OK
HangupString=+++ATH
InitDelay=50
InitResponse=OK
InitString=ATZ
InitString1=
Name=Modem
NoCarrierResponse=NO CARRIER
NoDialToneDetection=ATX3
NoDialToneResp=NO DIALTONE
PreInitDelay=50
RingResponse=RING
Speed=57600
Timeout=60
ToneDuration=70
UseLockFile=0
Volume=2
VolumeHigh=M1L3
VolumeMedium=M1L1
VolumeOff=M0L0
WaitForDialTone=1

[Modem1]
AnswerResponse=
AnswerString=
BusyResponse=
BusyWait=
ConnectResponse=
DLPResponse=
Device=
DialString=
Enter=
EscapeGuardTime=
EscapeResponse=
EscapeString=
FlowControl=
HangUpResponse=
HangupString=
InitDelay=
InitResponse=
InitString=
InitString1=
Name=
DLPResponse=
Device=
DialString=
Enter=
EscapeGuardTime=
EscapeResponse=
EscapeString=
FlowControl=
HangUpResponse=
HangupString=
InitDelay=
InitResponse=
InitString=
InitString1=
Name=
NoCarrierResponse=
NoDialToneDetection=
NoDialToneResp=
PreInitDelay=
RingResponse=
Speed=
Timeout=
ToneDuration=
UseLockFile=
Volume=
VolumeHigh=
VolumeMedium=
VolumeOff=
WaitForDialTone=

[WindowPosition]
WindowPositionConWinX=247
WindowPositionConWinY=0
WindowPositionStatWinX=567
WindowPositionStatWinY=0
.......................................................................



I append the result of $strace -v -e trace=all -f -o kppp.log kppp
to this mail.

Sincerely
Wolfgang


On Mon, 5 Mar 2012, Philippe Vouters wrote:

Dear Jacques, Wolfgang,

I just downloaded the adequate RPM for Fedora 16 containing kppp.
From
a root account, I ran strace over kppp. Because there are numerous
open calls, I edited the file and searched for the "kppp" string.

First kppp attempts to open /root/.kde/share/config/kppprc. From
this
it knows what it must do, therefore directly proposing for me a
connection my old tiscali provider. Next as I previously specified /dev/modem in this configuration, it opens /dev/modem and tries to
connect. Because the modem is no longer connected to an analog
phone
line, kppp fails with a NO DIALTONE. So its sounds so far very
sane.

Yours truly,
Philippe

[root@victor ~]# cat /root/.kde/share/config/kppprc
[Account0]
AccountingEnabled=0
AccountingFile=
Authentication=4
AutoDNS=1
AutoName=0
BeforeConnect=
BeforeDisconnect=
CallbackPhone=
CallbackType=0
Command=
DNS=
DefaultRoute=1
DisconnectCommand=
Domain=
ExDNSDisabled=0
Gateway=0.0.0.0
IPAddr=0.0.0.0
Name=tiscali
Password=100556
Phonenumber=0860015555
ScriptArguments=
ScriptCommands=
StorePassword=1
SubnetMask=0.0.0.0
Username=pvou0030@xxxxxxxxxx
VolumeAccountingEnabled=0
pppdArguments=debug

[General]
DefaultAccount=tiscali
DefaultModem=536EP
NumberOfAccounts=1
NumberOfModems=1
PPPDebug=0
QuitOnDisconnect=1
ShowLogWindow=0
pppdTimeout=60

[Graph]
Background=255,255,255
Enabled=true
InBytes=0,0,255
OutBytes=255,0,0
Text=0,0,0

[Modem0]
AnswerResponse=CONNECT
AnswerString=ATA
BusyResponse=BUSY
BusyWait=52
ConnectResponse=CONNECT
DLPResponse=DIGITAL LINE DETECTED
Device=/dev/modem
DialString=ATDT
Enter=CR/LF
EscapeGuardTime=255
EscapeResponse=OK
EscapeString=+++
FlowControl=Hardware [CRTSCTS]
HangUpResponse=OK
HangupString=+++ATH
InitDelay=73
InitResponse=OK
InitString=AT&F E0 Q0 X4 S0=0 &D2 &C1 &S0 V1 W2
InitString1=AT+GCI=3D
Name=536EP
NoCarrierResponse=NO CARRIER
NoDialToneDetection=ATX3
NoDialToneResp=NO DIALTONE
PreInitDelay=53
RingResponse=RING
Speed=115200
Timeout=12
ToneDuration=36
UseLockFile=0
Volume=2
VolumeHigh=M1L3
VolumeMedium=M1L1
VolumeOff=M0L0
WaitForDialTone=1

[WindowPosition]
WindowPositionConWinX=5
WindowPositionConWinY=24
WindowPositionStatWinX=279
WindowPositionStatWinY=354






[Index of Archives]     [Linux Media Development]     [Asterisk]     [DCCP]     [Netdev]     [X.org]     [Xfree86]     [Fedora Women]     [Linux USB]

  Powered by Linux