Linux-Networking Digest #948, Volume #11 Tue, 20 Jul 99 05:13:40 EDT
Contents:
Re: routing problem ([EMAIL PROTECTED])
Re: DHCP for 2 addresses on the same ethernet card? (Rudolf Potucek)
Re: are 4 nics practical? (Beat Rupp)
samba and windows network - incompatible namespaces? (Ben Blish)
Re: netscape (Cris Johnson)
Re: Question regarding syslogd and syslog.conf (Sitaram Chamarty)
Re: routing problem (Frank Waarsenburg)
Re: ISP uses PAP, but not always?? (Paul Winkler)
Re: SuSE 6.1 Dialup, Help Needed... (root)
NE1000 Network card (D.T.)
Re: ISP uses PAP, but not always?? (Frank Waarsenburg)
Cannot register service: RPC ([EMAIL PROTECTED])
Re: How to set up ADSL in Linux? ("Chris")
Firewall setup problem (nico)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: Re: routing problem
Date: Tue, 20 Jul 1999 06:39:44 GMT
Hi, thanks for your reply!
IP Forwarding is enabled. It still don't work.
Stephan
In article <NCglWERz#[EMAIL PROTECTED]>,
"Bono" <[EMAIL PROTECTED]> wrote:
>
> >
> >The linux box is able to ping to everybody on the other subnet. But
my
> >winnt machine can only ping to machines on the same subnet. It seems
to
> >me it's a problem on the linux box.
> >How can I find the wrong/missing setup?
>
> IP Forwarding should be enabled.
>
> Bono
>
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: [EMAIL PROTECTED] (Rudolf Potucek)
Crossposted-To: comp.security.firewalls
Subject: Re: DHCP for 2 addresses on the same ethernet card?
Date: 20 Jul 1999 06:46:59 GMT
I seem to have been misunderstood, to a certain degree. I have the a gate
machine set up to get a DHCP address on eth0 from the ISP and have a
comletely separate network on eth1 behind the gate. The ISP is willing
to supply 2 dynamic IPs.
That being said, it ircs me I can't actually make that work without
getting a hub and another ethernet card just to splice the connection
outside my machine and be able to get the second IP.
Why do I need it? I don't really, unless I ever wanted to play on
battlenet at the same time as my roommate, because fixed port connection
routing cannot be masqueraded too well ... but once curiosity is aroused!
I admit I haven't done enough reading yet, but is it impossible to
get the 2nd IP because the DHCP protocol uses hwaddr or is it just
because the client won't do it?
Rudolf
Scott Marlowe ([EMAIL PROTECTED]) wrote:
: Rudolf Potucek wrote:
: > Hey!
: >
: > Has anyone managed to set up a DHCP client to get 2 dynamic IPs on the
: > same physical interface? My problem is, I have an ADSL modem linked
: > to a linux gate and a couple of machines behind the gate on 10.x.x.x. For
: > obvious reasons I don't want to get a hub and an additional network card
: > just to get the 2nd dynamic IP.
: >
: > I tried:
: >
: > > ifconfig eth0 up
: > > ifconfig eth0:0 up
: > > dhcpcd eth0
: > > dhcpcd eth0:0
: >
: > This works fine for the first copy of the DHCP client demon (eth0), but the
: > 2nd copy (eth0:0) just conks out with
: Why do you need both IPs to be assigned by DHCP? Is it possible to alias an
: interface already running DHCP and add a static IP? If not, then you'll have
: to add a second NIC and set some routes to your internal network versus
: external network.
--
------------------------------
From: Beat Rupp <[EMAIL PROTECTED]>
Subject: Re: are 4 nics practical?
Date: Tue, 20 Jul 1999 08:47:39 GMT
>>>>>>>>>>>>>>>>>> Urspr=FCngliche Nachricht <<<<<<<<<<<<<<<<<<
Am 19.07.99, 19:39:38, schrieb Greg Leblanc <[EMAIL PROTECTED]> zum=20
Thema Re: are 4 nics practical?:
> In article <7museg$6no$[EMAIL PROTECTED]>,
> "Beat Rupp" <[EMAIL PROTECTED]> wrote:
> > Ok, I have a home server with 2 nics (LAN, cable modem) and the whol=
e
> > network is BNC. Now I want Fast Ethernet and I don't wanna buy some
> > expensive dual speed hub (the thing is, that 2 parts of the network
> have to
> > stay 10mbits)
> >
> > Now I have one practical solution, which requires the least amount o=
f
> new
> > cabling: the current server gets two additional nics and will become=
> some
> > sort of router/hub. Now that this shouldn't be a problem on the
> software
> > side with Linux I suspect that this could decrease network speed
> drastically
> > on the hardware front. It's a Pentium 166, 64mb, Asus X-P55T2P4. The=
> nics
> > would be: 10/RJ45/ISA for cable modem, 2 x 10/BNC/ISA for the 10 mbi=
ts
> parts
> > of the network and finally one 100 PCI nic for the new, fast part.
> >
> > Is it possible to copy large amounts of data (let's say hundres of
> > megabytes) between these "subnets" without slowing everything down?
> >
> Sounds to me like you'll be beating up that server pretty badly. Are
> these high quality NICs with good onboard processors? I'm guessing=20
that
> you have one host-to-pci bridge, and one pci-to-isa bridge, so
> everything is running off of one data channel to the CPU. If these=20
are
> all high quality nics (not necessarily new, but definately not cheap
> ne2000 clones) then I don't think you'll have much trouble unless=20
these
> file transfers are going all the time. I've seen 4 port dual speed=20
hubs
> for about $80, so they're really not all that expensive. You might=20
look
> at that, depending on how much you're willing to spend.
> Greg
The ISA nics are SMC EtherCards 16C (or something, they have BNC,=20
RJ46, AUI; that's what the C stands for: combo. Maybe there is another=20
number in the Product name, I'm not sure but 16 is there.)
I have the experience that if I copy something on my network from SMC=20
to SMC (in winslow), the cpu is used for 100% and work is not smooth=20
anymore. In Linux I just can't get over 500kb/s but it's smooth to=20
work. Conclusion: I don't know whether the cards have a high quality=20
chip (although they cost about $80-90 USD).
What sounds interesting is a dualspeed hub for $80. Where did you see=20
this one? Here in Switzerland these things are expensive (the cheapest=20
I found was about $160!).
Thanks for helping me!
Beat
------------------------------
From: [EMAIL PROTECTED] (Ben Blish)
Subject: samba and windows network - incompatible namespaces?
Reply-To: [EMAIL PROTECTED]
Date: Tue, 20 Jul 1999 07:16:21 GMT
I'm having a severe problem with samba. I'll try to explain it
all; I apologize for the wordiness, but I'm hoping to supply enough
info here to get a solution back, assuming one actually exists. So:
I'm trying to make samba work on a new redhat 6 system with a network
that consists of 1 win98 pc, 1 win95 pc, and 1 win nt4/sp3 pc. And a
redhat 6 pc.
The network controller, more or less, is the NT4 machine.
There is no password on the network on any of the machines except
the Linux box. The log in on the linux box is root, and the password
is zxcvbn
The log in on all the Windows machines is "Ben Williams".
Now, I have the machines up to where they ping each other happily.
I can even see the Linux group (toast) on the 98 machine, where I am
trying to get the initial connection up. But when I try to access
the shared dir on the linux machine by clicking into the group
(toast) and then the share (weasle), I get a dialog that says I
must supply a password. Now, on the Linux machine, although I tried
to set it for no password on root and etc, I failed. I set the password
length to 0, but it still won't let me get rid of the paranoia. Ok,
so the password I use is "zxcvbn" which is easily typed and easily
remembered, though completely superfluous and annoying. :( That's the
ONLY password anywhere in the system; it's what I use when something
won't let me get away without a password. So that's the only password
I expect to encounter.
Win98's dialog says I'm trying to access \\WEASLE\IPC$
Well, maybe, but all *I* said was "weasle" (lower case, I might add).
Anyway, you can feed this dialog anything on this $00ff00 earth and it won't
give you the resource. I've tried making all manner of users on the Linux
box, tried the examples on the HOWTO site, tried the examples on the beta
book site at www.mcp.com (the redhat book, chapter 14) and nothing I do
seems to make it work. From little droppings along the wayside in the Redhat
book, it says I have to make an account on the Linux book for "the client"
on the other machine. Is that the "Ben Williams" login that the machines all
start with? Because if it is, the namespace of the Linux box seems to puke
because (I think, I'm guessing and I'm frustrated out of my mind after 15 hours
of screwing with this obstinate thing) there's a space in it... the linuxconf
thing won't allow you to change a password if you have a space in the name, for
one (which I take as a hint, though it could just be another of the many bugs
I've found in the redhat distribution).
Can anybody help me here? I can do anything at all to the Linux box. I can't
change the login or the blank password on the other three machines. The only
requirement is that it be able to move files between the win98 machine and the
Linux box, both ways - I get the impression that because the win95 (OSR1) uses
unencrypted passwords and the win98 and NT boxes use encrypted passwords, that
samba can't deal with them all - that's ok, as long as the 98 box can work.
Some other info: I've got the password level and username level both set to 24
to eliminate presumed password case problems. I can reboot one or both machines
and the problems do not change in character.
I'll watch linux.samba, alt.linux and comp.os.linux.networking for responses,
but I would VERY much appreciate it if those responses would also be emailed
to me at...
[EMAIL PROTECTED]
...because the news server here in the sticks where I live is just the pits,
I miss tons of messages because its down, or just has a lousy feed. Or both.
Thanks in advance for any help - I'm at my wits end here.
Ben
------------------------------
From: [EMAIL PROTECTED] (Cris Johnson)
Subject: Re: netscape
Date: 20 Jul 1999 07:18:15 GMT
Alan,
I'm not advocating operating using numerical IP addresses - it is the only
way I have succeeded in getting the Collabra part of Netscape to connect
to my news server. Navagator and my non-Netscape network tools all work
fine with DNS names under Linux, and I was hoping someone might have some
better insights than I as to running Collabra "by the book".
------------------------------
From: [EMAIL PROTECTED] (Sitaram Chamarty)
Crossposted-To: redhat.config,comp.os.linux.misc
Subject: Re: Question regarding syslogd and syslog.conf
Date: 19 Jul 1999 22:34:34 -0700
On 19 Jul 1999 14:31:07 GMT, rob <[EMAIL PROTECTED]> wrote:
>How do i specify daemons? Like i'd like to see all su messages.
All the gory details can be found in "man syslog.conf". The short
answer to your question, meanwhile, is
daemon.* /var/log/messages
This will get you all messages logged with the facility code of
"daemon". Note that a program that that you consider to be a
daemon (like sendmail) may not use this "facility" in its logging.
(sendmail for instance, may use the "mail" facility, since that
exists also). The documentation for any program designed to run
as a daemon will usually tell you very clearly what facility it
uses to log messages to syslog.
HTH. HAND.
------------------------------
From: Frank Waarsenburg <[EMAIL PROTECTED]>
Subject: Re: routing problem
Date: Tue, 20 Jul 1999 09:42:12 +0200
It depends of what PING responds. If it says
Network Unreachable
then it means you did not set the default gateway at your Win machines and
they don't know where to deliver the packets.
If it says
Destination host unreachable
it means that somewhere a router does not know where to go. In this case,
that would be your Linux box and is unlikely since it is on both subnets.
If it says
Request timed out
then the other end does not know how to reach you (typically because it has
no default gateway) or is dead.
Troubleshoot as follows:
1) Ping from WINNT on subnet 1 to the Linux interface on subnet 1. This
should work. If not, your Linux is not set up correctly (since the NIC is
working).
2) Ping from WINNT on subnet 1 to the Linux interface on subnet 2. Here,
the results may vary. If you get "Network unreachable", do a Route Add
Default Gateway <Linuxinterfaceonsubnet1> at your NT box. If you get
Destination host unreachable, then Linux does not route between the
subnets. If you get Request timed out; same. If you get a response, Linux
is routing fine.
3) Now, ping WINNT on subnet 2. If you solved any problems occurring from
step 2, you have 2 possible results: Request timed out. In this case, you
probably need to Route Add Default Gateway <Linuxinterfaceonsubnet2> to
your 2nd NT box (Or something similar. Anyway, change it in the NT's NIC IP
settings). Or: it works.
Frank
[EMAIL PROTECTED] wrote:
> Hi,
>
> My winnt box is connected to a linux box. The linux box is connected to
> a winnt server which has two network cards and acts as a router between
> two subnets.
>
> /---- subnet 1 -----\ /---- subnet 2 ---- ...
> WINNT <-> LINUX <-> WINNT <-> ...
>
> The linux box is able to ping to everybody on the other subnet. But my
> winnt machine can only ping to machines on the same subnet. It seems to
> me it's a problem on the linux box.
> How can I find the wrong/missing setup?
>
> Any help appreciated!
>
> Thanks in advance!
> Stephan
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.
------------------------------
From: Paul Winkler <[EMAIL PROTECTED]>
Subject: Re: ISP uses PAP, but not always??
Date: Tue, 20 Jul 1999 02:52:56 -0400
Clifford Kite wrote:
> One of the ^Ms that follow 38400 is from the CONNECT "" chat expect/send
> and this is very likely what confuses the ISP into presenting a menu
> rather than skipping the menu and starting PPP instead. The ^Ms are
> carriage returns, one from the CONNECT "" and one from the ISP as a
> spacer to begin the menu.
(snip)
> A solution is to not send a carriage return. You can do this by replacing
> CONNECT "" with CONNECT \d\c . The new CONNECT expect/send inserts a
> one-second delay and doesn't send a carriage return. The \d shouldn't
> be necessary but sometimes it seems to help. The ~ '' expect/send
> probably isn't necessary but it shouldn't hurt either.
Drat. That sounded like a really good explanation. Sadly, it doesn't
seem to help... I edited my chatscript as per your suggestions and found
no difference, I'm still getting annoyingly frequent "annex command line
blah blah blah." (It took me seven attempts to get in here to reply!)
Any other ideas?
Thanks anyway,
PW
------------------------------
From: [EMAIL PROTECTED] (root)
Subject: Re: SuSE 6.1 Dialup, Help Needed...
Date: 20 Jul 1999 07:49:02 GMT
In article <[EMAIL PROTECTED]>,
peter van den Heuvel <[EMAIL PROTECTED]> writes:
> Iby wrote:
>>
>> I've installed SuSE 6.1 five days ago and have problems with ppp.
>> When I use minicom to dial, it works, dials, nagotiate connection, plays
>> little tune and says "CONNECTED 57600". Great...
>> Then it trys to send user/password informations and it takes waaayy toooooo
>> loooong till finnaly fails. Sometimes I can even see "Starting PPP" but
>> thats it, and some crazy characters fillup the screen.
>> Ok., let's try kppp,wvdial or whatever... But they can't even recognize
>> modem.
>> (It is ISA PnP 56k V.90 ZOOM faxmodem - not Winmodem!).
>> Back to xterm, run minicom, only initialize modem, than run kppp (or other
>> dialer)...
>> Now they work and set up speaker volume, dial, nagotiate communications...
>> GREAT...
>> but wait... oh no! Failure again....
>> Ok. I'm not quiter!
>> Boot NT and go online, see what IRC offers...
>> Nothing.... Only bunch of guys who have heard about something
>> called "fdisk" and "partition"...
>> Anyway, let's surf and investigate, internet is so huge...
>> There must be something....
>> Check ZOOM site..
>> Try Altavista...
>> Try Yahoo...
>> ......
>> ....
>> ..
>> .
>> This is about what I was doing lest few days and now I must admit that
>> I am frustrated...
>> Can anyone help me? I have tried to modify scripts, I've tried a lot
>> differnt things like changing strings in kppp with what I've taken from
>> *.INF or what ever I could find online that was for ZOOM modems, and now
>> I've no clue where to look anymore...
>> Please help, help help....
>>
>
If it's any help I had many problems like yours and finally solved them
(pretty much by accident to be honest) by installing kernel 2.0.36
instead of 2.2.5 - both supplied with suse 6.1. suse reckon this is a
hardware problem but I haven't yet been able to track it down. At least
everything works (well, not vmware because it needs 2.2.x ...)
Paul Douglas
------------------------------
From: [EMAIL PROTECTED] (D.T.)
Crossposted-To: alt.os.linux,alt.os.linux.redhat,linux.redhat.install
Subject: NE1000 Network card
Date: Tue, 20 Jul 1999 07:55:45 GMT
I have loaded Redhat Linux 5.2 onto my old 486dx4 100 successfully
except for the network connection.
I have a NE1000 compatible card installed with works OK under Netware
Lite. When I was asked for the type of network card installed in the
machine there was no option for NE1000 cards only NE2000.
Is there any instructions/ Howtos on how to install a NE1000 card in
Redhat 5.2 or do I have to obtain a more up to date card.
The card has successfully been running on the machine when Windows 95
was installed. The computer was recently retired from Win 95 and
replaced with a Celeron 400 so I am trying setup a Linux machine to
network with 2 other Win 95 machines
thanks for any advice or instructions
Dennis T
------------------------------
From: Frank Waarsenburg <[EMAIL PROTECTED]>
Subject: Re: ISP uses PAP, but not always??
Date: Tue, 20 Jul 1999 10:01:28 +0200
I'm still convinced that this causes the problem. Either the CR's, or there is
too much time between the actual connect and the LCP negotiation. You could try
the 'silent' or 'quiet' options. They cause pppd to send a REQ only once, or not
at all, waiting for the ISP to start negotiation. I had a similar problem and it
turned out to be a timing issue..
Frank
Paul Winkler wrote:
> Clifford Kite wrote:
> > One of the ^Ms that follow 38400 is from the CONNECT "" chat expect/send
> > and this is very likely what confuses the ISP into presenting a menu
> > rather than skipping the menu and starting PPP instead. The ^Ms are
> > carriage returns, one from the CONNECT "" and one from the ISP as a
> > spacer to begin the menu.
> (snip)
> > A solution is to not send a carriage return. You can do this by replacing
> > CONNECT "" with CONNECT \d\c . The new CONNECT expect/send inserts a
> > one-second delay and doesn't send a carriage return. The \d shouldn't
> > be necessary but sometimes it seems to help. The ~ '' expect/send
> > probably isn't necessary but it shouldn't hurt either.
>
> Drat. That sounded like a really good explanation. Sadly, it doesn't
> seem to help... I edited my chatscript as per your suggestions and found
> no difference, I'm still getting annoyingly frequent "annex command line
> blah blah blah." (It took me seven attempts to get in here to reply!)
>
> Any other ideas?
>
> Thanks anyway,
>
> PW
------------------------------
From: [EMAIL PROTECTED]
Subject: Cannot register service: RPC
Date: Tue, 20 Jul 1999 09:27:22 GMT
Reply-To: [EMAIL PROTECTED]
My other linux system behaves strange. It is a stand-alone, dial-up networked
system. After booting, when the prompt for the first logon is already displayed,
it tells in the same line:
Cannot register service: RPC: Unable to receive;
errno = Connection refused
As far as I understand, RPC is something NFS related. I did not change anything,
as far as I know. How does this happen ? After the message, normal logon is
possible.
The system: p120, SuSE 6.0, Kernel 2.2.10, Teles ISDN card.
=====================================================
Answers please in this newsgroup!
[EMAIL PROTECTED]
=====================================================
------------------------------
From: "Chris" <[EMAIL PROTECTED]>
Subject: Re: How to set up ADSL in Linux?
Date: Tue, 20 Jul 1999 16:15:41 +0800
Hi Mugur, is the 3COM509B connected to an ADSL modem? What brand is the ADSL
modem then?
Mugur <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Try a cheap 3Com509B. They can be had everywhere and yes, they are
recognized by
> Linux.
> At least that's what I have with ADSL.
> Mugur
>
> --
> ______________________________________
> Mailto: mugurd at nortelnetworks.com.
> PLS do not hit "reply" directly.
>
>
------------------------------
From: nico <[EMAIL PROTECTED]>
Subject: Firewall setup problem
Date: Tue, 20 Jul 1999 16:05:26 +0800
Hi all,
I want to setup a firewall with Linux 5.2 . My company is using a
private Class C network 192.168.1.0 and I have assign us the address
202.66.82.60 for our gateway external interface and 202.66.82.59 for our
external MS Exchange server.
So, How to set the policy and how many network card should I need?
How to make the DM-Zone work with the private network?
Should I need to tell my ISP about the Exchange server IP address?
thanks experts....
nico
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.networking) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Networking Digest
******************************
