Linux-Networking Digest #200, Volume #11 Wed, 19 May 99 07:13:50 EDT
Contents:
Re: Win95->Linux PPP can't see past subnet (Andrew Sun - UCE revokable account)
Re: What are good 100baseTX cards for RH 5.2 (Vlad Petersen)
Re: Weird Samba/NFS file caching problem (Vlad Petersen)
chat script needed ("donnell")
BOOTPC and kernel 2.2.3 (Jean-Michel VANSTEENE)
How to force 10Mbit (not 100Mbit) on PCnet card ? (Zenon Fortuna)
Re: Firewall question (strong ruleset) ("David Means")
Re: samba refuses to broadcast ("--==[bolMyn]==--")
The World Wide Expo 1768 ([EMAIL PROTECTED])
Re: ip_allow ("David Means")
tulip device driver ([EMAIL PROTECTED])
Re: Problem with Chat Script (Bill Unruh)
Re: Samba frustrations (Fischl Anton)
Re: ISDN interface card and RH6 setup? (Lim Chee Onn)
neighbour table (Sven Steinberg)
Routing question. ("MarkT")
----------------------------------------------------------------------------
From: Andrew Sun - UCE revokable account <[EMAIL PROTECTED]>
Crossposted-To: comp.protocols.ppp
Subject: Re: Win95->Linux PPP can't see past subnet
Date: Wed, 19 May 1999 04:49:36 GMT
Robin Munn wrote:
>
> I'm setting up a Linux box as an experimental PPP server here at school.
> It's running mgetty on Debian 2.1 (slink). I have mgetty set up properly
> -- I've dialed in, logged in, everything's fine on the hardware side.
> Now I'm trying to get PPP working, and that's where the "fun" starts.
> I've got mgetty set up with an AutoPPP line in /etc/mgetty/login.options
> so that pppd starts up automatically. So I go over to the Windows 95 box
> I'm using to dial up, create a new connection in Dial-Up Networking,
> enter the phone number, and use the default settings for IP
> (server-assigned DNS and IP numbers) in the new DUN connection's
> properties box. Dial it up, enter a username and password, get through,
> everything looks good. Except for one "little detail"...
>
> The Windows box (I'll call it "winbox") can't see past the class C
> subnet it's on. The Debian box (I'll call it "thorn") has IP a.b.65.12
> and is assigning IP's a.b.65.101 and a.b.65.102 to incoming calls on
> each of its two modems. (I'll hook up more modems once I can get this
> whole thing working). Now thorn is on a 20-bit subnet: its netmask is
> 255.255.240.0. At IP a.b.64.1 is a router that connects this subnet to
> the rest of the school's network (and thus to the Internet). a.b.64.2
> and a.b.64.3 are the primary and secondary DNS servers for this subnet.
> thorn can see the a.b.64.0/20 subnet just fine and can also see the rest
> of the network and the Internet, but winbox can't see anything outside
> the a.b.65.0/24 subnet -- can't even see a.b.64.1, so it can't see the
> Internet either. winbox can't see the DNS servers either, so I've been
> pinging things by IP. Machines on the a.b.65.0/24 subnet can ping
> winbox, and winbox can ping them right back. But anything outside the
> a.b.65.0/24 subnet can't see winbox, and winbox can't see them. They can
> see thorn just fine, though, and thorn can ping them just fine.
>
> Running winipcfg on winbox I get:
>
> PPP Adapter
> Adapter Address: 44-45-53-54-00-00
> IP Address: a.b.65.102
> Subnet Mask: 255.255.255.0 <-- should be 255.255.255.255, see below
> Default gateway: a.b.65.102 <-- shouldn't this be a.b.65.12?
>
> route -n on thorn gives:
>
> Destination Gateway Genmask Flags Metric Use Iface
> a.b.65.102 0.0.0.0 255.255.255.255 UH 0 0 ppp0
> a.b.64.0 0.0.0.0 255.255.240.0 U 0 26 eth0
> 127.0.0.0 0.0.0.0 255.0.0.0 U 0 0 lo
> 0.0.0.0 a.b.64.1 0.0.0.0 UG 1 5 eth0
>
> In /etc/ppp/options I have a "netmask 255.255.255.255" line, and that
> shows up just fine on thorn's routing table. But what winbox tells me in
> winipcfg doesn't seem right. A netmask of 255.255.255.0 and an incorrect
> gateway *would* kind of explain why winbox can see everything with an IP
> of a.b.65.(whatever) but can't see anything else... Should I change the
> netmask in thorn's /etc/ppp/options to be 255.255.240.0? Or would that
> have no effect?
>
> I'd appreciate help from anyone who has any ideas about what could be
> causing this. I've been working on it for several days now and am no
> closer to understanding the cause of this problem.
>
> Thanks.
>
It appears that your Win95 is behaving as if it had no
default route (despite what winipcfg says).
What does the "route print" command on Win95 show?
Is "Use default gateway on remote network" enabled on Win95 for
this PPP connection?
Can Win95 reach Internet addresses anyway (unrelated to your a.b.?.?)?
Can you verify (check modem lights)
if any traffic is going out via PPP from Win95,
when you ping unreachable places?
In general, netmasks are not relevant for PPP connections,
and setting them should have no direct effect (there are exceptions
with bad implementations though).
With Win95, the PPP netmask you see must have made its way into a
Win95 routing table entry. This misfeature is likely to be causing
your Win95 to be able to reach a.b.65.0/24.
Also, Win95 routing entries that point to PPP interfaces
make reference to the connection's local IP address
(rather than the next hop, remote IP address).
This appears to be the normal, but it is another MS oddity.
andrew
--
"Using & Managing PPP," March 1999, http://www.oreilly.com/catalog/umppp
------------------------------
From: Vlad Petersen <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux,comp.os.linux.redhat
Subject: Re: What are good 100baseTX cards for RH 5.2
Date: Tue, 18 May 1999 22:27:08 -0700
Frank Looper wrote:
> Dean and Mary Guenther wrote:
> >
> > I have just installed RH 5.2 on a Micron 486. What is a good choice for
> > a new 100BaseTX (non-PCI) card for this machine? -- Dean
>
> IF nothing else works, try the NE2000 card. That covers most 10baseT
> cards, but I have no idea how it'll work on your 100baseT one (mine's
> 10/100 and it works great)
>
> Frank/Floops
Intel makes (or used to make) some 10/100 ISA cards. I am not near the
Linux system which uses this card and don't remember its exact name but
it was some kind of EtherExpress line. I don't know whether it's as fast
as PCI 100baseT cards but I was able to network that system with others
which used 3Com PCI-based 100baseT cards and that was sufficient enough
for me.
--
Vlad Petersen | <vpetersen at kensington dot com>
Vancouver, BC | *Good pings come in small packets*
SIGSIG -- signature too long (core dumped)
------------------------------
From: Vlad Petersen <[EMAIL PROTECTED]>
Subject: Re: Weird Samba/NFS file caching problem
Date: Tue, 18 May 1999 22:38:33 -0700
[EMAIL PROTECTED] wrote:
.....
> My problem is that files edited on the server from the laptop do not
> update properly when accessed from the Windows machine. When I type
> "make", make determines that the file has changed, and so invokes the
> compiler properly to recompile the file. Unfortunately, the file
> contents appears not to have updated, because I get the same compile
> errors. The best way I've found to get around this is to use the DOS
> "type" command to type the file to the screen. Only then does the
> file completely update and compile properly (or at least with
> different errors).
......
> The problem goes away if I log in to the server and edit it there.
> (But this solution is undesirable.) This led me to believe that the
> problem was with NFS. However, remounting with the "noac" option (no
> attribute caching) has no effect. Nor does explicity sync'ing the
> file systems on either the NFS client or server ("sync" from shell).
>
> Any suggestions? I'm at my wit's end.
It sounds like it's a Windows-related problem. Could it be the file does
not get updated immediately after you "save" in a Windows editor and is
instead kept as a cached copy in local buffers, and because Windows
file/drive caching is always enabled but that's not the case in DOS
(unless you run some kind of smartdrive.exe thingie) the file gets
synced instantly when you use that DOS "type" command. There is some
setting you can add to Win95 system.ini file to limit or disable file
caching but I don't remember what it was. No matter whether what I am
proposing here leads you to a solution or not, could you please post
here what caused it when you find out?
--
Vlad Petersen | <vladimip at uniserve dot com>
Vancouver, BC | *Good pings come in small packets*
SIGSIG -- signature too long (core dumped)
------------------------------
From: "donnell" <[EMAIL PROTECTED]>
Subject: chat script needed
Date: Tue, 18 May 1999 21:28:40 -0700
new to linux using redhat 5.2 k2.0.36 on a i486
can someone post a generic chat script that has ogin and password that I
could study and possibly edit into a working chat script? I'm new to this
would like to see the proper format of a chat script so I can learn to write
them
thanks
donnell
------------------------------
From: [EMAIL PROTECTED] (Jean-Michel VANSTEENE)
Subject: BOOTPC and kernel 2.2.3
Date: Mon, 17 May 1999 09:46:13 +0200
Hi,
BOOTP works fine with the kernel 2.0.36 from Redhat 5.2
the ifcfg-up does a ifconfig eth0 0.0.0.0 broadcast 255.255.255.255
netmask 0.0.0.0 and a route add -host 255.55.255.255 which are ok.
BOOTP doesn't work with my kernel 2.2.3. the same ifconfig
is rejected with a SIOCIF... for broadcast and netmask options
Could you tell me what happens? Maybe a config option i missed ?
I tried with new net-toools with no luck.
Thanks for your help.
------------------------------
From: [EMAIL PROTECTED] (Zenon Fortuna)
Subject: How to force 10Mbit (not 100Mbit) on PCnet card ?
Date: 18 May 1999 22:22:42 -0700
I have the PCnet/PCI II 79C971A card, which is 10BT/100TX.
When I connect it to a hub (with some 10Mbit connections active)
the PCnet card still works with the 100Mbit speed.
How to *force* the card to slow down and to work with 10Mbit
by default ?
My system has the 2.0.35 kernel, with the CONFIG_PCNET32 option
to link the driver to the kernel.
TIA,
Zenon
------------------------------
From: "David Means" <[EMAIL PROTECTED]>
Subject: Re: Firewall question (strong ruleset)
Date: Tue, 18 May 1999 22:52:49 -0700
Francois Magnan <[EMAIL PROTECTED]> wrote in message
news:LJf03.12618$[EMAIL PROTECTED]...
> Hi,
>
> I have been looking at a sample strong firewall ruleset and I
> wonder why it blocks incoming packets for which the remote tcp port is
> a privileged port like in the following rule:
>
> ipchains -A input -p tcp -i $EXTERNAL_INTERFACE \
> -s $ANYWHERE $PRIVPORTS -d $IPADDR -l -j DENY
>
> where PRIVPORTS="0:1023".
>
> I understand why I should block incoming packets going to the
> privileged ports on my firewall but why should I be concerned about
> the above??? How insecure would my system be if I omit this?
> Francois Magnan
> Departement de Mathematique & Statistiques
> Universite de Montreal
> email: [EMAIL PROTECTED] (MIME, NeXTMail Ok!)
Is this rule at the end of a list of rules, most of which accept packets
that
come in from outside from*particular* privileged ports ? If so, then the
way to read this is: "if it hasn't matched something already, and claims to
be coming from a privileged port, ditch it."
------------------------------
From: "--==[bolMyn]==--" <[EMAIL PROTECTED]>
Crossposted-To: linux.samba
Subject: Re: samba refuses to broadcast
Date: Tue, 18 May 1999 03:23:53 GMT
Two things:
1. Make sure that you "specify remote announce"
2. And have a valid active guest account. Believe it or not but if you have
an invalid guest account, you will not see your Samba server in your
Windows clients. I found this info in Samba readme files.
Kai Krakow wrote:
> Hi!
>
> Since I recompiled my 2.2.5 kernel Samba doesn't work properly...
>
> I can still connect to the Samba services etc but Samba cannot announce
> itself as domain controller and master browser... The log file says:
> "udp_sendto 192.168.1.255: operation not permitted" or something similar...
>
> That means the UDP protocol refuses to broadcast into my local network to
> find other samba browsers... This means that samba's master election doesn't
> work...
>
> What could have went wrong? Which kernel options do I have to take care of
> to make broadcasting working again?
>
> Or could this be related to a config problem with samba?
>
> TIA
>
> x-post: linux.samba, comp.os.linux.networking
> please set appr. f'up2
>
> ciao
> Kai
--
Bolek,
URL: http://www.bolek.com
e-mail: [EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To:
comp.os.linux.development.system,comp.os.linux.hardware,comp.os.linux.m68k,comp.os.linux.misc,comp.os.linux.portable,comp.os.linux.powerpc,comp.os.linux.setup,comp.os.linux.x,comp.os.lynx
Subject: The World Wide Expo 1768
Date: 18 May 1999 22:45:29 GMT
ngrihwfolww
------------------------------
From: "David Means" <[EMAIL PROTECTED]>
Subject: Re: ip_allow
Date: Tue, 18 May 1999 22:38:46 -0700
root <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
> I have a Linux box (RH5.2) running as a mail gateway and running IP
> Masquerading. After placing my ISP's mail relay into the relay_allow
> file and placing my internal class C (192.168) along with 127.0.0.1 into
> the ip_allow my internal users can send mail out with no problem.
>
> How go I get outside mail into the shop via the same box?
I think you can allow mail from all sources by emptying the file
/etc/mail/relay_allow
> My DNS MX is pointing at the box. When I mail from an outside source, It
> bounces with the proverbial '551 relay not allowed' kind of error.
>
> I guess I'm looking for how to turn off the relay filter and allow all
> mail from all sources into the gateway, Or am I to have my MX entry
> point to a second IP and use tunneling to route it directly to my
> internal sendmail hub?
>
------------------------------
From: [EMAIL PROTECTED]
Subject: tulip device driver
Date: Wed, 19 May 1999 05:55:26 GMT
Eeek.
Just upgraded to kernel 2.2.5, running RH5.2
I have 2 nic cards -- a 3Com 3c509 10/100 which works beautifully (for
my cushy dsl connection) and a Linksys 10/100E which uses the DECTulip
device driver for my LAN/firewall.
The linksys is not working. No messages when the kernel boots, just
doesn't configure the darn thing. Using ifconfig is useless, merely
reports the following:
/sbin/ifconfig eth1 192.168.2.1
SIOCSIFADDR: Operation not supported by device
eth1: unknown interface: Operation not supported by device
I've compiled and recompiled my kernel, the damn driver is there. It
just ain't workin. Anyone got any suggestions?
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
From: [EMAIL PROTECTED] (Bill Unruh)
Subject: Re: Problem with Chat Script
Date: 19 May 1999 06:35:14 GMT
In <[EMAIL PROTECTED]> Bob Martin <[EMAIL PROTECTED]> writes:
]Rand Simberg wrote:
]>
]> OK, now that I think I've got my modem working without stepping on my
]> mouse, I'm trying to set up a proper connection to my ISP. Per the
]> instructions on the PPP web page, just to start, I'm sending the
]> following single command line, without a chat script.
]>
]> Command:
]>
]> pppd /dev/ttyS2 115200 debug connect "/usr/sbin/chat -v ''
]> ATD13108960011 CONNECT '' ogin:
]> us,ppp,simberg assword: *********"
Are you sure it is on the same command line.
I would also put '' around the
'us,ppp,simberg'
(Is that really what they want as a username?)
...
]> May 18 17:21:08 localhost chat[771]: netcom login:
]> May 18 17:21:08 localhost chat[771]: -- got it
]> May 18 17:21:08 localhost pppd[769]: Connect script failed
]> May 18 17:21:09 localhost pppd[769]: Exit.
]>
]> Note that it expected "ogin:" and it got it, but it quits right after
]> that instead of sending the user name. Anyone have any idea what's
]> going on?
------------------------------
From: Fischl Anton <[EMAIL PROTECTED]>
Subject: Re: Samba frustrations
Date: Wed, 19 May 1999 10:12:23 +0200
Pavel Grinfeld schrieb:
>
> Hi,
>
> I think I'm a pretty experienced Linux user but I seem to be completely
> missing this Samba thing.
> I followed the instructions for installation and setup but I have no
> clue how to even test it.
>
> For instance, under 95 I go to Tools->Map Network Drives
> Then what? What do I enter for the path?
\\SambaServer\DirNameInSmbConf
(e.g. "\\SambaServer\homes")
> Then I go Add Printer->Network Printer
> Then what? What's the Path.
\\SambaServer\PrinterNameInSmbConf
> If I do Browse, I get Network Neighborhood... Entire network but nothing
> under there.
Are the Workgroups of Samba and 95 the same? You should see the
Server and all
dirs and printers.
> Many thanks in advance!
>
> Pavel
------------------------------
From: Lim Chee Onn <[EMAIL PROTECTED]>
Subject: Re: ISDN interface card and RH6 setup?
Date: Tue, 18 May 1999 23:27:45 +0800
Brian wrote:
>
> Hello,
>
> What do I need to do to setup RH6 to use my Eicon Diva 2.0 PCI ISDN
> interface card? (U.S. U interface). I've been hunting around and believe
> I need to use the HiSax driver, but have no idea how to set it up. Where
> can I find information on getting this all going? I couldn't find anything
> in any of the usual HowTo documents.
>
> Thanks!
> Brian
A quick look at the kernel sources documentation shows that the EICON
DIVA is supported as 'Type 11'. I do not know much more as I do not yet
have an ISDN connection here. Look in
/usr/src/linux/Documentation/isdn/README.HiSax for more information on
how to configure Linux to use your card.
I believe that there is a Linux-ISDN homepage somewhere but it has
slipped my mind for the moment. Try doing a search at www.deja.com for
"linux isdn hisax" and see what it turns up.
Cheers.
--
=========================================================
A successful man makes more money than his wife can spend
A successful woman finds the man above
=========================================================
Alex C. O. Lim
Future Trend Computer Services
http://www.ftrend.com.my
=========================================================
------------------------------
From: Sven Steinberg <[EMAIL PROTECTED]>
Subject: neighbour table
Date: Wed, 19 May 1999 10:07:55 +0200
Hello,
I=B4ve got a router with 3 NW-Devices and an IP-Packet-Filter. If there i=
s
a lot of network traffic i get the following error: =
neighbour table overflow
and the net is very slow.
It is possible that some packets can reach the router over two devices.
Could someone explain the function and the logical position of the
neighbour table?
Where can i get further informations?
thanks
Sven
-- =
BETEILIGE DICH AN DER PETITION GEGEN SPAM =
http://www.politik-digital.de/spam/
------------------------------
Reply-To: "MarkT" <[EMAIL PROTECTED]>
From: "MarkT" <[EMAIL PROTECTED]>
Subject: Routing question.
Date: Wed, 19 May 1999 10:35:44 GMT
Can someone tell me what my route table should look like with the following
situation:
I have 1PC running Linux with two network cards installed.
One 192.168.0.25 mask 255.255.255.0 (for intranet)
one 24.4.46.125 mask 255.255.248.0(for internet)
My first upstream router on the internet is 24.4.40.1
I have the 24. address plugged into a utp hub that is plugged into a cable
modem next to a NT40 webserver with a 24.4.41.201 address with mask
255.255.248.0 NT is currently active as webserver.
Right now, I can ping both internal network cards, and also ping any local
address 192. I cannot ping the other 24.4.41.201 that is plugged into the
same hub as the 24.4.46.125 network card. I checked the subnet 255.255.248.0
and both of the 41 and 46 addresses are on the same subnet.
I don't think I have the gw or default gw setup right.
Any suggestions?
Thanks,
MarkT.
------------------------------
** 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
******************************
