Linux-Networking Digest #457, Volume #12 Fri, 3 Sep 99 02:13:42 EDT
Contents:
Linux Mandrake failing to detect NE2000 ([EMAIL PROTECTED])
Re: RFI - Analog Modem Channel Bonding ("Daniel O'Neil")
connecting to the internet ("Semegne Tafesse")
Re: My Dissapointment to find Linux not a viable solution (Patrick Farrell)
Graphic interface network connection monitor/logger (Vlar Schreidlocke)
multiple dns on one IP address
100base-Tx crossover cable, rtl8139 module options ("Lucas Fisher")
Re: ppp not working anymore (Glitch)
Enterprise Management Center - EMC2 ([EMAIL PROTECTED])
Internet transfer timeouts (Paul Trost)
router performance. (Matt Templeton)
Re: Can an ISP detect masquerading? (Chris)
Caldera OpenLinux 2.2 NFS install ([EMAIL PROTECTED])
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: Linux Mandrake failing to detect NE2000
Date: Fri, 03 Sep 1999 04:02:20 GMT
At least I think that it is failing to recognize
it, I am getting a message during boot up that
says:
Intializing Interface eth0:
Interface eth0 nitialazation delayed
I ran netcfg, linuxcfg, tried Mandrake's new boot
image(it doesn't work at all), and examined every
file I can think of.
I know that the card is good, it worked fine under
Win98, right up to the point I installed Linux,
and all other parts of the OS seem to be fine. Any
help would be appreciated.
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: "Daniel O'Neil" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware
Subject: Re: RFI - Analog Modem Channel Bonding
Date: Thu, 2 Sep 1999 17:34:05 -0400
http://linux-mp.terz.de/ claims to have a patched version of pppd for
multilink. However, i've never tried it myself, so why don't you and report
back to us? =)
Daniel O'Neil
remove the --NOSPAM-- from my email address to reply
Ken McCord <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Does anyone out there have any experience with bonding analog modems
> together to improve throughput (ie two 28.8k modems bonded together to
> get approximately 56k throughput)?
>
> Some of the technologies I've been looking at are:
>
> - Diamond's SupraSonic dual v.90 modems with 'Shotgun' technology.
> Anyone aware of a Linux driver for it?
>
> - EQL. Looks like you need the ISP to have either a Linux box or a
> Livingston Portmaster with EQL support.
>
> - slirp. Need shell account on ISP and it runs on their box (not
> something I can run on a Linux box being used as a router at home).
>
> - Multi-link PPP. Anyone aware of Linux support for MLPPP?
>
> Thanks,
>
> Ken McCord
------------------------------
From: "Semegne Tafesse" <[EMAIL PROTECTED]>
Subject: connecting to the internet
Date: Fri, 3 Sep 1999 00:02:52 -0400
1) How can I connect to the internet with 5.2 Red Hat Linux. My computer
is connect with internet on win98 operating system. How about
Linux.
2) When I install, Red Hat 5.2, I used custom but what to practice in
Networking. How can I do this? I do not have another computer or network
card, but is it possible to practice networking?
Please give your help.
Thank you
Semegne
------------------------------
From: Patrick Farrell <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.advocacy,comp.security.firewalls
Subject: Re: My Dissapointment to find Linux not a viable solution
Date: Thu, 02 Sep 1999 22:58:02 -0500
Reply-To: [EMAIL PROTECTED]
Watchguard makes a firewall called a Firebox that is a red box basically
running linux. The interface to setup runs completely on windows and is very
simple to manage. It has 3 10/100 interfaces. One trusted, one outside, and
one optional which is still behind the firewall. The optional is where you can
put web servers and such with their own security and if someone manages to
break it because you tend to have to leave web/ftp servers more open, they
still can't jump into the trusted side.
What you could do, is put a PC running MimeSweeper (e-mail virus scanner) on
the optional side. This becomes your smtp mailer, it's basically a relay, mail
goes to it, gets unpacked, scanned, filtered, whatever, and then thrown to your
real mailer behind the firewall. Outgoing goes from your internal mailer to
the Mimesweeper box to the rest of the world. Seems like a good solution.
check out http://www.watchguard.com for info on the firebox.
Patrick Farrell
Tam McLaughlin wrote:
> I love Linux and have been using it for the last year as our email
> and internet server but have found we cannot consider Linux as a
> serious option as a firewall, email/internet server and a virus
> checker for all incomming emails.
>
> I may be wrong and hope to be proved wrong but the companies in
> Glasgow/Edinburgh I have spoken to cannot support our requirements
> with Linux. I will explain our setup and what I would like to see
> and hope that someone can give me some suggestions or point me
> in the direction of some company that can help.
>
> We have ~ 150 PCs running win95
> 4 servers running SCO
> about to install a router between 2 offices using a leased line
> and a card to allow remote access from laptops.
> linux box running delegate proxy and qmail connected to a direct
> telephone line to a single dial up account.
>
> We would like to allow laptop users in to our network for email access and
> server access. I believe the cisco router will allow us to specify which IP
> addresses are allowed in and which are not. We wish to protect our LAN from
> the outside.
>
> So, I have been informed that the linux server is not secure and we need a
> proper
> firewall. We also need a better method of virus scanning rather than
> updating
> each PC each month. I believe we could use samba for this with some type of
> network AV software.
>
> I know hat I could go out and buy an NT server with MSProxy or whatever and
> some
> email package, firewall-1 and mime-sweeper. But this would cost a hell of a
> lot
> or money which I dont know our company would be willing to pay for (ok, i
> know all
> about the importance how much is our data worth etc etc...).
>
> So, why is there Linux based solution. Why is there no AV scanning software
> that can
> run on Linux? If there is , does anyone know of a local company that can
> help us?
------------------------------
From: [EMAIL PROTECTED] (Vlar Schreidlocke)
Subject: Graphic interface network connection monitor/logger
Date: Fri, 03 Sep 1999 03:19:33 GMT
Is there a realtime network connection monitor with a graphic
interface for Linux (Red Hat 6.0 w/ Gnome or KDE)? I'm looking for
something equivalent to Xnetstat and/or Nukenabber (Win98 programs)
that I can look at in realtime to see whos connected and on what port,
etc., possibly hackers.
------------------------------
From: <[EMAIL PROTECTED]>
Crossposted-To: redhat.networking.general
Subject: multiple dns on one IP address
Date: Fri, 03 Sep 1999 03:31:06 GMT
I've seen this question asked here before but it didn't fully clear up my
confusion, so I'm asking for more specifics here.
I've got one IP address and am trying to host several domains, including A
names and MX records.
I already have the following files:
/etc/named.conf
/var/named/myfirstdomain.com
/var/named/216.95.101
set up to provide DNS for "myfirstdomain.com".
If I now want to host "myseconddomain.com" on the same IP address, is the
following all I have to do?
1) Add a zone record for "myseconddomain.com" in named.conf that looks
similar to the one for "myfirstdomain.com".
2) Create a zone file called /var/named/myseconddomain.com that looks
similar to the one for "myfirstdomain.com", complete with MX record.
3) Leave the /var/named/216.95.101 file alone because I cannot have more
than one PTR record for a particular IP address. This means that people
who do reverse lookups on the dotted quad associated with
www.myseconddomain.com will get back www.myfirstdomain.com.
4) Restart the named service.
5) Update records with InterNIC.
Is this correct so far? I understand that my MTA (qmail) would have to
know about the two domains and deal with this accordingly.
If you could email me in addition to posting a reply, I'd appreciate it.
Thanks,
Loren
[EMAIL PROTECTED]
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: "Lucas Fisher" <[EMAIL PROTECTED]>
Subject: 100base-Tx crossover cable, rtl8139 module options
Date: Thu, 2 Sep 1999 23:29:50 -0500
I have a linux computer with a Realtek 8139 NIC and a Win98 computer with a
3Com 905B NIC, both of which are capable of 100Base-Tx. I have connected
them with a standard crossover cable hoping for a 100mbps connection
between them, but this didn't happen. The connection is only 10mpbs. Why
is this? Can I force the NICs to run using 100base-Tx?
This brings me to my second question. What are the module options for the
rtl8139 or where can I find them? I need to know, so I can force the
Realtek NIC to run using 100base-Tx.
thanks,
Lucas
------------------------------
Date: Fri, 03 Sep 1999 00:51:24 -0400
From: Glitch <[EMAIL PROTECTED]>
Subject: Re: ppp not working anymore
Brandon wrote:
>
> I recently bought a new 56k modem along with changing living locations.
> THose are the only 2 factors that have changed. But yet for some reason
> ppp wont work in linux now.
>
> I'm using rh 5.1 and everything was working fine before i moved to
> school in my apt and before i got a new modem. And no it isnt a
> winmodem, I made sure it wasnt and it even says on the box it will work
> with linux, among other OSes. But its not that the modem wont dial, but
> linux even says that ppp isnt even in the kernel, but I havent even
> touched the kernel, I had no reason to.
>
> At startup I'm all of a sudden getting a lot of unresolved symbols
> dealing with networking stuff ,ppp.o, being one of them. But I dont see
> why since I havent even been using Linux for about 3 weeks so nothing
> could have even been changed. Linux coudlnt even be accessed since I
> hadnt used my rescue disk to put LILO back on the mbr since I had
> reinstall Windows recently.
>
> I do knwo the modem was installed on Com5 and irq 12 which i was able to
> move it to com3 so i woudlnt have to make a new com port in linux. And
> I used setserial to tell linux the modem uses irq 12 but I still dont
> get anything, not even from minicom. Even if minicom did work I'd still
> have problems with ppp b/c for some reason linux acts like ppp.o isnt
> even in the kernel. I dont even know why it was complaining of
> unresolved symbols at boot time b/c i compiled in ppp directly into the
> kernel so i woudlnt have to worry about loading it everytime i wanted it
> as a module. I have include output that I get when i try to do a
> 'insmod ppp.o'. it makes no sense to me but i put it in here in case
> someone can help me with it.
>
> If anyone needs any more info please let me know b/c I want to get this
> damn thing fixed. It only makes me mad b/c I know that nothing has even
> touched any files in Linux and before i got my new modem, and changed to
> a diferent phone # since Im at school now, linux was working fine before
> that. Even if the modem is fine somehow the damn ppp got messed up.
>
> ppp.o: unresolved symbol kill_fasync_Ra394f189
> ppp.o: unresolved symbol dev_close_Rbc2476b5
> ppp.o: unresolved symbol dev_alloc_skb_R753f384a
> ppp.o: unresolved symbol dev_kfree_skb_R56cb10ad
> ppp.o: unresolved symbol register_netdev_R6ccb67ee
> ppp.o: unresolved symbol netif_rx_Rb089a42b
> ppp.o: unresolved symbol tty_register_ldisc_R52b306d9
> ppp.o: unresolved symbol unregister_netdev_R7a53f724
> ppp.o: unresolved symbol n_tty_ioctl_Rba30d46c
>
> this comes directly from /var/log/messages
>
> thanks
> Brandon
>
I fixed the problem. I knew my new modem was on a new irq but I never
thought it could be on a different port. I checked the port that it was
on using 'cat /proc/pci' and realized the port the new modem is on is
different than what the old modem used. So I did a setserial command
and changed the irq and the port address and ppp worked fine. I'm still
getting a lot of unresolved symbol messages (one of them being
ppp.o)...how do i get rid of those? And also I'm getting a lot of
messages concerning networking mostly related to I believe my system not
being able to figure out what my hostname is even though I changed it to
'homeofglitch' but for some reason nothing seeems to recognize this
change, at least no utilities.
thanks
Brandon
--
"Bill Gates?, I dont know any Bill Gates. Oh, you mean 'by putting
every conceivable
feature into an OPERATING SYSTEM, whether you want it or not, is
innovation' Bill
Gates? Yeah, I know the monopolizer"
http://web.mountain.net/~brandon/main.htm
For Beginners in Linux, Emulation, Midis, Playstation Info, and
Virii.
------------------------------
From: [EMAIL PROTECTED]
Subject: Enterprise Management Center - EMC2
Date: Fri, 03 Sep 1999 04:23:03 GMT
Enterprise Management Center - EMC2
The purpose of the EMC2 Forum is to provide a common location, for
people who are interested in enterprise management
tools/policies/practices, to share expertise through the various online
discussion groups. Discussions include SNMP, RMON, CMIP, and TCP/IP.
Vendor specific information can also be obtained in peer group forums
for Unicenter, OpenView, Tivoli, Spectrum, Command Post, NetCool, and
NerveCenter.
Enterprise Management Center - EMC2 Forum
http://www.delphi.com/emc2
The EMC2 Web is designed to bring ESM / NMS technologies together
through a single interface. Since we have many talents that are brought
together, engineers, consultants, and users bring shareable information
to the enterprise management community that are discussed at the EMC2
Forum and archived at the EMC2 Web.
Enterprise Management Center - EMC2 Web
http://members.tripod.com/emc2web
Christopher Smiga
EMC2 Web Designer and EMC2 Forum Co-Manager
[EMAIL PROTECTED]
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Paul Trost <[EMAIL PROTECTED]>
Subject: Internet transfer timeouts
Date: Fri, 03 Sep 1999 04:31:09 GMT
Lately every time I have been downloading from ftp sites or even when
surfing the web page the transfer times out after a few seconds. Telnet
sessions over the internet will also timeout relativelt quickly. My
connection is still active though. I am running SuSE 6.2 with a X2 US
Robotics modem and wvdial to dial up with. On some sites it doesn't time
out but still only downloads at half or a quarter of the speed it should.
My ISP said everything looked fine on their end. Does anyone have any
ideas what could be causing these strange timeouts?
Thanks,
Paul
[EMAIL PROTECTED]
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: Matt Templeton <[EMAIL PROTECTED]>
Subject: router performance.
Date: Fri, 03 Sep 1999 17:04:36 -0700
I have a pentium 75 with 64 meg ram configured as a masquerading router.
When I am on the internet side of the linux router a ping packet may
have a round trip time of 20ms. Just on the other side of the router, I
get round trip times of around 230ms. Downloads seem to take a lot
longer. Is there any way to improve this performance?
------------------------------
From: [EMAIL PROTECTED] (Chris)
Subject: Re: Can an ISP detect masquerading?
Date: Fri, 03 Sep 1999 05:44:21 GMT
On Wed, 01 Sep 1999 21:35:27 -0400, The Dude <[EMAIL PROTECTED]> wrote
in comp.os.linux.networking:
>How exactly the ports in the range 60000-65535 are used ?
>When one of my masqueraded client pass through Linux to go on
>www.someplace.com his ip is changed with the real one.
>Now my question is the port source and dest remain the same ?
>-If yes from where those higher ports appear on outside
>-If not and that's mean that the dest port get translated to 60000...
>how come it can connect to server listening only on port 80 ?
When you connect from one machine to another, the port numbers don't have
to be the same. TCP/IP communication compares quite well to the situation
where multiple offices each have a PBX. When you (a programmer in company
A) want to call your friend (an accountant in company B), you pick up the
phone. Any phone. The PBX finds an unused outside line and connects it
to your extension. You dial Company B and ask the operator to connect you
to your friend's extension. He picks up and the connection is complete.
It doesn't matter which phone you use to make the call-- his extension is
always the same. He doesn't care which extension you are using to place
the call.
Lets assume that your machine is 192.168.1.2 and the masquerading server
(your gateway) is 192.168.1.1 on your side of the firewall and
111.222.23.32 on the internet side. For the sake of argument, we'll
assume you already did the DNS lookup and the address of www.someplace.com
is 123.234.12.21.
When your program tries to get a page from the remote web site, it opens
up a port on your machine. It can request a particular port or ask for a
random one. Your address and that port number go together to form the
"from" field of the TCP/IP packet. For example, if the outgoing port is
1030, then the from field would be 192.168.1.2:1030. Since www servers
(almost) always listen to port 80, the "to" field will be
123.234.12.21:80. Since your request has to go through a gateway, the
packet is actually sent to 192.168.1.1 even though the "to" field contains
a different addressee. The masq server receives the packet and decides
whether or not you are allowed to use the masquerade feature. Since you
are, it opens a more-or-less random port above 61000 (say, 61111) on the
111.222.23.32 interface and completes the call to 123.234.12.21:80. It
replaces the "from" field in the packet with the masqueraded address and
port and sends it out.
All the server at www.someplace.com knows is that it's getting a
connection from 111.222.23.32:61111. It serves the request and sends the
reply back to that address.
When the masq server receives the packet addressed from 123.234.12.21:80
to 111.222.23.32:61111 it looks in its masquerade data table and sees that
it has to send the reply packet to 192.168.1.2:1030 (via the 192.168.1.1
interface). The program using 192.168.1.2:1030 thinks the packet came
straight back from 123.234.12.21:80. Neither the user on 192.168.1.1 nor
the remote web site know that any masquerading has taken place.
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.os.linux.questions;,comp.os.linux.help;
Subject: Caldera OpenLinux 2.2 NFS install
Date: Fri, 03 Sep 1999 04:44:22 GMT
I'm trying to install OpenLinux 2.2 on a second machine connected to a
home network. This machine has no CD-ROM drive so I have attempted to
set up another Linux machine on the network to export its /mnt/cdrom
directory so that the install files can be accessed via NFS. I'm
using a LISA boot floppy, going through all the options up to the part
where you select an install method. When I select NFS/SMB I get a
message stating something about attempting to remote mount, and then
the error message
mount cludp_create: RPC: Program not registered
A HOWTO I read told me that this meant that nfsd and mountd were not
running on the server. I've confirmed that they are by using rpcinfo
and ps aux. I won't swear that my exports file is correct, it's a
single line something like
/mnt/cdrom <hostname>(ro)
but I wouldn't think that would make a difference if for whatever
reason the remote machine doesn't even think that NFS is available.
Can you give me any suggestions?
Thanks in advance,
J. Freeman
------------------------------
** 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
******************************
