Linux-Networking Digest #602, Volume #10 Tue, 23 Mar 99 08:13:42 EST
Contents:
IPCHAIN help for web server behind gatewat ([EMAIL PROTECTED])
Portforwarding-problem ("Kim �rkenrud")
Re: Yet another firewall question (Ng Wai Wing)
Re: ne card no longer detected ("Terry East")
Re: DNS+ISP=broken network ("Greg")
Re: newbie net card problem ("Terry East")
Re: Modeprobe can't locate module ("Terry East")
Re: Linux + 2 Ethernet cards ([EMAIL PROTECTED])
Re: Teles 16.3c ISDN card + kernel 2.0.36 (Udo Freitaeger)
Re: Win95 X-software suggestion (Dr Paul Kinsler)
Re: Help on PPP dial-up (Farid)
Re: samba newbie FYI ("Eddy Young")
Re: NFS problems with Linux 2.2.x server, freebsd client
([EMAIL PROTECTED])
Re: Dial up Linux (Mike Jagdis)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED]
Subject: IPCHAIN help for web server behind gatewat
Date: Tue, 23 Mar 1999 10:09:25 GMT
Hi, hoping someone can help me,
I've got a gateway running linux connected to an ADSL connection. I use an
ipchain MASQ to allow my 192.169.x.x boxes to use the internet. One of these
boxes is running linux with Apache. All of my internal boxes can bring up my
web page by typing the IP address of the www server into a browser
(192.168.0.2). I need Apache to service requests from the internet.
I figured the best way to do this would be to create an INPUT ipchain on the
gateway. So I tried this with the following command (with xx being my net
ip):
ipchains -A input -p tcp -s xx.xx.xx.xx 80 -j REDIRECT -d 192.168.0.2 80
Didn't work. Outside addresses cannot make a connection with the www server.
Beyond this, I tried a billion other things and every variation of each that I
could come up with.
Any help would be appreciated ...
Ryan C
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: "Kim �rkenrud" <[EMAIL PROTECTED]>
Subject: Portforwarding-problem
Date: Tue, 23 Mar 1999 10:09:20 +0100
Hi
I'm using RH 5.2 with kernel 2.2.3.
I thought I'd test Portfw but when I run xconfig these options aren't
highlightet so I just can't select them.
Is there something elese that I must select before. I have managed to
compile the features into the kernel before but now it's hopeless.
Any suggesions?
Please send me a mail also.
Kind regards
Kim
[EMAIL PROTECTED]
------------------------------
From: [EMAIL PROTECTED] (Ng Wai Wing)
Subject: Re: Yet another firewall question
Date: 23 Mar 1999 06:51:04 GMT
nobody ([EMAIL PROTECTED]) wrote:
: Alas, from the frequency of posts on this topic it seems this is a
: very confusing issue. I'm trying to set up the old localnet->PPP->isp
: firewall with RH 5.1 (upgraded kernel 2.0.36). I've followed the HOWTO
: and a Linux Network Toolkit book, and everything seems right. Without
: listing the full config, here's what I can do:
: 1) Ping the internet from the firewall
: 2) Ping the ethernet card from the firewall
: 3) Ping the Win95 box from the firewall
: 4) Ping the firewall's ethernet card from the Win95 box.
: What I can't do:
: 1) Ping the internet from the Win95 box
: 2) Ping the the PPP interface from the Win95 box
: The PPP interface IP address shows up in the routing table as the ISP
: router number instead of the dynamically allocated IP address, which
: seems confusing to me.
: My question is this: does this functionality point to a standard
: mistake (bad routing table?) when people set up firewalls? In the many
: linuxconfig dialogs there seems to be many things to tweak, and I
: suspect that it's something there that'll make it all happen.
: Thanks in advance...
: Jim
------------------------------
From: "Terry East" <[EMAIL PROTECTED]>
Subject: Re: ne card no longer detected
Date: Fri, 19 Mar 1999 13:48:17 +0200
I haven't seen your earlier posts so forgive me if I don't understand the
problem.
It looks like you've recompiled your kernel and either the linkup between
depmod,modprobe has gone awry or you haven't compiled enable loadable module
support and kernel module loader (kerneld) kerneld into your new kernel.
Try
1) modprobe eth0 - if this works (your ip address and routing tables won't
be setup) , then your either your startup routines are not correct (checkout
/etc/rc.d/rc.sysinit for redhat 5.x) or you dont have kernel module loader
compiled in. (for redhat 5.2 make sure that you $ cat /proc/version
/lib/modules/2.x.y/.rhkmvtag - this will ensure that rc.sysinit will find
the correct set of modules when it does the depmod -a )
2) it 1) fails try insmod /lib/modules/2.x.y/net/ne.o where 2.x.y is your
current kernel
If this fails you haven't enabled loadable module support or your kernel
compile was not a success. if it works, you probabll have depmod,modprobe
confusion
Good Luck - hope this helps
Terry East
T.L. (Terry) Branscombe wrote in message <[EMAIL PROTECTED]>...
>T.L. (Terry) Branscombe ([EMAIL PROTECTED]) wrote:
>
>: [snip]
>: Mounting local filesystems...
>: /proc on /proc type proc (rw)
>: SIOCSIFADDR: no such device
>: SIOCSIFNETMASK: no such device
>: SIOCSIFBRDADDR: no such device
>: SIOCADDRT: no such device
>: Mounting remote filesystems...
>: [snip]
>
>Just another note. These messages are the same as those generated by
>ifconfig. For some reason, Linux can no longer detect the eth0 device,
>while Win95 on the same machine still can.
>
>
>--
>+------------------+
>Terrence Branscombe
>[EMAIL PROTECTED]
------------------------------
From: "Greg" <[EMAIL PROTECTED]>
Subject: Re: DNS+ISP=broken network
Date: 23 Mar 1999 11:09:36 GMT
Alan,
I'm no guru :) but if I followed your question correctly a edit to your
/etc/resolv.conf might do the trick along with the /etc/hosts
add this line or edit existing
#resolv.conf
search <your domain.name>
nameserver <your local ip> # not sure if this is needed but try it as a
third entry.
in your /etc/hosts file add your ip to domain name, plus a alias if you
want.
#hosts # the alias
198.xx.x.xx foo.myname.org foo
198.168.xx.x your.win95.org win
if I'm not mistaken or I will be corrected the lookup should work this way
to find your local domain lookups.
also check your /etc/hosts.conf for this entry
#hosts.conf
order hosts,bind
multi on
There maybe a correct or easier way but give this a try in case its
the only reply you receive ; )
Greg.
Alan Mead wrote in message <[EMAIL PROTECTED]>...
>I slogged through comp.os.linux.networking looking for an answer to my
>question and I think Bill Clark answered it below but I'm going to ask
>all you guru's to verify this before I give up...
>
>I have 2 Win95 and 1 Linux machines on a 10 Mbps peer-based network at
>home. I added the Linux box as a Samba server and all was cool after
>assigning all boxes bogus IP's. I can dial the ISP from either Win95
>machine and it gets a new IP dynamically and everything is still cool
>(I guess the Windows dial-up and nic adapters can live with having
>different IP's).
>
>But a major reason for adding the Linux box was to be able to try
>web-stuff out at home and I cannot set up DNS to work at home. As a
>result, I use IP's but browsers sometimes refuse to connect to, say,
>http://192.168.1.2 or http://192.168.1.2/manual/ without dialing my
>ISP. Whenever I set up the Linux box as DNS server for a Win95
>machine, it breaks name resolution when dialed into my ISP.
>
>So my question is, how do I configure these machines so that the Linux
>box serves the DNS for my little net and the ISP's DNS is consulted
>for the rest of the world?
>
>Now, Bill writes in article <[EMAIL PROTECTED]>,
>
>>[a lot of stuff about cacheing and DNS not working with dynamic IP's]
>>
>>If you're going to set up a domain, do it right. Get a static IP.
>>
>>-Bill Clark
>
>Note that I don't anticipate anyone ever calling into this network so
>I don't care too much about other DNS servers and caches (except,
>maybe as they break or enable ftp call-back).
>
>And quickly, I'm not going the IP-Masq route of putting the Linux box
>on the net because I need to ftp frequently from the Windows boxes and
>I understand that IP-Masq breaks certain programs like ftp that need
>to "call-back"?
>
>Your advice is very much appreciated. Sorry for the length.
>
>-Alan Mead
------------------------------
From: "Terry East" <[EMAIL PROTECTED]>
Subject: Re: newbie net card problem
Date: Tue, 23 Mar 1999 13:27:37 +0200
use the dos diagnostic disk for your 8029 card to disable plug and play and
setup the irq and io-base.
insert these lines in /etc/conf.modules
alias eth0 ne
options ne irq xx where xx is the irq you've just setup.
Make sure you don't have conflicting irq ot base io-addresses
Next try modprobe eth0
Your card should now be visible.
Note: this will only setup the card , you still have to configure you ip
addresses etc. If you're not sure how to do this, restart the install
procedure and specify the io and irq under the ne2000 card detect
Good Luck
Terry East
dave canavan wrote in message
<[EMAIL PROTECTED]>...
>I have a realtek 8029 chipset based card and have just installed 2.0.32
>kernel on a PC. Whatever I seem to do, it just does not seem to want to see
>the network card either as an 8029 realtek (driver and instaructions form
>their site) or as an NE2000 compatible. If anybody can shed any light on
>this I would be forever in their debt
>
>Thanks
>Davey C
>
>
>
------------------------------
From: "Terry East" <[EMAIL PROTECTED]>
Subject: Re: Modeprobe can't locate module
Date: Tue, 23 Mar 1999 13:32:40 +0200
These are aliases for ipx and appletalk which you presumably just removed
from your kernel.
If you want to remove these messages (you don't need to , nothing is wrong)
add these 2 lines to /etc/conf/modules
alias net-pf4 off
alias net-pf5 off
Terry East
Michael Moffitt wrote in message ...
>Just built kernel-2.0.36
> during boot I get
>
> modprobe can't locate module net-pf4
> modprobe can't locate module net-pf5
> modprobe can't locate module net-pf4
> modprobe can't locate module net-pf5
> modprobe can't locate module net-pf4
> modprobe can't locate module net-pf5
>
>everything seems to work.
>
>can anybody tell what this is..
>
>please reply via email mailto:[EMAIL PROTECTED]
>
>thx,
>Mike
>
>
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Linux + 2 Ethernet cards
Date: Tue, 23 Mar 1999 08:47:25 GMT
In article <[EMAIL PROTECTED]>,
Raymond Doetjes <[EMAIL PROTECTED]> wrote:
> Read the Multi Homed howto. There you see how to add a eth1 as lilo
> parameters this does work.
>
> Raymond
>
I looked everywhere for this "Multi Homed howto," but couldn't find it. Does
anyone know where I can find this?
--Daniel
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: Udo Freitaeger <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: Teles 16.3c ISDN card + kernel 2.0.36
Date: Tue, 23 Mar 1999 10:44:07 +0000
John Wong wrote:
> Hi,
>
> Anybody got a Teles 16.3c ISA PnP ISDN card working with
> Linux kernel 2.0.36??
>
> I've compiled the ISDN subsystem as a module...
> HiSax drivers as modules... here's the relevant section in
> the .config file...
>
> CONFIG_ISDN=m
> CONFIG_ISDN_PPP=y
> CONFIG_ISDN_PPP_VJ=y
> CONFIG_ISDN_MPP=y
> CONFIG_ISDN_DRV_LOOP=m
> CONFIG_ISDN_DRV_HISAX=m
> CONFIG_HISAX_EURO=y
> CONFIG_HISAX_16_0=y
> CONFIG_HISAX_16_3=y
> CONFIG_HISAX_TELES3C=y
> CONFIG_HISAX_TELESPCI=y
> CONFIG_HISAX_S0BOX=y
>
> Booted with this kernel... got isapnp to load the following isapnp.conf..
>
> (CONFIGURE TAG2620/209717183 (LD 0
> (IO 0 (SIZE 2) (BASE 0x0300))
> (INT 0 (IRQ 10 (MODE +E)))
> (NAME "TAG2620/209717183[0]{TELES.S0/16.3c Plug&Play}")
> (ACT Y)
> ))
>
> ran isapnp and got ...
>
> Board 1 has Identity 76 04 03 02 01 00 20 32 0d: CIR2000 Serial No 67305985
>[checksum 76]
> Board 2 has Identity bd 0c 80 07 bf 20 26 27 50: TAG2620 Serial No 209717183
>[checksum bd]
> TAG2620/209717183[0]{TELES.S0/16.3c Plug&Play}: Port 0x300; IRQ10 --- Enabled OK
>
> BUT when i tried to load the HiSax modules by modprobe... i
> got the following errors...
>
> # modprobe hisax type=14 protocol=2 io=0x300 irq=10
> Initialization of hisax failed
> Initialization of hisax failed
> ISDN subsystem Rev: 1.44.2.9/1.41.2.11/1.48.2.27/1.28.2.2/none loaded
> Teles 16.3c: IRQ(10) getting no interrupts during init 1
> Teles 16.3c: IRQ(10) getting no interrupts during init 2
> Teles 16.3c: IRQ(10) getting no interrupts during init 3
> HiSax: Card Teles 16.3c not installed !
> Initialization of hisax failed
> ISDN-subsystem unloaded
>
> What could be the problem? any ideas?
>
> regards,
>
> John Wong
Hi,
Maybe you should try to configure the card using a different irq / io-address (in
/etc/isapnp.conf). In the beginning I had the same problems, but since I changed irq
to 10 and
io to 0x0580 every works fine.
Udo
------------------------------
Crossposted-To:
alt.os.linux.slackware,comp.os.linux.x,comp.os.linux.setup,comp.os.linux.misc,linux.redhat.misc
From: [EMAIL PROTECTED] (Dr Paul Kinsler)
Subject: Re: Win95 X-software suggestion
Date: Tue, 23 Mar 1999 09:34:53 +0000 (GMT)
In comp.os.linux.misc Jon Slater <[EMAIL PROTECTED]> wrote:
> I want to be able to run X on my Win95 box. I am connecting to a RedHat
> Linux 5.2 box.
I've used eXceed in the past, it seemed to do the job ok.
--
==============================+==============================
Dr. Paul Kinsler
Institute of Microwaves and Photonics
University of Leeds (ph) +44-113-2332089
Leeds LS2 9JT (fax)+44-113-2332032
United Kingdom [EMAIL PROTECTED]
WEB: http://www.ee.leeds.ac.uk/staff/pk/P.Kinsler.html
------------------------------
From: Farid <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.help,comp.os.linux.misc,comp.os.linux.setup
Subject: Re: Help on PPP dial-up
Date: Tue, 23 Mar 1999 07:20:18 -0500
Well,
I did as David said, and it seems that the chat failed because the
chat-script failed. And then I went on with "more chat-ppp1" (at the
/etc/sysconfig/network-script, which I get from the /var/log/messages)
to check the chat-script (mine is named chat-ppp1) and there something
like this:
'ABORT' 'KILL'
'ABORT' 'ERROR'
'ABORT' 'ETC'
'ABORT'.....
'ABORT'.....
'ABORT'.....
'ABORT'......
" " OK
...................
the contents of what I think is the chat-script begins with a lot of
'ABORT' for 5 or 6 lines, before something that looks make sense a bit
to me begins, at line 6 or 7 after the last 'ABORT' line (beginning with
the " " OK line).
Is there something wrong with my chat-script or is there something else.
All that I know (from looking at the /var/log/messages) is that the
connection failed because chat program failed, and the log messages is
something like this:
pppd (789): kernel 2.2.3 .....
chat (790): failed
ppp(792): exit
anybody? help me please. Thanks.
David Pace wrote:
> Farid wrote:
>
> > Hi,
> >
> > I need help in installing the PPP dial-up. I tried to setup PPP
> dial-up
> > through XWindow, but it seems that I cannot connect, although the
> line
> > to the ISP isn't busy.
> >
> > I configured the modem ok, and can dial to the ISP, the modem ring
> ok I
> > can hear the tone and the ISP server receive connection, but no
> > connection is established. I don't know which part isn't right. I
> got
> > the hostname, the domain, DNS server, phone number. Which part is
> still
> > wrong.
> >
> > Actually, I'm not that good with Linux or PC...I got the XWindow
> install
> > because I am using redhat 5.1 (installation set the XWindow for
> me)...I
> > try reading the documentation, but became worse, I don't understand
> a
> > thing. Tension...
> >
> > Huh, hope somebody can help to clear the things up to me. Thanks.
>
> Take a look at:
> cat /var/log/messages
>
> after you do the dial up.
> It will help show where the call is getting stuck.
> You could also post the tail end of the file to this newsgroup
> and somebody might offer more help.
>
> --
> David Pace Free Trading software: http://www.daveware.com
------------------------------
From: "Eddy Young" <[EMAIL PROTECTED]>
Subject: Re: samba newbie FYI
Date: Tue, 23 Mar 1999 11:13:11 +0400
Robert Rice <[EMAIL PROTECTED]> wrote in message
news:7d6sb0$is3$[EMAIL PROTECTED]...
> ok, after a few days of stressing out over this i finally found out
that
> samba does not support user level sharing in windows. so if anyone
is
> having the problem of viewing a list of users to allow to share in
windows,
> forget it, switch to share level security. thanx to those who tried
to
> help.
Robert,
Samba does support user level sharing in Windows, but you will still
need an NT domain controller to authenticate users.
Eddy
------------------------------
From: [EMAIL PROTECTED]
Crossposted-To: comp.unix.bsd.freebsd.misc
Subject: Re: NFS problems with Linux 2.2.x server, freebsd client
Date: Mon, 22 Mar 1999 10:41:12 GMT
In article <7d3jih$11r8$[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (Leslie Mikesell) wrote:
> In article <7d1djn$833$[EMAIL PROTECTED]>,
> <[EMAIL PROTECTED]> wrote:
> Keep in mind that you aren't going to see any difference in speed
> unless you are loading more than a couple of T1's, since Apache
> has no trouble doing that on a so-so pentium.
Actually, the people working on this project are currently pushing roughly
250 Megabits/sec through a rather large cluster of boxes (half of which is
internal, the other half outgoing), which pretty much saturate two outgoing
T-3's at peak. The main bottleneck at this moment is the speed with which
the internal traffic is routed outwards, hence the work on the efficent proxy
mechanism.
> There are some other free solutions to this as well. The one from
> http://www.eddieware.org/
> does load balancing and dead server detection.
I took a look at the page, but unfortunately that's a load-balancing solution
which works primarily by redirecting requests, with Apache as a front-end.
That's simply not going to work very well in this case. The site in question
already has dedicated load-balancing routers and a pretty robust distributed
network -- it's access to the data itself that's causing the bottleneck
(several Terabytes of data). The problem is essentially one of moving from a
constrained topology on the storage side of things (few NFS servers) to a
more distributed model (clusters of local storage controllers) -- while
*decreasing* the complexity of the (already-distributed) front-end machines,
which handle the network side of requests.
This is a pretty exotic problem, so I don't really expect there to be any
off-the-shelf solutions. I do appreciate the suggestion, though :).
> >If you're not running low on RAM, one way to cut that time by
> >over 99% is to create an MFS (FreeBSD) or RAMDISK (Linux), and write the
> >logfiles there.
>
> A good idea, but I don't want to lose these.
I can't speak one way or the other to the Linux side of things, but I've
never had a problem with data loss using FreeBSD'S MFS -- except in the event
of reboot, obviously. At my previous job, I was rotating logfiles on a
five-minute basis and so risk of data loss wasn't deemed important (at worst
only five minutes of logs would have been lost if a machine punted).
So as not to be misleading, I've never actually used an MFS _in_
_conjunction_ with logfile storage -- logfiles were stored locally in the
above scenario, and the MFS was used for other purposes. It just so happened
that we never needed to get the extra 3% performance increase that MFS would
have provided, as we always needed to add extra machines to our clusters for
other reasons anyway.
> I'm considering that. Perhaps a syslog type network write (I would
> rather lose an occasional line than bog things down if the
> file writer can't keep up.)
Please let me know how that works out, I'd be interested in the results.
> Les Mikesell
> [EMAIL PROTECTED]
-Bill Clark
============= Posted via Deja News, The Discussion Network ============
http://www.dejanews.com/ Search, Read, Discuss, or Start Your Own
------------------------------
From: [EMAIL PROTECTED] (Mike Jagdis)
Subject: Re: Dial up Linux
Date: 23 Mar 1999 12:38:25 GMT
In article <7d6t8h$pjb$[EMAIL PROTECTED]>, [EMAIL PROTECTED] wrote:
>
>WHY, oh WHY can I not connect at a faster speed than 24K?? I use USR V.90
>(NOT Winmodem) on BOTH ends but the speed is still 24K.
You will not get V.90 speeds unless the server end has a digital
line from the exchange (such as ISDN) with local modem emulation
hardware on your digital switch which can handle incoming V.90.
If both ends are analogue you can't get better than V.34. If your
phone line is less than perfect you might not even get that (and nor
would you get V.90 on it in that case).
>I just need to be able to connect @ 40+K. I am using this box for POP3 and
>Internet access just like any ISP would do. I work for a non-profit that
>needs to give internet access to internal staff.
The USR/3-Com Courier I-Modem *might* be able to do head-end V.90
on an ISDN2 line. If so (check first!) you *could* put in ISDN2
with an I-Modem at the server end - but if your phone line is
less than perfect you still might not get V.90 speeds (and remember
that V.90 is asymmetric too). It's probably worth considering
ISDN at both ends with a couple of cheapo passive ISDN cards.
Mike
--
A train stops at a train station, a bus stops at a bus station.
On my desk I have a work station...
.----------------------------------------------------------------------.
| Mike Jagdis | Internet: mailto:[EMAIL PROTECTED] |
| Roan Technology Ltd. | |
| 54A Peach Street, Wokingham | Telephone: +44 118 989 0403 |
| RG40 1XG, ENGLAND | Fax: +44 118 989 1195 |
`----------------------------------------------------------------------'
------------------------------
** 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
******************************
