Linux-Networking Digest #216, Volume #11 Thu, 20 May 99 16:13:51 EDT
Contents:
Two network cards ("Jim Scheffler")
Re: HTML based POP tools (Jim Richardson)
ip-routing / maquerading / dual internet gateway (Lieven Van Acker)
Firewall+DHCP (pump) Strange DHCP behavior HELP! (Francois Magnan)
Re: NFS and /usr (OldUncleMe)
Re: Eth0 and ppp0 on gw works. routing not working (Clifford Kite)
Problems compiling ppp-2.3.7 (Colin P. McNally)
File Corruption Problems on Samba ([EMAIL PROTECTED])
need help on webcam for remote monitoring (weihua sheng)
Re: PCI Modem - lost cause? ("Roberto Leibman")
Re: Modem i/o conflict problem ([EMAIL PROTECTED])
ftp server problems- Unable to load Interpreter, ([EMAIL PROTECTED])
Remote Printing ("LSI")
Re: RHL6, Samba 2.0.3 NO Workee! ("Celestino Okenve Ndo")
----------------------------------------------------------------------------
From: "Jim Scheffler" <[EMAIL PROTECTED]>
Subject: Two network cards
Date: Thu, 20 May 1999 13:08:28 -0400
I have @Home cable modem service that I would like to share over my
intranet. The problem that I am running into is that I have a two sided
network (intra and internet) and I can only get one NIC to work. I am
running Mandrake (5.2 w/KDE) Somebody told me that there is an RPM that I
need to load to use dual NIC's. Any help would be appreciated.
-Jim
------------------------------
From: [EMAIL PROTECTED] (Jim Richardson)
Crossposted-To: comp.os.linux.misc
Subject: Re: HTML based POP tools
Date: 20 May 1999 17:34:16 GMT
Reply-To: [EMAIL PROTECTED]
On Wed, 19 May 1999 16:57:50 +0300,
Mies, in the persona of <[EMAIL PROTECTED]>,
brought forth the following words...:
>
>>I'm fairly new at Linux (2.2) and have a couple of questions regarding
><snip!>
>>2- HTML based Clients
>>
>
>
>Well you could try IMP, I recall it handles IMAP and POP.
>It was http://www.horde.org/ I think...
>
>
>
Linux Journal has an article on this very subject in this months
issue. (carried over from last months.)
--
Jim Richardson
www.eskimo.com/~warlock
All hail Eris
"Linux, where do you want to go tomorrow?"
------------------------------
From: Lieven Van Acker <[EMAIL PROTECTED]>
Crossposted-To: be.comp.os.linux,nl.comp.os.linux
Subject: ip-routing / maquerading / dual internet gateway
Date: Thu, 20 May 1999 19:40:02 +0200
Hello,
I'm a network admin on a public high school, and I'm faced with this
problem:
CENTRAL LINUX GATEWAY
| | | | |
| | | | ---> eth0: Internet link 1: leased line, public fixed IP
addr (ISP1)
| | | ------> eth1: Internet link 2: cable-modem, dhcp configured
(ISP2)
| | ---------> eth2: DMZ
| ------------> eth3: Internal subnet 1, private address range,
masqueraded
---------------> eth4: Internal subnet 1, private address range,
masqueraded
What we want to reach is explained simply:
All incoming traffic ( to the gateway and DMZ ) has to come in via
internet link 1 (Leased Line).
All outgoing traffic except SMTP traffic has to go via internet link 2
(Cable-Modem).
All outgoing SMTP traffic has to be routed via the Leased Line.
Outgoing traffic from our private subnets is masqueraded before is
leaves via eth1 and is blocked to go via eth0.
Outgoing SMTP traffic is generated via a mail exchanger, located in the
DMS (with public DNS records)
THE TOOLS:
ipchains to control masquerading and blocking of IP traffic
static routing
THE PROBLEM:
How can I tell the gateway there are two interfaces to reach the
internet ?
Do a have to use a routing deamon i.s.o. static routing ?
Thanks in advance,
Lieven.
------------------------------
Subject: Firewall+DHCP (pump) Strange DHCP behavior HELP!
From: [EMAIL PROTECTED] (Francois Magnan)
Date: Thu, 20 May 1999 18:37:24 GMT
Hi,
I am trying to install a strong firewall script in my Linux
router. It is based on
http://rlz.ne.mediaone.net/ftp/firewall/rc.firewall.ipchains
I have RedHat5.2 with some RedHat6.0 rpms: kernel2.2.5, initscripts,
pump, dhcpcd, ipchains, ... (everything needed by the 2.2.5 kernel to
work).
I am using "pump" (in RedHat6 pump is used to call dhcp) to get my IP
address. I therefore added a line in my
/etc/sysconfig/network-scripts/ifdhcpc-done
file to call my rc.firewall script. But it doesn't get called!!!
Also, my script
(see http://rlz.ne.mediaone.net/ftp/firewall/rc.firewall.ipchains)
uses /etc/dhcpc/hostinfo-eth1 to determine it's ip address. But this
file doesn't contain the correct ip address.
It seems that dhcpcd (or pump) don't update the file
/etc/dhcpc/hostinfo-eth1
to give the correct ip address.
What is going on???
Thank you,
Francois Magnan
--
______________________________________________________
Francois Magnan
Departement de Mathematique & Statistiques
Universite de Montreal
email: [EMAIL PROTECTED] (MIME, NeXTMail Ok!)
------------------------------
From: [EMAIL PROTECTED] (OldUncleMe)
Subject: Re: NFS and /usr
Reply-To: [EMAIL PROTECTED]
Date: Thu, 20 May 1999 18:40:19 GMT
It was: Wed, 19 May 1999 18:35:58 GMT and with STARTLING insight,
"[EMAIL PROTECTED] (Jim Roberts)" posted "Re: NFS and
/usr" to "comp.os.linux.networking" :
-->In article <[EMAIL PROTECTED]>,
--> [EMAIL PROTECTED] writes:
-->> With RH50 if host allows root access ('/') I can read any directory. I
-->> want this. With RH60, I cannot read /usr (or /root and some others). I
-->> -can- read /etc, among others.
-->> If I mount host:/usr as /something, I -can- read /usr. Anyone else seen
-->> this?
-->It sounds like /usr is on another partition or disk. This is normal operation
-->for NFS. It works this way on solaris and SCO also.
-->I'm afraid I don't have a workaround for you. I have the same situtation
-->on my solaris file server.
-->Mabe someone else could help us both?
-->Jim Roberts
Not able to share root ( / ) on Samba, I created a soft link to it on the
root tree called 'bob', just for fun. Now I can share bob and the
connecting machines see all the root structure, uh, including bob, which
can be opened to show the root structure again, including bob.... But it
does work. May work for /usr on another drive with NFS, worth a try:
mkdir /bob #or some other name
ln -s /usr /bob
/ts
------------------------------
From: kite@NoSpam.%inetport.com (Clifford Kite)
Subject: Re: Eth0 and ppp0 on gw works. routing not working
Date: 19 May 1999 17:37:18 -0500
[EMAIL PROTECTED] wrote:
: I want "main" to actroute add -net 192.168.1.0 netmask as the router for
: the 192.168.1.* network (eth0).
: All these machines should be able to use "main" to hope onto the 'net.
: Main's eth0 config looks like this:
: route add -host 158.152.71.154 dev lo
: ifconfig eth0 192.168.1.1 netmask 255.255.255.255
: route add -net 192.168.1.0 netmask 255.255.255.0 eth0
: route add default gw 158.152.71.154 dev eth0
: Dev2's config is as follows:
: ifconfig eth0 192.168.1.3 netmask 255.255.255.255
: route add -host main dev eth0
: route add default gw main dev eth0
Neither dev2 or any other machine on your LAN except main is going
to get packets back from the Internet without using IP masquerading.
The return packets will be bound for reserved addresses and they'll be
dropped into /dev/null by the first properly configured router they hit.
If main has IP forwarding compiled into the kernel and turned on, then
main will send them on to the Internet but the result is effectively
"poof" they are gone forever.
--
Clifford Kite <kite@inet%port.com> Not a guru. (tm)
/* Editing with vi is a lot better than using a huge swiss army knife. */
------------------------------
From: [EMAIL PROTECTED] (Colin P. McNally)
Subject: Problems compiling ppp-2.3.7
Date: 19 May 1999 20:03:43 GMT
I recently upgraded to kernel 2.2.9 and glibc 2.0.6. After running the
configure script of ppp-2.3.7 and running the make command i get the
following error:
main.o: In function 'main':
/usr/src/ppp-2.3.7/pppd/main.c:460 undefined refrence to '__sigsetjmp'
/usr/src/ppp-2.3.7/pppd/main.c:689 undefined refrence to '__sigsetjmp'
/usr/src/ppp-2.3.7/pppd/main.c:788 undefined refrence to '__sigsetjmp'
and then a few make errors as it exits
It looks like i'm missing something, what is it??
--
______________________________________________________________________________
Colin McNally "I'm just this guy eh?"
"Aristotle once said "The only true knowledge is the knowledge that
you know nothing." however he also thought that frogs are spontaniously
generated from mud." - Colin McNally '96
"Knowledge is the only elegence."-Edward Emerson
____________________________________________________________________________
------------------------------
From: [EMAIL PROTECTED]
Subject: File Corruption Problems on Samba
Date: Wed, 19 May 1999 22:08:38 GMT
Hi All,
We're fairly new to Linux and Samba and have a baffling problem.
We're running Samba 2.0.3 as a file and mail server. This is running on
RedHat Linux 5.1.
This is on a Celeron 300 with 32 Mb RAM. We're using generic NE2000 ISA
NICs and 10BaseT UTP, connecting with 4 NT Workstation 4.0 and 10 Win
95 boxes. This is all set up as a workgroup. The Linux box is also
being used for internet access through a separate Linux firewall
installed on a separate stand-alone box.
Maximum cable run between any two machines, through the hub, is about
50 m. The hubs (2) are located in a wiring closet with the telephone
PBX stuff.
There is an intermittent file corruption problem when saving to and
retrieving from the Samba server shares. The file type doesn't seem to
matter. It also does not seem to matter which shares, i.e. pub or
private. The primary applications being used are Excel 97 and
Wordperfect 7 and 8.
The mail system seems to be OK.
This problem does not occur if the Linux box is disconnected from the
net and the NT boxes substituted as the file servers.There were no
problems running tests from 20 PCs against two NT workstations.
Samba was built with -DTCP_NODELAY and -DUSE_MMAP options ... no
improvements noted.
Windows 95 clients are more likely to be affected (they get an "Abort,
Retry, Fail cannot read disk ?" error ... retry doesn't allow a resume)
than Windows NT clients (which simply get file transfer errors and
corrupt the file). I could get about 5 NT clients running a
copy-to-the-server-and-get-it-back-and-compare-to-the-original test,
but only around 3 Windows 95 clients. Problem is much worse (occurs
earlier) if Samba debug level > 2.
We don't know if this is a Samba problem, a TCP problem, or a network-
busy problem.
I've read (on
http://www.germanynet.de/teilnehmer/101/69082/samba.html#4.3)
that /etc/hosts should be set up with the localhost entry as line 1 and
the local PCs entry as line 2. This was *not* the case (local PCs
entry was down around line 20) and I have rearranged the /etc/hosts
file as well as removed about 20 lines of comments. Also added
SO_SNDBUF=4096 and SO_RCVBUF=4096 parameters. Haven't had a chance to
try the test again, since these changes were made.
Also, there is no lmhosts file on the Samba server - would it help?
Mike Luckham
Jim Humphrey
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
From: weihua sheng <[EMAIL PROTECTED]>
Subject: need help on webcam for remote monitoring
Date: Thu, 20 May 1999 14:35:28 -0400
Hi, all,
I am working on a project on "vision through Internet".
I have a mobile robot, its operating system is Linux, we try to
mount a camera on it and try to transfer the camera output image to any
other PC(windows 98) on the Internet. So I want to make a strategy for
the project.
Can anyone familiar with this area give me some advice? As I know,
the Webcam in Internet is one kind of application. So, if we try to make
the mobile robot with linux as a Web Server, and do what the Webcam
does, maybe it can work. Can someone give me some reference on this?
Thanks a lot!
weihua
------------------------------
From: "Roberto Leibman" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.setup
Subject: Re: PCI Modem - lost cause?
Date: Wed, 19 May 1999 16:23:12 -0700
David A. Rogers <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> On Wed, 19 May 1999 09:50:51 -0700, Roberto Leibman
> <[EMAIL PROTECTED]> speaketh saying:
> >
> ><[EMAIL PROTECTED]> wrote in message
> >news:7hulmp$1f6$[EMAIL PROTECTED]...
> >> In article <7hgvds$n38$[EMAIL PROTECTED]>,
> >> Roberto Leibman <[EMAIL PROTECTED]> wrote:
> >> On the other hand, "what the modem is expecting the OS to do" also
raises
> >a
> >> philosophical disagreement. It's not the OS's responsibility to be
part
> >> of a peripheral. For this reason, probably nobody who *can* would be
> >working
> >> on it.
> >
> >I have to disagree with this, as the constraint in computing power moves
> >around, from i/o bound to processor bound to memory bound to disk bound
to
> >bandwith bound, the "right" partitioning philosophy changes as well. I
think
> >we are at a time where we most people have a lot of extra capacity in
their
> >CPU's (most of the time anyway), it makes engineering sense to download
some
> >of the processing chore to the less stressed CPU, particularly if this
can
> >bring the price of the peripherals down. System partitioning does not
have
> >hard rules such as "It's not the OS's responsibility", granted modularity
> >should be sought after, but other things do enter the equation. Also,
> >remember that 3COM (and others) are in business to make money, not to
cater
> >to "fringe" markets, I don't blame them for producing the WinModem, It's
a
> >GREAT idea on their part, I do blame computer makers for not clearly
stating
> >that its what in the box when you buy it clearly stating that you *want*
to
> >install Linux on it! We might be unhappy about it, but think of how many
> >people have gotten access to the net with the savings that WinModem's
> >produce!
>
> I don't buy it. I've got a 133 and I want all of those cycles
> being used for their intended purpose. Even if I had a 450, I don't want
some
> rube-goldberg peripheral slowing it down.
Free market, choice, isn't it great?
--
Roberto Leibman
Talaria Research, Inc.
http://www.talaria.com
Cxi tioj opinioj ne necese estas la opinioj de la administrantaro
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Modem i/o conflict problem
Date: Wed, 19 May 1999 22:42:59 GMT
Thanx for the reply,
I tried that also. One more thing I have noticed is if there is
any conflict with /proc/ioports it reports (see /proc/ioports).
I tried all the 4 COM ports, no luck
thanx
krishna
In article <H9D03.1525$[EMAIL PROTECTED]>,
"Curt" <[EMAIL PROTECTED]> wrote:
> Remove the board, and take a look at /proc/ioports to see what already
has
> those addresses taken.
>
> <[EMAIL PROTECTED]> wrote in message
news:7hun76$7h7$[EMAIL PROTECTED]...
> > HI,
> > I am facing a probelm while installing(isaconf) a modem(Digicom 56K
> > D-F-V-SP DI3658) on RH6.0 the following is the error message :
> >
> > =================
> > /etc/isapnp.conf:118 -- Fatal - resource conflict allocating 8 bytes
of
> > IO at 0x3E8 (see /etc/isapnp.conf)
> > /etc/isapnp.conf:118 -- Fatal - IO range check attempted
> > while device activated
> > /etc/isapnp.conf:118 -- Fatal - Error occurred executing request'
> > <IORESCHECK> ' --- further action aborted
> > ====================
> >
> > The PNPDUMP returns 4 possible IO address which are same as that of
> > ttyS0-ttyS3. I tried all of them, but the result is the same.
> >
> > Any Help ?
> >
> > thanx in advance
> > krishna
> >
> >
> > --== Sent via Deja.com http://www.deja.com/ ==--
> > ---Share what you know. Learn what you don't.---
>
>
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
From: [EMAIL PROTECTED]
Subject: ftp server problems- Unable to load Interpreter,
Date: Thu, 20 May 1999 18:42:17 GMT
Hello, how is everyone?
I am having trouble configuring my Linux machine as an FTP server.
Whenever I try to make a few hundred connections,I start getting the
"Unable to load interpreter" message. This appears to be because I run
out of memory.
Strangely enough, it never seems to use the swap space ... It always
shows like maybe 4 blocks being used.
I have 256 MB of RAM... Can anyone give me a hint on how to configure
this thing? I need to be able to support hundreds or even thousands of
FTP connections at once. I guess the way it works normally is that
inetd spawns an ftpd process for each connection. I did increase the
number of connections in the inetd.conf file to 5000 (nowait.5000).
Any help would greaty appreciated ...
Thanks,
Will H.
[EMAIL PROTECTED]
--== Sent via Deja.com http://www.deja.com/ ==--
---Share what you know. Learn what you don't.---
------------------------------
From: "LSI" <[EMAIL PROTECTED]>
Subject: Remote Printing
Date: Thu, 20 May 1999 15:01:42 -0400
Assume that I cannot use printtool to configure a remote printer and must do
it manually. I have set up the entries in the /etc/printcap to look like so:
printer7:\
:sd=/var/spool/lpd/printer7:\
:mx#0:\
:sh:\
:rm=lsi:\
:rp=printer7:
where printer7 is a printer on a SCO machine and I created the
/var/spool/lpd/printer7 directory. When using lpr -Pprinter7 the files are
spooled to the /var/spool/lpd directory and are never forwarded to the SCO
machine. On a local printer that is set up I noticed that the files are
spooled the the printer directory under /var/spool/lpd/*. lpc status
printer7 shows that the spooler and daemon are enabled and that there is a
file spooled. One question is - are the spooled files SUPPOSED to end up
under the individual printers' directory under /var/spool/lpd/*? What else
must you configure manually besides the printcap file to ensure remote
printing? Thanks in advance for any responses...
------------------------------
From: "Celestino Okenve Ndo" <[EMAIL PROTECTED]>
Subject: Re: RHL6, Samba 2.0.3 NO Workee!
Date: Wed, 19 May 1999 21:31:33 +0200
Steve Smoogen wrote in message ...
>Hi what kind of problems are you experiencing? (Text errors would be the
>most helpful.) Thanks
>
>
>On Mon, 17 May 1999 11:58:36 -0700, W.A. Scheer <[EMAIL PROTECTED]> wrote:
>>I've done a 'workstation' load of RHL and added SAMBA using RPM. I've
>>tried everythiing under the sun and can't seem to get the danged thing
>>working!
>>
>>The smbclient util seems to work OK, but I can't use smbmount worth a
>>darn. Taking the opposite approach, setting up the smb.conf doesn't get
>>me any MS98 readable shares - they don't even show up in the browsw
>>list.
>>
>>This is my first REAL try at using SAMBA. Any pointers? Suggestions? is
>>this 'broken' under RHL6? I'm using the following books as my
>>references:
>>
>>1) Teach Yourself SAMBA in 24 Hours (Sams)
>>2) Samba (SSC)
>>3) Linux Network Toolkit (IDG)
>It all seem very straightforward - but 'no go'
>>TIA
>>W.A. Scheer
--
>SJS -- Red Hat Technical Support
>[Please be aware I cannot always answer email directly emailed at me. I
> try to answer on the news groups for more people to see the information
> and correct me if I am wrong :)]
Try to edit smb.conf.
Copy smb.con.default to smb.conf save the old one. Then try to modify by
hand
following some tips included in man and books. After three tryings you will
get something (printer or directories or mounted devs visible). If no
visible, from clients try to give the path \\your_linux_machine_name
Be sure linux can resolve names or edit hosts file in /etc.
Good Luck.
------------------------------
** 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
******************************
