Linux-Networking Digest #909, Volume #10         Sun, 18 Apr 99 20:13:36 EDT

Contents:
  Mail virus filter (Jure Krasovic)
  Re: 3c509B and 3c905B in same box ("Jan Johansson")
  Re: MS PPTP  w/ Linux? (Kenneth Springhetti)
  Re: Newbie Seeks Networking Direction (Michael J Surette)
  Re: DHCP server, PC has two NICs (Stephen Carville)
  ppp problem (fairly simple question i think) (Mogul 55)
  Re: Dial on Demand Linux Server (jeff)
  Re: dhcp and 255.255.255.255 a new twist (Stephen Carville)
  Re: Beowulf clusters ("Randy")
  IRQ printing & PLIP with 2.2.x kernel (Brian Hall)
  Re: minicom - users cannot create lock files (Clifford Kite)
  Memory, darn it (Jon McKee)
  Using smbmount ("Jing Duan")
  Re: Web Mail Server for Linux? (Jim Roberts)
  Re: Cable Modems ("--==[bolMyn]==--")
  Re: ..damn ("Gary Cameron")
  Re: NT faster than Linux? ("Larry Brasfield")

----------------------------------------------------------------------------

From: Jure Krasovic <[EMAIL PROTECTED]>
Subject: Mail virus filter
Date: Sun, 18 Apr 1999 18:55:46 +0200

This is a cryptographically signed message in MIME format.

==============ms3C2CB6DACBB58D724A36BE55
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Hi !

Is there someone who knows the software checks atachment files in mail
for viruses. I would like to run it under linux.

Thanks and best regards!

                    Jure

==============ms3C2CB6DACBB58D724A36BE55
Content-Type: application/x-pkcs7-signature; name="smime.p7s"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="smime.p7s"
Content-Description: S/MIME Cryptographic Signature
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==============ms3C2CB6DACBB58D724A36BE55==


------------------------------

From: "Jan Johansson" <[EMAIL PROTECTED]>
Subject: Re: 3c509B and 3c905B in same box
Date: Sun, 18 Apr 1999 18:56:27 +0200

PCI cards are _always_ PnP. IRQ is determined by the system bios, and the
Adress is configured on device initialization.



------------------------------

From: Kenneth Springhetti <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup,comp.protocols.tcpip
Subject: Re: MS PPTP  w/ Linux?
Date: Sun, 18 Apr 1999 20:53:26 +0000

Hi, I noticed you are working with Linux PPTP. Can you point me in the
right direction? I need to figure out how to connect my linux box to the
PPTP server at work. I have read lots of technical info on the protocols,
but I still cannot for the life of me get the software to work. I am sure
this is a smiple thing to do, but I cannot get it to work properly.

I have a cablemodem internet connection, I have ppp installed, and I have
the pptp package. Is there a location I can find this info at?

Thanks
-Ken Springhetti
[EMAIL PROTECTED]


Sami YOUSIF wrote:

> Does anyone know if there is way to connect an MS client to a linux
> server using the PPTP used by MS OS's in VPN mode?
>
> I do know that it is possible to have a linux-linux VPN; the quesion is
> can it be done when one of the endpoints is an MS product....hopefully
> using the built-in PPTP....
>
> --
> -
> ---
> Sami Yousif
>
> mailto:[EMAIL PROTECTED]
> http://www.mav.net/teddyr/syousif/ Personal Page
> http://www.alug.org/   Amarillo Linux Users Group
>
> [eMail sent to any of my addresses is subject to the Conditions outlined
>
> in http://www.mav.net/teddyr/emailtos.shtml]
>
> [Note: I no longer support ARNet (arn.net) as an ISP nor WTAMU
> (wtamu.edu) as an educational institution nor LEK (lektech.com) as a
> Computer Supplier] {http://www.mav.net/teddyr/access/banned.shtml}
>
> [heard somewhere: "You have the right to remain clueless. Anything you
> know may be used against you in a court of law"]
>
> Another day, so many more LARTS to go. [BOFH, BUFH, JOAT]
>
> "Understanding is a three edge sword: Our side, Their Side, and the
> Truth" Babylon 5
>
> <time is on my side>
>
> Tuesday, January 19th 2038, 03:14:07 UTC: Are YOU Ready?


------------------------------

From: Michael J Surette <[EMAIL PROTECTED]>
Subject: Re: Newbie Seeks Networking Direction
Date: Sun, 18 Apr 1999 18:18:18 -0400

Earl Beck wrote:

> Hi All,
>
> Heres what I have:
> A Packard Bell 486 SX2 50 Mhz (I know I know) running Win 3.11 and a
> P133 running Win98 that I built. Not too shabby for a poor college
> student!!! Both have RedHat 5.2 installed. I am wondering what would be
> the best way to network them and using what OS? I would like to
> incorporate at least one computer as a Linux Workstation and or Linux
> Server. Also need to know what kind of cheap NICs will be compatible
> both in Win 3.11 or 98 and Linux and whether I should run em through a
> bridge or just pier to pier. I can do the cabling myself so thats not a
> concern. Thanks on any guidance provided......!
>
> Humbly Curious!!!

I've had no difficulty at all with any isa ne2000 compatible cards under
any of these os's.  If you're not going to run more than the two stations
then a direct cross-connected cable (pin 1 -> pin 3 & pin 2 -> pin 6 ...
one pair,  same in the other direction) will pull you through.  You can
even go full duplex if you want.

HTH


------------------------------

From: Stephen Carville <[EMAIL PROTECTED]>
Subject: Re: DHCP server, PC has two NICs
Date: Sun, 18 Apr 1999 11:15:19 -0700

Derek Shaw wrote:
> 
> The ISC website says "Linux: Partial functionality: only systems
> with a single network interface are supported" in regards to
> their V1 (stable) version of the DHCP server
> (http://www.isc.org/dhcp-v1.html).  Unfortunately, this is the
> situation I am faced with.
> 
> I am reluctant to use beta code (their v2 does not appear to be
> out of beta yet).  Does anyone have any recommendations, or
> experience with the v-2 code?

Use the 2.0 code.  I am using it right now on a test bed system for dhcp to
(mostly) NT/95 clients.  SO far teh only glitches have been bugs in my perl
code to glue dhcp to dns.  The dhcpd server has performed flawlessly.

-- 
Stephen Carville
[EMAIL PROTECTED]
====================================================
It's all right to have geniuses build systems for use by idiots, but 
the path from laboratory to marketplace needs to go through the 
proving ground of prudent engineering.
                                        Peter Coffee

------------------------------

From: [EMAIL PROTECTED] (Mogul 55)
Subject: ppp problem (fairly simple question i think)
Date: 18 Apr 1999 15:48:08 GMT

ok i just used linuxconf to set up my ppp connection.  However it will only
activate when i boot.  in linuxconf there is a connect button but i click it
and nothing happens also it does not disconnect correctly from the same place. 
 when it loads at boot time the connection is fine and i am able to browse the
net with out a problem.  

Can someone please help
Thanks in advance
[EMAIL PROTECTED]

------------------------------

From: jeff <[EMAIL PROTECTED]>
Subject: Re: Dial on Demand Linux Server
Date: Sun, 18 Apr 1999 12:07:06 -0400

You'll need to set up IP forwarding and masquerading on your linux host
to allow your Win98 machines to dial out via your linux host.

Aladdin Evad wrote:
> 
> I'm trying to setup a Linux Server on an Old 486 to provide a Internet
> connection to all my Computer on my network. . I got it Samba up and working
> and I can see all my computers on the network. I also Got my Linux box to
> dial out and connect to my ISP. I wanted to know what do I have to do to
> make my Win98 machines use the Linux server to provide Internet access. Do I
> need to set up IP routing? Any info will be help full.
> 
> Thanks
> 
> Aladdin

-- 
jeff
being superstitious is bad luck!

------------------------------

From: Stephen Carville <[EMAIL PROTECTED]>
Subject: Re: dhcp and 255.255.255.255 a new twist
Date: Sun, 18 Apr 1999 11:24:42 -0700

Larry Ash wrote:
> 
> I have configured dhcpd on a RH 5.2 box with three network interfaces.
> (Don't ask, that part is a long story.)
> DHCP correctly sees the inbound interface and assigns an address
> apropriate for the inbound subnet.
> 
> It then needs to respond to the client sending to broadcast  address of
> 255.255.255.255.

Windows right? :-)

> The problem, the routing code wants to send this address out one
> interface. (Actually loopback always gets a packet also) It needs to go out,
> as a minimum, the inbound interface, or lacking that, all three. If I enter
> three route commands the machine selects one of the three (usually the one
> that the default route is one) and sends it there.
> 
> The question, is there a way to make a linux machine flood a broadcast to
> all interfaces.

I haven't tried this but maybe dhcrelay will work for you.  Have it relay
all requests on two of the ethernet interfaces to the IP address on the
third.  Then (I think) the broadcast will be properly handled.

-- 
Stephen Carville
[EMAIL PROTECTED]
====================================================
It's all right to have geniuses build systems for use by idiots, but 
the path from laboratory to marketplace needs to go through the 
proving ground of prudent engineering.
                                        Peter Coffee

------------------------------

From: "Randy" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.setup
Subject: Re: Beowulf clusters
Date: Sun, 18 Apr 1999 11:22:54 -0400

RedHat offers a product called Extreme Linux which is specifically set up to
install and maintain Beowulf Clusters.  I have no direct experience with it,
but coming from RedHat, it may very well be worth a look.

Bob Jones <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
>Anybody have any experience setting up and maintaining a Beowulf
>cluster? I'm attempting one at my school as a project, and I'd like to
>have any advice or input that you can give. I currently have 10 486DX2
>25MHz machines to use in the cluster; any advice on how I should
>properly set them up and all? (I have a slightly working configuration
>right now, but I want to incorporate the best ideas and use the best
>software and drivers that I can.) Thanks in advance for any advice/help!
>
>_    /
>  \  /
>   X
>  /   \_
>/
>



------------------------------

From: [EMAIL PROTECTED] (Brian Hall)
Crossposted-To: comp.os.linux.setup,
Subject: IRQ printing & PLIP with 2.2.x kernel
Date: Sun, 18 Apr 1999 17:13:36 GMT
Reply-To: [EMAIL PROTECTED]

I am currently running the 2.2.5ac6 kernel. Next time I reboot it will be
the 2.2.6ac1 kernel. I am trying to set up IRQ printing, in the hopes that
it will allow faster or lower CPU load PLIP networking with the laptop I
expect to arrive this week. I have not been able to configure the lp0 port
to use IRQ 7. I have tried the tunelp-0-2.1.131.tar.gz package, with no
effect; I still get an ioctl error:

../tunelp /dev/lp0 -i 7
tunelp: ioctl: Invalid argument
/dev/lp0 using IRQ -1

Suggestions? I would prefer to use PLIP networking between my laptop &
desktop for now; I don't want a PCMCIA headache to deal with. Plus I've
pretty much blown my budget for awhile with the laptop purchase anyway.

------------------------------

From: kite@NoSpam.%inetport.com (Clifford Kite)
Subject: Re: minicom - users cannot create lock files
Date: 18 Apr 1999 16:14:05 -0500

Peggy ([EMAIL PROTECTED]) wrote:

: I can't start minicom as a user, the error i rcv is "cannot create lock
: files"
: the user have access to /dev/ttyS0-1 and the mode is correct.
: i have also set the minirc.dfl files to be in the same group containing the
: users and the mode seem right
: -rw-rx-r--

The permissions on the lock file directory need to give users write
permission, or you can not specify a lock file directory in which case
there is no lock file created.  Alt-o, Serial port setup, B.


--
Clifford Kite <kite@inet%port.com>                       Not a guru. (tm)



------------------------------

From: [EMAIL PROTECTED] (Jon McKee)
Subject: Memory, darn it
Date: Sun, 18 Apr 1999 09:59:08 -0600

   Hello everybody,

I am having trouble getting Linux to recognize 128mb installed RAM.  I
have edited the lilo.conf file to include:
append="mem=128M"
but it still reports my total memory to be 64mb.

I have checked several FAQ's and searched the linux.org site but cannot
find any information on this subject.  Can somebody help?

Jon McKee

------------------------------

From: "Jing Duan" <[EMAIL PROTECTED]>
Subject: Using smbmount
Date: Sun, 18 Apr 1999 16:56:02 GMT

Hi,

I have problem on using smbmount.  I hope someone could help me out.

I am using Slackware 3.6, kernel 2.0.35 and Samba 2.0.2.  I have used kernel
2.2.4 and Samba 2.0.3, but the problem is the same.

I have two pcs, one for Linux and one for Win95.

I want to share the D drive in Win95 machine with Linux.  I name the Win95
machine as "main", the Linux as "hp" and the network as "home".  So, I have
hp.home and main.home on my network.  The D drive is named as "main-d"

I can use smbclient to access the d drive.  I type,

smbclient  //main.home/main-d

Even though it complains "startlmhosts: Can't open lmhosts file
/etc/lmhosts. Error was No such file or directory", it works after I type in
a password.  I can access to the d drive.

I type

smbmount //main.home/main-d /mnt

to connect my drive, but after I type in the password, it complains:

mount error: No such device
Please look at smbmount's manual page for possible reasons.

I have read the news in related newsgroups and tried different methods, but
I could not get it working.

Does anyone know what is wrong?

My Windows 95 is a plain standard alone OS.  Do I need some kind of server
for win95 to use the smbmount?

Your help will be most appreciated.

Jing Duan
[EMAIL PROTECTED]




------------------------------

From: [EMAIL PROTECTED] (Jim Roberts)
Subject: Re: Web Mail Server for Linux?
Crossposted-To: comp.os.linux.misc
Date: Sun, 18 Apr 1999 19:14:50 GMT

In article <[EMAIL PROTECTED]>,
        "Gray McCord" <[EMAIL PROTECTED]> writes:
> I run sendmail and Apache on my 2.0.35 Linux server and was wondering if
> anyone knew of a solution to permit viewing users' email via a web
> connection to the Apache server.  For those that have seen it, I'm looking
> for something similar to the MS Exchange Server facility that lets users
> access their Exchange mail via IE instead of using the Outlook client.
> 
> Thanks,
> 
> Gray
> 

As mentioned, AtDot is a good WebMail client. Go to 
cgiresource.com and have a look. Many choices.

-- 
Jim Roberts         Never enough time!
[EMAIL PROTECTED]

------------------------------

From: "--==[bolMyn]==--" <[EMAIL PROTECTED]>
Subject: Re: Cable Modems
Date: Sun, 18 Apr 1999 17:45:03 GMT

Hmm,

My peak download from @Home Comcast was 300+ Kb/s with a steady flow (meaning
it stayed there for a while rather than just hit 300 and than roll back).  The
best results I had so far downloading from Netscape's FTP site (that's where
I broke my speed record).

Eugene wrote:

> cable modem is an external device which connects to your computer via a
> network card
> From my experience, the peak download speed is about 100k/s, peak upload -
> 30k/s
> Here in Canada cable companies give cable modems for free for as long as you
> subscribe for the service (they take the cable modem back if you cancel it).
>

--
Bolek,

URL: http://www.bolek.com
e-mail: [EMAIL PROTECTED]




------------------------------

From: "Gary Cameron" <[EMAIL PROTECTED]>
Subject: Re: ..damn
Date: Sun, 18 Apr 1999 14:01:22 -0400

    Sure.  Just use Samba for the Win95/NT machines.  The other question is
how
dumb are the dumb terminals?  Do they at least have a CPU?  If so, is it at
least a 386?
If they are really just terminals, then they cannot be be tied into a
network in any meaningful way.  You would just use
a rocket port serial box to run them off the linux box.

WingNut4 wrote in message ...
>... obviously i hit the wrong button...
>
>anyway my end-of-line question was exactly how effective would running a
>linux server be to cater for approximately 50 computers?(half of which
would
>probably be running Win32, the other half would likely be dumb terminals)
>would there be any problems linking it up to other servers using another OS
>such as NT?
>
>Cheers
>
>
>



------------------------------

From: "Larry Brasfield" <[EMAIL PROTECTED]>
Crossposted-To: linux.samba,uk.comp.os.linux
Subject: Re: NT faster than Linux?
Date: Sun, 18 Apr 1999 23:32:38 GMT

Julian T. J. Midgley <[EMAIL PROTECTED]> wrote in message 
news:7f62e1$uk1$[EMAIL PROTECTED]...
> In article <U9tR2.73$[EMAIL PROTECTED]>,

The post to which I first responded in this thread
had claimed "NT is using over 4X the memory"
and made no other points except to conclude,
(incorrectly) that the NT and Linus tests were not
run on the same hardware.

> Larry Brasfield <[EMAIL PROTECTED]> wrote:
>
> >Depending on what "****" is, that may be overstated.
> >I don't see why they did not tell NTS to use 960 MB,
> >but it could be an adjustment that reflects differences
> >between how Linux kernel usage is reported and
> >what the maxmem parameter means to NT.  Still,
> >960 is close enough to 1024 that it takes a lot of
> >wishful thinking to make much of the difference.
>
> Indeed, if you had read the rest of this thread you would have
> discovered the other considerably more damning differences.

I found much of the thread that I could see (on my
ISP's news server with aggressive aging) to be
speculative or more questioning than conclusive.
Please notice my response was narrow in scope.
I addressed a factual mistatement with a factual
rebuttal.

> >>  Let them Test an EQUIVALENT system, hardware to hardware!!!!!!
> >
> >Under "Configuration and Tuning" they claimed:
> >    We used the same Dell PowerEdge 6300/400 to test
> >    both Windows NT Server 4.0 and Red Hat Linux 5.2
> >    upgraded to the Linux 2.2.2 kernel.
> >
> >Near the end of the report, Mindcraft states:
> >    Our test results should be reproducible by others
> >    using the same test lab configuration, the same
> >    Dell computer, and the software configurations
> >    and modifications documented in this report.
> >
> >I think their results deserve credence for what
> >they mean on that system with those benchmarks.
>
> Nonsense.  They had failed to configure the Linux system correctly,
> not only with respect to the memory, but in several more damaging ways
> discussed elsewhere in this thread.  In other words, they were not
> comparing a Linux Server with an NT server.  They were comparing a
> poorly configured Linux Server running inappropriate versions of the
> webserver with a highly tuned NT server.

Your rebuttal of my "nonsense" does nothing but
elaborate how something different might mean
something new.  Nobody has shown that the
systems tested do not perform as claimed.

> Their benchmark is in no way
> valid as a comparison of the performance of real webservers- it is
> only valid as a measure of the incompetence of Mindcraft.

Sounds like time for somebody to perform
some "real webserver" benchmarking.  It
should be easy for the NT side since it can
be presumed that Mindcraft's settings for
that OS are servicable, if not optimal.

One of the interesting outcomes of Mindcraft's
report is that they claim to have sought help
with the problems they perceived.  It would
be quite a coup to show that they got such
help and elected to ignore it or that advice
on setting up Linux servers with SMP and
RAID is effective and easy to find.

> >Any benchmarking effort is subject to bias in what
> >benchmarks are chosen, so the ~3x result might not
> >apply in applications that matter to most people.  I
> >would not be surprised, given the development
> >environment in which Linux gets tweaked, if it has
> >been optimized more for smaller systems.  But I
> >think it is a mistake to write off Mindcraft's result
> >as "rigged" without better evidence.
>
> There is evidence.  You appear to have decided to
> post before reading it.

You got me.  I responded to a narrow point without
first researching hundreds of posts that have gone
by over the last few weeks.  My comment about
evidence relates to the proposition that Mindcraft
intentionally crippled the Linux system or intentionally
misconfigured it.  The memory disparity issue does
not cut it as evidence.

> I am willing to stake a tenner than no-one will be able to reproduce
> that result if they run the same test on a machine of identical spec,
> but this time bothering to configure the Linux set-up properly, using
> the most appropriate software.

That would not be reproducing the result.  That would
be running a different test.  Reproducability means
that enough information has been given to allow
others to validate the "exact" same test by setting
up the same conditions and getting the same result.

> As someone who administers both NT and Linux webservers I can assure
> you that it is demonstrably fallacious to suggest that NT is faster,
> or makes a better server in any way whatsoever.

That is a very broad claim.  I don't have the experience
to know how those systems generally compare, but I
am amazed that anybody has enough to make such
a categorical statement.

> >(Bias warning: I am an ex-employee of Microsoft who does not believe
> >in the "Evil Empire" notion.  I am also a Linux fan and user, but not
> >a "convert".)
>
> Bias warning: I am one of the system administrators for the Free
> Software Foundation's machines at *.gnu.org.

I try not to infer bias from such facts.  I prefer
to evaluate what people say on the merits.

Interestingly, in the article cited by Mr. Tennent at
   http://lwn.net/1999/features/MindCraft1.0.phtml
one of the "Non-issues" mentioned was this:
    Some complaints have been raised about the
    test being run on a 4GB server, even though
    the Linux kernel, in its default form, can only
    use 960M of that. Patches can be applied to
    make 2GB available fairly easily. But, in any
    case, they claim that NT was limited (with the
    maxmem parameter) to 1G of memory, so this
    aspect of the test was fair. It would have been
    more straightforward of them, however, to have
    simply remove the other 3G from the system.

Perhaps, instead of castigating me for pointing
out that Taylor's issue, ("NT is using over 4X the
memory"), is false, you should thank me for being
among those willing to help keep the controversy
founded on facts.

--
--Larry Brasfield
Above opinions may be mine alone.
(Humans may reply at unundered [EMAIL PROTECTED] )




------------------------------


** 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
******************************

Reply via email to