Linux-Networking Digest #691, Volume #11         Sun, 27 Jun 99 12:13:39 EDT

Contents:
  Synchronizing Samba partition & W/95 box ([EMAIL PROTECTED])
  Re: Netscape, POP3 server, DNS ("Carl R. Friend")
  Re: RH 6.0 and netatalk (Rod Smith)
  Re: Remote fsck possible for Linux? (Peter F. Curran)
  clustering under linux? (Dionysus)
  Mirc linux mandrake dcc send ("ICON")
  Re: Linux PPP & Leased Line problem (Rob van der Putten)
  Re: LinkSys NIC LNE100TX:  Anybody have any problems with this card? (Rod Smith)
  Re: LinkSys NIC LNE100TX:  Anybody have any problems with this card? ("Tony C")
  Re: tcp_timestamps causing massive network slowdown problems under 2.2.x ("Gene 
Heskett")
  How to use NT Printer & files from Suse 6.1 ([EMAIL PROTECTED])
  Re: SOCKS5 and RH6 (Nathan Valentine)
  Re: PROXY (Frank Hahn)

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

From: [EMAIL PROTECTED]
Subject: Synchronizing Samba partition & W/95 box
Date: Sun, 27 Jun 1999 09:30:17 -0400

Hi,
  I need to do some dual Windows/Linux development, and decided the
obvious option
was to maintain the code on the Linux system, and use Samba to mount
that partition
as a drive from the Win/95-SR1 box.

  However, when I modify files via Linux, the changes don't appear on
the W/95 box
for several minutes.  Apparently, Windows is caching the file, and
doesn't know it
should refresh.

  Is there a setting I've obviously missed to cause samba to announce
that files have changed
and clients should refresh, or something I should be doing on W/95 to
force refreshes?

  TIA,
    Clif
--
........................... Clif Flynt ..........................
... Tcl/Tk for Real Programmers - Academic Press Professional ...
.... http://www.cflynt.com  ............   [EMAIL PROTECTED]   ....
. In theory there is no difference between theory and practice  .
........................ In practice, there is. .................



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

From: "Carl R. Friend" <[EMAIL PROTECTED]>
Subject: Re: Netscape, POP3 server, DNS
Date: Sun, 27 Jun 1999 10:06:12 -0400

George Lane wrote:
> 
> The problem: When the server is not dialed up and I try to fetch the
> downloaded messages from my server, using Netscape Navigator 4.08
> running on a Linux workstation. Clicking Messenger Mailbox to download
> the messages, Netscape connects to my server, but it also seems to
> generate some type of DNS request, because my server dials up the ISP,
> thus defeating the purpose of having the server in the first place :)

   What you might try doing is hanging a tcpdump session on the diald
snoop device to see exactly what is waking diald up in the first place.

   I run a similar setup on my home lan and don't have your problem,
but I'm also authoritative (in a DNS sense) for my entire LAN. Your
Netscape client may be looking for your "outside" address instead of
your "inside" one, bypassing your DNS, and waking diald to get the
address via DNS.

-- 
+------------------------------------------------+---------------------+
| Carl Richard Friend (UNIX Sysadmin)            | West Boylston       |
| Minicomputer Collector / Enthusiast            | Massachusetts, USA  |
| mailto:[EMAIL PROTECTED]                |                     |
| http://www.ultranet.com/~crfriend/museum       | ICBM: N42:22 W71:47 |
+------------------------------------------------+---------------------+

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

From: [EMAIL PROTECTED] (Rod Smith)
Subject: Re: RH 6.0 and netatalk
Date: Sun, 27 Jun 1999 14:06:42 GMT
Reply-To: [EMAIL PROTECTED]

[Posted and mailed]

In article <[EMAIL PROTECTED]>,
        "Mark" <[EMAIL PROTECTED]> writes:
> I recently upgraded from RH 5.2 to RH 6.0, I had netatalk installed, and had
> my linux box running as a print server for my LAN, after the upgrade I ran
> into a error when i tried to print from any of my mac's, the error prints
> out on the page instead of my print job.
> 
> Error: /invalidfont in -dict-  Operand stack:  Courier-Bold  --dict:12/25-- 
> 
> btw i am running netatalk-1.4b2+asun2.1.3-4 and I couldn't install the dev
> package because of conflicts with the built in appletalk support in RH 6.0

This isn't a netatalk problem; it's a Ghostscript problem.  I had somewhat
different font problems with the Ghostscript in RH 6.0.  I solved them by
replacing both the Ghostscript binaries and the Ghostscript fonts package
with the ones I'd been using before, which I obtained from:

http://www.users.dircon.co.uk/~typhoon/

-- 
Rod Smith
[EMAIL PROTECTED]
http://www.channel1.com/users/rodsmith
NOTE: Remove the "uce" word from my address to mail me
Author of _Special Edition Using WordPerfect for Linux_, from Que;
see http://www.channel1.com/users/rodsmith/books.html

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

From: [EMAIL PROTECTED] (Peter F. Curran)
Subject: Re: Remote fsck possible for Linux?
Crossposted-To: comp.os.linux.setup,comp.os.linux.hardware,comp.os.linux.misc
Date: Sun, 27 Jun 1999 13:30:34 GMT

In article <[EMAIL PROTECTED]>,
        md5¤Ï¼s§[EMAIL PROTECTED] (Abe Lin) writes:
>Hi, guys.
>For Solaris we can get a console remotely in this setup in the
>*locked* machine room 20 kms away from us:
>
>1.One NT box runs pcanywhere. Serial(s) go to two Solaris machine.
>2.Using pcanywhere from home. Do whatever you like to save the machine
>when it's done. Single user mode. fsck. Anything.
>
>Do we have something for Linux? (With a Linux only solution...)
>Cannot find a useful document on this. I'll keep on dejanews search,
>but nothing yet showed up....
>
>TIA a lot.

There are HOWTO or other docs concerning serial consoles I 
believe.  You could then run the serial line to another
box which has dialup access.  Using a serial terminal program
like "minicom", you should be able to do what you describe.



-- 
     Peter F Curran
     Rensselaer Polytechnic Institute


dough knot male: [EMAIL PROTECTED]
Use address in Organization line, finger
for PGP key.  Antispaam test in progress.


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

From: [EMAIL PROTECTED] (Dionysus)
Subject: clustering under linux?
Date: Sun, 27 Jun 1999 12:39:23 GMT

has anyone had any experience with linking a small number (i.e. up to
half a dozen) 486s together to form a cluster? If so, what software
did you use, MOSIX, Beowulf or something else? I have heard a few
people recommend both but only in very vague terms (as they were short
conversations, and i don't even know if beowulf is a version of linux,
like MOSIX, or a linux program that will run under any distribution).
Anyway, all advice, tips etc will be greatly appreciated. I am
intending to start with 2 computers and gradually add more to the
cluster, so a system that allows growth of the cluster would be very
useful. I am intending to link them via ethernet cabling (with one
computer (the server) linked to my windows95 box (i would have got rid
of win95 and gone entirely to linux but unfortuantly i have too much
of a software base built up for win95...)) Anyway all help is much
appreciated. Thanks in advance.

-dionysus


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

From: "ICON" <[EMAIL PROTECTED]>
Subject: Mirc linux mandrake dcc send
Date: Sun, 27 Jun 1999 09:30:00 -0500

I have my mandrake box as a firewall using ip masquerading. i have 3 win
boxes behind that one using cable connection. only problem is that  I can
not dcc send from the win boxes. I can recieve without any problems. can
anyone give me any help

ICON



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

From: Rob van der Putten <[EMAIL PROTECTED]>
Subject: Re: Linux PPP & Leased Line problem
Date: 27 Jun 1999 16:21:21 +0200

Hi there


[EMAIL PROTECTED] wrote:

> Yes, I did. In fact, I can connect two Linux boxes via a leased
> line easily, but not to a Cisco router. I know some succeed
> and some fail with Cisco routers, and unfortunately I am
> the one that failed.

These are equipted with a synchronous or an asynchornous serial port.
Make shure you have the right type.


Regards,
Rob
-- 
+------------------------------------------------------------------------+
|                Rob van der Putten, [EMAIL PROTECTED]                 |
|              http://www.sput.webster.nl/spam-policy.html               |
+------------------------------------------------------------------------+

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

From: [EMAIL PROTECTED] (Rod Smith)
Subject: Re: LinkSys NIC LNE100TX:  Anybody have any problems with this card?
Date: Sun, 27 Jun 1999 14:29:42 GMT
Reply-To: [EMAIL PROTECTED]

[Posted and mailed]

In article <[EMAIL PROTECTED]>,
        [EMAIL PROTECTED] (Lewis) writes:
> Folks,
> 
> I am building up a new computer, and converting my old Pentium 166/MMX
> to a Linux box (my new computer will have Linux on it as well, but it
> will primarily be a Windows98 box for my wife).
> 
> I plan to network the two computers, and am trying to determine which
> NIC cards to use.
> 
> I am considering a Linksys Fast Ethernet Kit which contains two
> Linksys LNE100TX cards and a hub.  It seems like a good value.

I've got one of these cards in one of my computers, and I've got a LinkSys
hub, but I didn't buy them as part of a kit.

> I have been to RedHat's web site, and see that they do not support
> this card, but I have been to LinkSys' web site, and see that they do
> support it for Linux.
> 
> I sure that a few of you have used this kit.  Any words of wisdom on
> this kit?
> 
> 1.  Does it work for Linux? (I know that Linksys states it does, but I
> want to hear it from somebody who is actually using it!)

Yes, it'll work with Linux.  You may need to replace one file in the Linux
kernel source tree and recompile, though.  This is because recent LinkSys
10/100 boards use a PNIC chipset (it's labelled "LinkSys" on the chip
itself, though), rather than the original DEC 21x4x.  The latest Tulip
drivers support the PNIC chipset, but they have yet to appear in any
kernel as of 2.2.9 (the latest I've checked).  You can obtain these latest
drivers from:

http://cesdis.gsfc.nasa.gov/linux/drivers/tulip-devel.html

The drivers on LinkSys' web site are the same (or possibly slightly
older).  LinkSys didn't write Linux drivers themselves, though the source
code includes a comment to the effect that LinkSys did donate a PNIC-based
board.

> 2.  Any "gotchas"?

See above.  Also, be sure that you're getting a 10/100 hub, as well as
10/100 NICs.  I've heard of kits that bundle 10/100 NICs with cheaper 10
Mbps-only hubs.  I don't know if the LinkSys kit is like this, but it
might be.

> 3.  Any suggestions for any other NIC's and hubs that support 100BaseT
> (that don't cost an arm and a leg!)?

Intel and 3COM NICs are both reputed to work well in Linux, and without
updating drivers, at least for recent kernels.  They're more expensive,
though (typically $60-$80 apiece).

Most of the inexpensive 10/100 NICs, including models from Netgear,
Kingston, SOHOware, and others now use clones of the Tulip chips,
supplied by PNIC, Macronix, ASIX, and perhaps others.  The actual chips
are often labelled with the board's manufacturer, though, so it's hard to
tell that this is the case until you see the Linux kernel bootup message
confirming the fact.  Most of these board should work fine with Linux
when using recent drivers, but may not work with the kernel "out of the
box."  I've seen reports of both success and failure for each of the
major Tulip clone board brands, though, so there may be subtle
differences from model to model, incompatibilities with some models with
other hardware, or plain old-fashioned user error.

The most popular exception to this rule I'm aware of is the D-Link
DFE-530TX, which uses a VIA Rhine chipset.  It's supported by
2.2.x-series kernels without any driver update.  The kernel source for
this board indicates that the drivers sometimes need to move data to
align it, though, which causes a speed hit.  I don't know how serious
this problem is.  I do have one of these boards in one system, though,
and it's been reliable.

No matter what model you get, I recommend purchasing from a store that
will let you return the whole mess for a refund or exchange with another
model, in case you have problems.

-- 
Rod Smith
[EMAIL PROTECTED]
http://www.channel1.com/users/rodsmith
NOTE: Remove the "uce" word from my address to mail me
Author of _Special Edition Using WordPerfect for Linux_, from Que;
see http://www.channel1.com/users/rodsmith/books.html

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

From: "Tony C" <[EMAIL PROTECTED]>
Subject: Re: LinkSys NIC LNE100TX:  Anybody have any problems with this card?
Date: Sun, 27 Jun 1999 08:04:37 -0700

I am using Linksys LNE100 cards in both my Linux PC and my Win98 PCs.
Everything is working just fine. I am running RH 52. straight out fo the box
with the most recent tulip drivers. I bought the Linksys FENSK05 kit which
includes a 10/100 Mbs 5 port hub. Like everyone else I have a high speed
internet line (384 DSL) and am using IP Masquerading to allow the Win PCs to
share the connection.

Cheers
TC

Lewis wrote in message <[EMAIL PROTECTED]>...
>Folks,
>
>I am building up a new computer, and converting my old Pentium 166/MMX
>to a Linux box (my new computer will have Linux on it as well, but it
>will primarily be a Windows98 box for my wife).
>
>I plan to network the two computers, and am trying to determine which
>NIC cards to use.
>
>I am considering a Linksys Fast Ethernet Kit which contains two
>Linksys LNE100TX cards and a hub.  It seems like a good value.
>
Snipped....




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

Date: 27 Jun 99 10:46:45 -0500
From: "Gene Heskett" <[EMAIL PROTECTED]>
Subject: Re: tcp_timestamps causing massive network slowdown problems under 2.2.x

Unrot13 this;
Reply to: <[EMAIL PROTECTED]>

Gene Heskett sends Greetings to John Gardner;

On seeing this message, it surely waddled like a duck to me.  However,
in order for it to work, I had to use this syntax:

echo \000 -n > /proc...

which did change the value of tcp_timestamps to 0.  The 'cat 0 >' was an
error as it couldn't find the file 0.

Going online immediately after a reboot, and a check of its value after
the reboot said it was now a 0, not a 1.  I used NS4.61 to grab a couple
of things from metalab.  No efective difference for me, the speed was
still 1200 cps over a 50kb connect, with lots of stalls.

This was kernel 2.2.10 and the new pppd that goes with it.

So this is not my problem.  But I'll look at the daemon code too.

Can IPTraf give me any useful info from a dialup?

 JG> In article <[EMAIL PROTECTED]>, Gary Elmes
 JG> <[EMAIL PROTECTED]> writes
>>I have just had a couple of fun days trying to work out why my http
>>downloads slow to a crawl from certain sites.
>>
>>I solved the problem by setting /proc/sys/net/ipv4/tcp_timestamps =
>>0. Is this a known problem?
>>
>>I found that, with timestamps on, packets from problematic sites
>>(www.isc.org was the one I tripped over) were showing up as
>>fragmented by tcpdump, even though I have the kernel compiled to
>>reassemble fragmented packets. In addition, the remote http host
>>took yonks
>>(minutes at a time) to respond.
>>
>>Typically, a 1460-byte packet would be fragmented into 1448 and 12.
>>
>>Disabling timestamps solved the problem - no fragmenting, no
>>slowdown, nice whizzy downloads. I now have:
>>
>>    cat 0 > /proc/sys/net/ipv4/tcp_timestamps
>>
>>in my rc.local, and everything is fine.
>>
>>I reproduced the problem under both 2.2.1 and 2.2.10, and with all
>>manner of network-related bits both compiled in and left out of the
>>kernel (all in an effort to diagnose the problem). I reproduced it
>>with both lynx and communicator, as well. My connection to the
>>internet is via dial-up ppp; and stuffing with all the available ppp
>>options
>>(compression, mru, mtu, etc.) did nothing to solve the problem.
>>Other machines - both win95 and linux 2.0.36 - using this box as a
>>gateway to the internet did not have the problem (timestamps were
>>not set on their packets).
>>
>>Should somebody be fixing something in the timestamping code, or is
>>my "fix" just working around the real problem?

 JG> I came across a similar problem a few days ago. When using telnet
 JG> or rlogin from Linux and leaving the connection for five minutes
 JG> without any traffic on the line, a lockup occurs from the
 JG> application point of view. 

 JG> Examining the trace shows TCP timestamps being used. However,
 JG> Linux fails to update the sequence number after the next
 JG> character is typed and the remote stack keeps sending the same
 JG> data. Sometimes Linux recovers, sometimes the connection is lost.

 JG> If timestamps support is removed (in the same way that you did),
 JG> the stack works correctly. 

 JG> The stack employed by the OS I'm responsible for, is developed
 JG> from Mentat's streams implementation. Solaris also uses Mentat
 JG> TCP and this stack is (presumably) well tried in the field. 

 JG> In your case, is there retransmission of data, or is it purely a
 JG> fragmentation problem?

 JG> Certainly, turning off timestamps is only an avoidance. 

 JG> Regards,
 JG> -- 
 JG> John Gardner
 JG>                         


Cheers, Gene
-- 
  Gene Heskett, CET, UHK       |Amiga A2k Zeus040 50 megs fast/2 megs chip
    Ch. Eng. @ WDTV-5          |A2091,GuruRom,1g Seagate,CDROM,Multiface III
                               |Buddha + 4 gig WDC drive, 525 meg tape
                               |Stylus Pro, EnPrint, Picasso-II, 17" vga
         RC5-Moo! 690kkeys/sec isn't much, but it all helps
email gene underscore heskett at iolinc dot net
-- 


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

From: [EMAIL PROTECTED]
Subject: How to use NT Printer & files from Suse 6.1
Date: Sun, 27 Jun 1999 15:39:53 GMT

I've compiled the kernel with SMB support built in, played around with
Samba, read a few how-to files and got nowhere. Most of the
documentation assumes I want Linux to be a server.

Can someone please point me to some useful instructions at the level of
"What do I set to TCP/IP addresses and sub masks to?" as all the help
I've read so far assumes I know things like this. I have got the NT
machine to talk to other NT machines over a TCP/IP network (I just made
up the addresses).

What diagnostic tools are available beyond the Network Sniffer which I
have tried to tell me how much success I am having and what I need to
change if not?

Best regards,
Fred.


Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

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

From: Nathan Valentine <[EMAIL PROTECTED]>
Subject: Re: SOCKS5 and RH6
Date: Sun, 27 Jun 1999 15:38:00 GMT


Bernhard,

        I've found the answer! First, the "Received request with incompatible
version number: 71" is because I had Netscape's HTTP proxy set to the
port of my SOCKS proxy. So I cleared all of my proxy settings except the
SOCKS box. Then, like you said, I had to set up a caching nameserver for
my home machines. I suppose you could probably funnel DNS lookups thru
SOCKS, but it is probably a bigger pain than it's worth. Now everything
works just fine. 

        BTW, I did an AltaVista search on the SOCKS error messages and got some
links to NEC's SOCKS mailing list. The problem that we have been having
appeared in their mailing list like ten times! I guess very few other
people went to USENET for help. Go figure. <shrug>


-- 
Nathan Valentine - [EMAIL PROTECTED]          AIM: NRVesKY
=====================================================================
University of Kentucky Linux Users Group
http://www.uky.edu/StudentOrgs/UKLUG

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

From: [EMAIL PROTECTED] (Frank Hahn)
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup
Subject: Re: PROXY
Date: Sun, 27 Jun 1999 15:44:14 GMT
Reply-To: [EMAIL PROTECTED]

On Thu, 24 Jun 1999 22:25:59 -0400, Marlon <[EMAIL PROTECTED]> wrote:
>[EMAIL PROTECTED] wrote:
>> I linux box(198.168.200.1) doing PPP.  How do I get WIN95
>> (198.168.200.2) Browser to connect to the internet through the linux
>> box?
>> I can telnet from win95 to linux.
>
>I too am trying to do the same thing.  I have a Linux box with IP
>192.168.1.254 with my Win95 logged into in with samba.  My Win95's IP is
>192.168.1.2.
>
>How do I get my Win95 client to share my PPP connection that Linux has
>up and running?
>
After setting up IP Masquerading on the Linux machine, you need to
tell the Windows machine to use the Linux machine as its gateway.
Go into the TCP/IP setup on the Windows machine.  There is a box
to enter the gateway address.  This needs to be the IP address of
the Linux machine.

In this same area on the Windows machine, there is a place to put
the IP numbers of your DNS machine.  I placed the numbers provided
by my ISP here also.

-- 
Frank Hahn

Immigration is the sincerest form of flattery.
                -- Jack Paar

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


** 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