Linux-Networking Digest #860, Volume #11         Sun, 11 Jul 99 12:13:33 EDT

Contents:
  Re: /etc/ppp/peers/... permission denied (addendum) (Michael Steffens)
  Re: local ppp no longer working with kernel 2.2.x ("Eric G. Stern")
  Re: Tweaking Linux / Becoming 'guru' (Monte Phillips)
  Re: Networking for Dummies ("Juan Carlos")
  Red Hat as Print server ("steve epstein")
  Re: Could Microsoft Cheat On The New Mindcraft Benchmark? (Hartmann Schaffer)
  Re: DHCP and host name resolution (Hartmann Schaffer)
  Re: ppp pap auth failing (Clifford Kite)
  Samba & win98 long file names ("JimA")
  Re: How To Map A Drive From Window95/98 (mike murray)
  Re: DDNS Server for Linux (Steve Horejsi)
  Re: How to make Linux _NOT_ use a gateway at all??!!  Help! ("Eugene")
  Re: PC Anywhere from private net through IPChains (Dont Spam Me)
  Re: PPP connect: ping & traceroute but no data?? (Clifford Kite)
  Re: multiple NIC's ("Eugene")
  Re: Supported Token-Ring Cards?? (Steve Horejsi)
  Re: Multiple connections at the same time (combined) ("Eugene")
  Re: Linksys Card (Rod Smith)
  Devise or resource busy ("Ronald J. Yacketta")
  partitions on RH 5.2 ("Francisco Cáceres")
  Re: Weird fast(er) connection... ("Eugene")
  dhcpcd and ADSL (Mugur Dragus)
  Re: RH 5.2 IP Masq problem ("Allan Wingenbach")
  HELP: linux and dhcpcd in a NT network ([EMAIL PROTECTED])

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

From: Michael Steffens <[EMAIL PROTECTED]>
Subject: Re: /etc/ppp/peers/... permission denied (addendum)
Date: Sun, 11 Jul 1999 14:36:32 +0200

Sorry for replying to my own posting immediately,
but I forgot to mention something:

Michael Steffens wrote:
[...]
> My question is: How can pppd, despite its suid bit
> being set, have difficulties in reading the provider
> specific options file? How can this be fixed?

This could happen, if the file were located on an
NFS device. However, this is not the case, it's local.

Thanks again!
Michael

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

From: "Eric G. Stern" <[EMAIL PROTECTED]>
Subject: Re: local ppp no longer working with kernel 2.2.x
Date: Sun, 11 Jul 1999 08:12:26 -0500

I found the problem.  The automatic firewall rules that Debian provides
doesn't seem to be configured to work with my setup on 2.2 even though
it uses ip-chains for 2.2.

                        Eric Stern


Eric G. Stern wrote:
> 
> kernels 2.0.28 through 2.0.36.  When I tried to upgrade to the 2.2.x
> series
> of kernels I can't get the ppp link to come up.

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

From: [EMAIL PROTECTED] (Monte Phillips)
Crossposted-To: comp.os.linux.misc,alt.os.linux
Subject: Re: Tweaking Linux / Becoming 'guru'
Date: Sun, 11 Jul 1999 12:59:57 GMT

Get in the linux developers newsgroups and on their mailing lists.
Also some of the HOWTO's are written (it seems) for the very advanced.
Take a look at samba-confHOWTO sometime or even the NET-3 and its
references.  In fact those references at the end of the HOWTO's are
the place to look.



Chris Harshman <[EMAIL PROTECTED]> wrote:
>Can anyone recommend good reading (O'Reilly titles,
>URLs, whatnot) with respect to tweaking Linux for
>performance?  I've been a systems administrator for
>4 years now, and I think I've performed ably.  
>My boxes don't crash, are always available, and
>are hacker-resistant.  =)  
>But now I'm being called upon to provide "enterprise-
>class" performance out of Linux machines, in general
>and in specific circumstances (e.g., tweaking the
>kernel to provide specialized performance gains in
>areas such as file I/O).  
>
>Where might I begin to fill my brain with such information?
>I'm not a coder, but I know how to code when necessary.
>(No, not `./configure ; make ; make install` - when I
>need something done and I can't find software to do it,
>I write it.  However, it's not always the best optimized
>or the prettiest code you've ever seen ;)
>
>For example, say you were on the team responsible for
>tweaking a Linux box to go head-to-head against
>an NT server for a rematch of the Mindcraft benchmarks.
>Where might you turn for tips/tricks/advice?


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

From: "Juan Carlos" <[EMAIL PROTECTED]>
Subject: Re: Networking for Dummies
Date: Sun, 11 Jul 1999 10:13:51 -0700

Guess I wasn't very clear.  I have a linux box running oracle at home, that
I access from my Windows 98 machine, also at home, through a telnet session
over ethernet.  My employer is asking us to install RAS on our home PC's so
we can dial into work to do support.  I am leery of doing this because of
the grief I went through getting my own network setup.  In reading the
instructions, it sounds like I have to setup RAS much like I did my home
network.  I am just wondering if the two will play nicely together.


Joseph Cheek wrote in message <7m97gn$4mf$[EMAIL PROTECTED]>...
>hi juan,
>
>not sure what you mean.  you want to know if you can use win98 ras to
connect to a linux network?
>
>joe
>
>In comp.os.linux.networking, "Juan Carlos" <[EMAIL PROTECTED]> wrote:
>
>> After a weekend of getting my Win 98 conect to linux using ethernet, my
>> employer now tells me I must use RAS on Windows to keep me captive 24
hours
>> a day.  I know this is not a linux-specific question, but can this
co-exist
>> with my linux setup, or should I tell my boss to take a hike?  Please
help
>> me keep my job.
>
>
>--
>LinuxNews Beta -- http://linuxnews.cheek.com/
>Think of us as DejaNews for Linux.



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

From: "steve epstein" <[EMAIL PROTECTED]>
Subject: Red Hat as Print server
Date: Sun, 11 Jul 1999 08:46:24 -0500

I am running a Red Hat file/print server using Samba and Win98 workstations.
When I first try to print, it takes nearly a minute to get the print menu.
Evidently something is starting up, as the next time I go to print it is
near instantaneous. If I wait a while, the print service evidently shuts
down. How do I tell what is running, and how do I keep it running, at least
whenever someone is logged in?

Thanks

steve dot epstein at worldnet dot att dot net



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

From: [EMAIL PROTECTED] (Hartmann Schaffer)
Subject: Re: Could Microsoft Cheat On The New Mindcraft Benchmark?
Crossposted-To: 
omp.os.ms-windows.nt.advocacy,comp.os.linux.advocacy,comp.infosystems.www.servers.unix,comp.os.linux.misc
Date: Sat, 10 Jul 1999 19:58:02 GMT

In article <7m7kmm$2pl$[EMAIL PROTECTED]>,
        [EMAIL PROTECTED] (Joseph T. Adams) writes:
> ...
> Nope.  Read about the Roman Republic and Empire.  Several of its
> conflicts, including the Punic Wars IIRC, were similar in scope, and
> even destructiveness to human life (adjusting for the smaller world
> population at the time), than WWI.  And this in spite of the fact that
> Rome was unchallenged, and unchallengable, by any single external
> power.  Usually, opponents like Attila and Hannibal were able to
> profit from the extreme discontent of folks in conquered territories
> that Rome already considered to be under control. 

When Attila came into the picture, Rome was far from unchallengeable.
The (West) Roman Empire was in the process of breaking apart.

-- 

Hartmann Schaffer

It is better to fill your days with life than your life with days

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

From: [EMAIL PROTECTED] (Hartmann Schaffer)
Subject: Re: DHCP and host name resolution
Date: Sat, 10 Jul 1999 20:03:26 GMT

In article <[EMAIL PROTECTED]>,
        "Whiplash!!!" <[EMAIL PROTECTED]> writes:
> This may be a dumb question but I'm real new to this so here goes...
> 
> I just setup a Linux DHCP server. I also set up several linux clients to
> pull addresses from that server. After booting the linux clients, I run
> into several problems, I cannot do a 'rsh servername' anymore and Gnome
> complains that I need to make a new entry in hosts.
> 
> I assume this is because I had an entry in hosts that assigned my
> hostname and others to particular IP addresses. so I removed the hosts
> file and of course got the same errors.
> 
> My question: When using DHCP, how do you assign the hostname to the new
> IP on the fly? More importantly, how do other computers find you by your
> hostname when your IP is constantly changing? I believe that windows
> uses broadcasts. Can Linux be configured to do the same?

Check out the DHCP documentation.  There are a number of criteria you
can use to assign an IP address to the requester.  I think the ethernet
address where the request originates from is one of them.

-- 

Hartmann Schaffer

It is better to fill your days with life than your life with days

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

From: kite@NoSpam.%inetport.com (Clifford Kite)
Subject: Re: ppp pap auth failing
Date: 11 Jul 1999 09:32:30 -0500

[EMAIL PROTECTED] wrote:

: i have a remote access account at work, and they have recently changed
: their dial up setup.
: i believe the new system is using PAP, i'm able to successfully connect
: with windows NT without any scripting and just putting in the
: name/password they gave me

: however, i cannot now connect with linux (i have redhat 5.1)

: when i dial up their number with minicom, this is the output i get:

: ~~ÿ}#.!}!.} .}"}&} }*} } }#}$.#}%}&
: Göâ}'}"}(}"}1}$}%ô}3}2}!cmdnnjce-sap-016F~~~

: when looking in the logs after running pppd, i see this:

[edited]

: Jun 20 19:24:43 localhost pppd[1385]: Using interface ppp0
: Jun 20 19:24:43 localhost pppd[1385]: Connect: ppp0 <--> /dev/modem
: Jun 20 19:24:47 localhost kernel: ppp_tty_ioctl: set xmit asyncmap a0000
: Jun 20 19:24:47 localhost kernel: ppp_tty_ioctl: set rcv asyncmap
: ffffffff
: Jun 20 19:24:47 localhost kernel: ppp_tty_ioctl: set flags to f070003
: Jun 20 19:24:47 localhost kernel: ppp: frame with bad fcs, excess = 9088

Lots of fcs errors.
..

: Jun 20 19:25:17 localhost pppd[1385]: No response to PAP
: authenticate-requests
: Jun 20 19:28:16 localhost pppd[1385]: Terminating on signal 15.
..

: is the server indeed expecting PAP? or one of the micro$oft propietery
: protocols

It's hard to be sure without pppd debug messages but I think your company
has likely upgraded it's PPP implementation to broken.  Try using the
pppd option "asyncmap a0000" to match what the peer asked for itself -
a broken ACCM algorithm is common.  The fcs errors should go away if
this guess is correct.  Try "asyncmap 0" if "asyncmap a0000" fails.

: also, i'm not sure why i'm not seeing the  'sent' 'rcvd' messages from
: the pppd  even though i have debug on

There should be pppd debug messages, here they are in /var/log/debug,
but the location varies with distribution.  The log files are configured
in /etc/syslog.conf; if none of the files you see there contain the pppd
debug messages, then try adding, say,

daemon.*                                       /var/log/ppp-log

to syslog.conf .

If you need to post log messages again then please consider getting rid
of the pppd option "kdebug <something>" that is causing the fcs errors to
be reported.  These messages are rarely useful and make reading the log
more difficult.

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



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

From: "JimA" <[EMAIL PROTECTED]>
Subject: Samba & win98 long file names
Date: Sun, 11 Jul 1999 07:45:14 -0700

What do I set in the smb.conf to get win98 to see log file name
format on the samba sever?

In don't see it in the HOWTOs. Right now this is how I have
smb.conf set;

# Samba config file created using SWAT
# from win98 (10.8.11.7)
# Date: 1999/07/08 19:55:36

# Global parameters
 workgroup = workgroup
 server string = Samba
 interfaces = 10.8.11.1 127.0.0.1
 security = SHARE
 encrypt passwords = Yes
 use rhosts = Yes
 protocol = LANMAN2
 printcap name = /etc/printcap
 os level = 70
 preferred master = Yes
 domain master = Yes
 wins server = 10.8.11.1
 guest account = samba
 force create mode = 0700
 hosts allow = 10.0.0.0 127.0.0.0
 hosts deny = none
 short preserve case = No
 mangle case = Yes

[Workgroup]
 path = /group
 read only = No
 guest ok = Yes

[printers]
 comment = All Printers
 path = /tmp
 read only = No
 create mask = 0700
 guest ok = Yes
 print ok = Yes




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

From: mike murray <[EMAIL PROTECTED]>
Subject: Re: How To Map A Drive From Window95/98
Date: Sun, 11 Jul 1999 09:08:22 -0500

go to network neighborhood.
Can you see the Linux box?
If so, then click on it  and from tools ( i think) at the top. in hte
pull down list, map a network drive.

Later


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

From: Steve Horejsi <[EMAIL PROTECTED]>
Subject: Re: DDNS Server for Linux
Date: Sun, 11 Jul 1999 10:09:08 -0700

 I can't see the beginning of this thread on the news server I'm using, so at
the risk of repeating something said earlier on, there's always BIND 8.x. It
supports all of the nifty parts of DDNS and is free; heck, it's probably on the
media that came with your distribution.

 BIND is not the easiest thing in the world to get running if you've never
dealt with a DNS before but it's not malicious, merely obscure.

 Most of the Internet runs on BIND, so 10K addresses wouldn't be any problem.
There is reasonably good documentation with it, but the recognized wellspring
on knowledge is still 'DNS and BIND' by Albitz and Liu. (O'Reilly)



                            -=[ Steve ]=-



Phil Green wrote:

> Are there any other products?  For example, is there anything like NetID
> from Bay Networks in the Linux/Unix world.  It would be important to be able
> to support like 10,000 dhcp addresses/dns entries, etc. and be able to do
> this across the enterprise.
>
> Please let me know if there is anything like that our there.
> TIA

<...snip...>


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

From: "Eugene" <[EMAIL PROTECTED]>
Subject: Re: How to make Linux _NOT_ use a gateway at all??!!  Help!
Date: Sun, 11 Jul 1999 15:05:30 GMT

read networking howto at www.linux.org/help

Evan A. Reidell <[EMAIL PROTECTED]> wrote in message
news:7m9fgs$[EMAIL PROTECTED]...
> I don't have a gateway machine or a Cisco router, and I'm
> trying to make a telnettable Linux box --- I thought this
> would be easy!!!
>
> With others like a Win98 machine, NeXT machine, Sun
> workstation, SGI Indy, or Macintosh, I can pretty much set the
> IP address to 200.0.0.5 (or whatever) and the netmask to
> 255.255.255.0, then connect it (via a hub) to another
> 200.0.0.x machine, and I've then been able to "telnet
> 200.0.0.5" (well, not Win98 or Mac) or "ftp 200.0.0.5"
> (well...)
>
> ANYWAY
>
> the Linux machine keeps asking for GATEWAY numbers and such,
> and I can't figure out how to fake it out ---
> and it refuses to sit there and pretend that it's 200.0.0.5
> so I can ftp to it from another machine!
>
> I'm sure this is a pretty dumb question, but please answer it
> anyway!!!  Email or post, I just want to ftp into my Linux
> box!  (Redhat 5.2)
>
> evan reidell
> ereidell at mit dot edu
>
>
>



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

From: [EMAIL PROTECTED] (Dont Spam Me)
Crossposted-To: comp.os.linux.security
Subject: Re: PC Anywhere from private net through IPChains
Date: Sun, 11 Jul 1999 15:09:36 GMT

I forget which ports, but PC Anywhere uses multiple ports simultaniously, and 
it might even mix TCP and UDP. To determine that, just get a connection going 
inside your firewall and run netstat on the machines. Assuming there are no 
connectionless communications going, this should reveal all.

You might want to look at VNC also. This free package lets you remotely access 
a system in a similar manner to PC Anywhere, and uses only 1 socket. (Well, 2, 
because it allows a java capable browser to connect on one port, or the normal 
client on another, but only 1 would need to be port forwarded.) The main 
differences between this free package and PC Anywhere are that PC Anywhere has 
better authentication (multiple user/passwords, VNC has only a single 
password), and PCA has additional functions like file transfer.

In article <[EMAIL PROTECTED]>, Raymonds Doetjes 
<[EMAIL PROTECTED]> wrote:
>You need a portforwarding package for this.
>I don't know if PCAnyWheres protocoll uses several sockets simultaniously
>like f.i IRC, and FTP do. If that is the case, then you need a masq module
>otherwise you will be fine.
>
>May I suggest rinetd for the protforwarding this is a good package is
>stable and configuraring is easy. PORTFORWADING in the kernel is still
>somewhat buggy.
>
>Raymond
>
>[EMAIL PROTECTED] wrote:
>
>> I have a successful setup using IPChains and IPMasq to allow multiple
>> machines with 192.168 addresses to share a single routable address. Ftp
>> and RealAudio were taken care of with the ipmasq kernel modules.
>>
>> Any idea on how to get PC Anywhere through? I get the sinking feeling it
>> embeds an address in the outgoing packet. Without an ip masq module to
>> handle this, are there any other options. Or am I wrong and can this be
>> dealt with using autofw or portfw?
>>
>> Thanks.
>

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

From: kite@NoSpam.%inetport.com (Clifford Kite)
Subject: Re: PPP connect: ping & traceroute but no data??
Date: 11 Jul 1999 10:23:11 -0500

Paul and Linda Hardin ([EMAIL PROTECTED]) wrote:

: I'm running ppp on my linux box and connect to my isp just fine. DNS is
: working, I can traceroute and ping by name or ip.

: However, when I telnet, ftp, or use a browser I get all the 'connected
: to host' messages but no data. For ftp and telnet this means no
: 'username' prompt while netscape leaves me at 'waiting for data'.

The kernels 2.2.x, x<8, may have had problems of this kind.  There was a
significant revision of the kernel PPP support code, ppp.c, in 2.2.8 .
I also know of at least one poster that had this kind of problem with
both 2.2.5 and 2.2.7, in it the problem went away by using the pppd
novj option.  A kernel upgrade is the correct solution; I'm using 2.2.9
with no problem.

--
Clifford Kite <kite@inet%port.com>                    Not a guru. (tm)
/* Microsoft is a great marketing organization.
 * It _has_ to be */

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

From: "Eugene" <[EMAIL PROTECTED]>
Subject: Re: multiple NIC's
Date: Sun, 11 Jul 1999 15:13:25 GMT

edit /etc/conf.modules

add lines:

alias eth1 <nic_module>
options <nic_module> io=xxxx irq=yy

where <nic_module> is the name of your network card module (look under
/lib/modules)
irq is the network card's irq
io is the io address

For example, I have ne2000 - compatible network card (module: ne), which has
irq 11 and io 0xd400 (the 0x preffix is used to specify hex numbers)

alias eth1 ne
options ne irq=11 io=0xd400



Matt <[EMAIL PROTECTED]> wrote in message news:[EMAIL PROTECTED]...
> I am running Red Hat 6.0, 2.2.5, and I have been trying to get my second
> ethernet card up so I can use IP MASQ.
>
> I have a 3c905 for the eth0, which works fine, and is connected to the
> cable modem.
>
> But the non-working one is a 3c509, and PnP is turned off!  I have tried
> used the "append" thing in the lilo.config file I have tried playing
> around with "linuxconf"  but when I boot the machine it says that it is
> delaying eth1, and there is a "FAILED" in red next to it, when all the
> others are "OK".  Any help would be greatly appericated.  I am sorry if
> I missed anything, I am kinda new to this.  Thanks again.
>
> -Matt
>



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

From: Steve Horejsi <[EMAIL PROTECTED]>
Subject: Re: Supported Token-Ring Cards??
Date: Sun, 11 Jul 1999 09:50:26 -0700

Sean;

 Linux support of Token Ring is limited to the IBM 'Tropic' chipset which was
also used by other manufacturers. In general, all ISA cards and all
(gasp!) MicroChannel cards (with the exception of the 'Streamer' models) will
work. The PCI ones use a different chipset and won't work with the TR support in
Linux.

 From a PCMCIA perspective it gets a little trickier. I know that the 'Turbo'
cards won't work, but I believe the older 16/4 ones should.

 I have Red Hat (5.2 and 6.0) running on several machines with the old ISA cards
without having lost any great volumes of hair getting them working.


 Hope that helps...


                                    -=[ Steve ]=-


Dele9831 wrote:

> Does anyone know what token-ring cards are supported under Linux (RedHat 6.0
> running on an IBM 600e Thinkpad w/docking station)?  The Hardware HOWTO says
> that IBM and Madge cards are, but I have fairly new ones of both (PCI, Wake-On
> LAN for the IBM), and it cannot seem to see it.  I haven't tried the old IBM
> token ring ISA cards yet, and for obvious reason don't really want to.  Has
> anyone else run into this problem?
>
> Sean
> [EMAIL PROTECTED]


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

From: "Eugene" <[EMAIL PROTECTED]>
Subject: Re: Multiple connections at the same time (combined)
Date: Sun, 11 Jul 1999 15:18:33 GMT

yes and no
if you add the second cable modem the total bandwidth will double (at least
in theory, I'm not sure how well this theory works for cable modems since
they use shared bandwith)

BUT your connection speeds will NOT double, they'll remain the same. You
will be able to take advantage of increased bandwidth only if you open
multiple connections at the same time.

Jeremy McNicoll <[EMAIL PROTECTED]> wrote in message
news:rxRh3.25375$[EMAIL PROTECTED]...
> Hi:
> Is it possible to double your throughput, by combining 2 cable connections
> to get mega bandwidth to the same machine?
>
>
>



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

From: [EMAIL PROTECTED] (Rod Smith)
Subject: Re: Linksys Card
Date: Sun, 11 Jul 1999 14:53:41 GMT
Reply-To: [EMAIL PROTECTED]

[Posted and mailed]

In article <[EMAIL PROTECTED]>,
        [EMAIL PROTECTED] (JunkDTectr) writes:
> On Sat, 10 Jul 1999 01:25:59 GMT,  In comp.os.linux.networking, 
> [EMAIL PROTECTED] says...
>> 
>> I had the same problem. You MUST get the later driver, at least .90 .
> 
> I can't say why it works for me, I don't know of anything I've 
> done out of the ordinary.  Seems like I may have run modprobe 
> and/or maybe insmod.  
> 
> I have 2 of the things in my stock RH5.2 system using the stock 
> Tulip v0.89H modules.  No compiling, nothing.  I'm kind of 
> peculiar that way, I like to keep things as stock as possible.

First, there are at least two major versions of the Linksys LNE100TX.  One
uses the "genuine" DEC Tulip (21140, I believe) chipset.  This board works
fine with the older 0.89H driver that comes with kernels through 2.2.9 (at
least; I've not checked later kernels).  The second variety uses a PNIC
clone of the DEC Tulip.  This version may not work with that driver (my
card certainly doesn't).

Second, I've heard that PNIC-based Tulip clone boards do sometimes work
with the older Tulip drivers *IF* their firmware is of a certain
revision.  I don't know the details on this, but I've heard that some
people had problems with such boards with some firmware revisions but not
others.  It's possible that the Linksys is like this, but I've never
looked into it -- using a newer Linux driver is easy enough for me.

-- 
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: "Ronald J. Yacketta" <[EMAIL PROTECTED]>
Subject: Devise or resource busy
Date: Sun, 11 Jul 1999 11:21:17 -0400

Hello all!!
I have a bit of a problem here
I have two nic's in my linux box , one is a rtl8139 and the other is a
tulip.
When I boot I see error messages saying that eth0 and eth1 are delayed.
I tried to load the mod manualy and get a error stating "Device or resource
busy"

any advise?

Thanxs in advance

Ron




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

From: "Francisco Cáceres" <[EMAIL PROTECTED]>
Subject: partitions on RH 5.2
Date: Sun, 11 Jul 1999 11:22:48 -0300

Hi,

It is posible to change de size of a partition on RH 5.2 ? I installed the
server and the instalation made a partition for usr/local and now Iam
running out of space on that device...

Francisco
[EMAIL PROTECTED]



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

From: "Eugene" <[EMAIL PROTECTED]>
Subject: Re: Weird fast(er) connection...
Date: Sun, 11 Jul 1999 15:25:09 GMT

it might have to do with modem compression.
all modems can compress data for faster transfer. Compression might have
been turned off. Most people never bother with it since most files you
transfer are zipped already and don't compress at all. What kind of file was
it that you downloaded?

P.S. Linux does kick as, but it can't magically make your modem work faster
then its maximum bandwidth.

dpc <[EMAIL PROTECTED]> wrote in message
news:YwRh3.40$[EMAIL PROTECTED]...
> I've been having some interesting download speeds ever since I set up my
> Linux box as server/router to my other NT box.
> Ever since I yanked the modem from NT and put it in Linux, I have
> experienced abnormally high download speeds on my 56K modem.  Abnormally
> high being the transer was peeking out around 10-11KB/sec.  I was hoping
> perhaps somebody here might have an explanation of this.   The setup is as
> follows.
>
> Linux box w/56K modem doing ipmasq for routing.
> The two are connected via crossover ethernet cable running at 100Mbps
> My only thought was that somehow, because when the NT box is d/l'ing, it's
> going over the ethernet cable and getting it's speeds...mixed up somehow.
> However, as I write this, I d/l'ed a 595KB file in 74 seconds or 8.04
> KB/sec. So the speeds DO seem to be accurate.
> On Windows, I'd never been able to get over 6KB/sec!  And now Linux has
held
> 11KB/sec transfers over a period of an hour!?
>
> Is this merely Linux kicking ass again or just my imagination?
>
> Thanks for any ideas,
>
> dpc
> Remove AllYourClothes to reply directly to me.
>
>
> 0000000000000000000000000000000000
> http://www.deja.com
> Chances are, your question
> has already been answered.
> ==============================
>
>



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

From: Mugur Dragus <[EMAIL PROTECTED]>
Subject: dhcpcd and ADSL
Date: Sun, 11 Jul 1999 15:25:36 GMT


Anyone got Caldera 2.2 (kernel 2.2.5) working with ADSL?
It seems that the dhcp client (ver 1.3.17) is unable to get the IP
address (I have
though both Rx and Tx packets!). Moreover, it's shutting down the eth0
card (3Com509, correctly installed).
>From what I read in other NGs, it seems that I'm missing some files (I
wander why they were not created when I first choose the option DHCP?!).



Pls help, someone!
Mugur


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

From: "Allan Wingenbach" <[EMAIL PROTECTED]>
Subject: Re: RH 5.2 IP Masq problem
Date: Sun, 11 Jul 1999 15:56:10 GMT

Thanks for the response.
The alias string is set properly, without io arguments.  I haven't tried
irq - will try tonight.  Drivers were compiled as modules.

Ted Goodwin <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> Alias string in conf.modules? Try loading modules with no io and irq
arguments?
> Did you compile as modules or in the kernel for the card? If you compiled
as
> module try loading it by hand using the full pathname, that should narrow
the
> problem down.
>
> Ted
>
> Allan Wingenbach wrote:
>
> > ****Newbie Warning!!****
> > Hi Everyone.
> > I'm trying to get my Red Hat Linux 5.2 box setup to do IP Masquarding.
Have
> > read and followed HOW-TOs on IP masuq and installing 2 network cards in
same
> > computer.  Before I setup the IP Masquarding by following directions in
the
> > miniHOW-TO, both network cards were in and working fine.  I could ping
and
> > communicate (http) with clients on both subnets.
> > Now after I've recompiled the Kernel to add IP masq support and
installed
> > it, I rebooted as per the HOW-TO and my 2nd network card (eth1) is gone!
> > I've tried everything suggested in the Ethernet HOWTO about using 2
ethernet
> > cards, but to no avail.  Using ifconf eth1 returns "unknown interface".
> > modprobe eth1 returns "can't locate module eth1".
> > Both Cards are 3Com 3c509b.  eth0 is at 0x300 IRQ 10, eth1 would be at
0x210
> > IRQ 11.
> >
> > Any Ideas?
> >
> > Allan
>
>
>



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

From: [EMAIL PROTECTED]
Subject: HELP: linux and dhcpcd in a NT network
Date: Sun, 11 Jul 1999 15:55:04 GMT

Hi all,

I've been reading a lot about dhcpcd and linux but
so far I was unable to setup my linux box
correctly in an NT network. I'm running Slack4.0
and the ethernet card is well configured.
dhcpcd is version 1.3.17.

dhcpcd seems to work fine, since I get an IP
from the server, as well as a DNS, net mask, etc.
But I still can't telnet from other Win98 machines
in the same network into my linux box, nor can I
telnet into other unix machiines outside this
network. In the dhcpcd-eth0.info the variables
HOSTNAME and NETWORK are not set.

Surprisingly, I can ftp to the NT server, but not
beyond that.

Any help would be appreciated
Sincerely yours

Antonio Santos





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

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


** 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