Linux-Networking Digest #428, Volume #12 Tue, 31 Aug 99 17:13:30 EDT
Contents:
Re: Netmeeting & IP Masquerading ("Moses Kamai")
Re: Firewalling Question ("Robert_Glover")
Re: Routing in Ring (Cluster) ("Robert_Glover")
Re: Help setting up Samba ("Chris L")
Re: eth0: tulip.o: init_module: device or resource busy (Cliff)
Re: Is this possible? ("Chris L")
Re: Is this possible? (Mike)
Re: eth0: tulip.o: init_module: device or resource busy ("David Murray")
diamond homefree phoneline (John-Paul Allen)
How to deny FTP-access to a whole domain ("gonZo theGreaT")
Re: Quick question 3C509B ([EMAIL PROTECTED])
Re: Can't mount Windows drives (s)
Re: What domain do I use? ("Jan Geertsma")
Re: Force D-Link 530-TX to use 100 MBit / Half Duplex ("Torsten Pinkert")
Error message: tcp/telnet:unknown service (Sender)
Re: samba passwd encrypt question (Andrew Williams)
NFS caching (bill davidsen)
Re: CPU needed to make use of 100 mbit ("Pat Crean")
ppp and samba (Francisco =?iso-8859-1?Q?Jos=E9?= Toledano =?iso-8859-1?Q?Alcal=E1?=)
Re: Network card - newbie
neighbor table overflow? Error? ("Scott MacDonald")
Re: eth0 and modules (Howard Mann)
Re: networking slows down (Tom Eastep)
Re: linux box vs switched hub (David C.)
----------------------------------------------------------------------------
From: "Moses Kamai" <[EMAIL PROTECTED]>
Subject: Re: Netmeeting & IP Masquerading
Date: Tue, 31 Aug 1999 12:44:35 GMT
Sorry you missed the posts here on this sore subject of Microsoft's product.
Bottom line: without a little help from a third-party product - you can't.
Here is the post again.
Some time back, some posts were made on getting Netmeeting to work from an
internal LAN through a Linux firewall (or any firewall for that matter) to
the internet. I finally found a product that works with NM 3.01. I have
not checked the security implications yet but at least it works.
The company, thus NOT free software, provides a switching type program that
basically performs proxy services for NM 3.x. The company is called
Equivalence, Inc. (http://www.equival.com.au/). The latest version, 1.2 -
14 Aug 99, is the I tested with Slackware 4.0 (Linux 2.2.10.).
TURBO1010 <[EMAIL PROTECTED]> wrote in message
news:7qfls3$2rl$[EMAIL PROTECTED]...
> I got netmeeting 3.1 to work, no problem if I call someone on the net.
But
> if they want to contact me, oviously they can't because they're calling
the
> Linux box. My question is, and maybe this could be done, how can I get IP
> chains, to forward the requests of netmeeting to my internal lan, on a
> certain computer(s). Example, a request comes in on eth0 for netmeeting,
to
> be forwarded to eth1. So, if a request comes in on a port for netmeeting,
> then it shoud be forwarded to 192.168.1.3 eth1. I get the concept, just
> don't know what ports, or how to tell IP chains to forward the requests.
> Anyone run into something similar?
>
> Thanks in advance for any input.
>
>
>
>
>
>
------------------------------
From: "Robert_Glover" <Please_reply_to@newsgroup>
Subject: Re: Firewalling Question
Date: Tue, 31 Aug 1999 17:21:54 -0000
What you're looking for is ipportfw
------------------------------
From: "Robert_Glover" <Please_reply_to@newsgroup>
Subject: Re: Routing in Ring (Cluster)
Date: Tue, 31 Aug 1999 17:30:33 -0000
I'm not sure I can help you there, but have a look at the file
Documentation/proc.txt. It comes with the kernel source code.
[EMAIL PROTECTED] wrote in message ...
>
>Hi! Can you help me with setting this to different value.
>I have searched the /proc/sys/net and /proc/net dirs,
>but I have not hound anything usefull.
>
>
>Help :],
>
>
>In article <7qgnmt$[EMAIL PROTECTED]>, Robert_Glover wrote:
>>That would be because you are exceeding the maximum number of hops
>>(2^4 hops).
>>
>>[EMAIL PROTECTED] wrote in message ...
>>>Hi!,
>>>
>>>I have a cluster with 32 computers in a ring.
>>>The problem I have is, that, I can not access
>>>a computer, that is routed more than over
>>>15 computers.
>>>So, if I try to access 1->17, than the packets
>>>are lost (killed in computer 18).
>>>
>>>can anyone help me out here?
>>>
>>>
>>>---------------------------------------------
>>>traceroute to n17 (10.0.0.17), 30 hops max, 40 byte packets
>>> 1 n32 (10.0.1.32) 0.221 ms 0.12 ms 0.114 ms
>>> 2 n31 (10.0.0.31) 0.231 ms 0.186 ms 0.179 ms
>>> 3 n30 (10.0.1.30) 0.298 ms 0.25 ms 0.245 ms
>>> 4 n29 (10.0.0.29) 0.378 ms 0.318 ms 0.312 ms
>>> 5 n28 (10.0.1.28) 0.427 ms 0.4 ms 0.385 ms
------------------------------
From: "Chris L" <[EMAIL PROTECTED]>
Subject: Re: Help setting up Samba
Date: Tue, 31 Aug 1999 10:24:09 -0500
First make sure swat is installed, then type "man swat" and follow the
directions on editing /etc/inetd.conf and services, lastly restart inetd
(killall -1 inetd) and then connect to swat (http://192.168.0.1:901).
Chris L.
Justin Colson wrote in message <7qapft$bv6$[EMAIL PROTECTED]>...
>I have been following the mini howto at
>http://www.sfu.ca/~yzhang/linux/samba/toc.html but the Linux box is
>192.168.0.2, I subsituted it in where ever it has 192.168.0.1 in the howto
>but when I get to the SWAT stuff and it says type in http://192.168.0.1:901
>in your browser I type http://192.168.0.2:901 but get a server not found
>dialogue. Where am i going wrong?
>
>
------------------------------
From: Cliff <[EMAIL PROTECTED]>
Subject: Re: eth0: tulip.o: init_module: device or resource busy
Date: Tue, 31 Aug 1999 01:31:36 -0700
Hello Robert,
Instead of using modprobe which has more dependencies, you may try the
following: (for redhat 6.0)
insmod /lib/modules/2.2.5-15/net/tulip.o
and see what you come up with...
Cliff
Robert Inskeep wrote:
> I try starting eth0: whether in the bootup as a module OR at the command
> prompt (/sbin/ifup eth0) and I get a message OR modprobe tulip.o
> (whatever) and I receive the message
> 'init_module: device or resource busy'. What does this mean? Where do
> I go from here?
>
> --
> Robert Inskeep
> [EMAIL PROTECTED]
------------------------------
From: "Chris L" <[EMAIL PROTECTED]>
Subject: Re: Is this possible?
Date: Tue, 31 Aug 1999 10:57:54 -0500
Mike wrote in message
<[EMAIL PROTECTED]>...
>Is this possible?
>
>Here is the scenario I'd like:
>
<snip ASCII art>
>Now, I'd like it so that when Box B or Box C requests a page, it will be
>served from Box A's Apache server over the LAN. I'd like the web files
>to be stored on Box B's drive (and accessed through NFS or something),
>because it has a lot more room on its hard drive.
>
>When Box A requests the same page with its browser, I want it to be
>served from the internet.
>
>Boxes B and C should never connect to the internet.
>
>The same address needs to be used for all 3 browsers so that I don't
>need two different scripts (one for the web & one for at home), i.e.
>
>http://www.bigmweb.com/foo/bar.php3
>
>I'd copy the directory structure of my provider's drive on Box A (with a
>share from Box B) and set up everything the same way.
>
>
>Mike
I'm not quite clear on what you are trying to do... Are you trying to
set up Box A as the proxy server to B & C? If so then how about squid.
Chris L.
------------------------------
Subject: Re: Is this possible?
From: [EMAIL PROTECTED] (Mike)
Date: Tue, 31 Aug 1999 18:41:50 GMT
[EMAIL PROTECTED] (Chris L) wrote in <[EMAIL PROTECTED]>:
>
>Mike wrote in message
><[EMAIL PROTECTED]>...
>>Is this possible?
>>
>>Here is the scenario I'd like:
>>
><snip ASCII art>
>>Now, I'd like it so that when Box B or Box C requests a page, it will be
>>served from Box A's Apache server over the LAN. I'd like the web files
>>to be stored on Box B's drive (and accessed through NFS or something),
>>because it has a lot more room on its hard drive.
>>
>>When Box A requests the same page with its browser, I want it to be
>>served from the internet.
>>
>>Boxes B and C should never connect to the internet.
>>
>>The same address needs to be used for all 3 browsers so that I don't
>>need two different scripts (one for the web & one for at home), i.e.
>>
>>http://www.bigmweb.com/foo/bar.php3
>>
>>I'd copy the directory structure of my provider's drive on Box A (with a
>>share from Box B) and set up everything the same way.
>>
>>
>>Mike
>
>
> I'm not quite clear on what you are trying to do... Are you trying to
>set up Box A as the proxy server to B & C? If so then how about squid.
>
>Chris L.
>
>
>
No, not a proxy. I want to be able to test my PHP scripts at home without
having to upload them, then test them, then edit, then upload, then test,
then edit ad infinitum until they work right. I don't want to tie up my
phone line.
The other part is that I want the environment to look the same so that the
file structures and web addresses that the scripts use are the same as on
the "real" server that is on the web (my Internet Presence Provider).
Otherwise I have to have include in the scripts something like "if you're
running on this server, look 'here' for what you need, but if you're
running on that server, look 'there' for what you need." I don't want to
have to have that extra code in them. I also have to have two different
types of static pages: one that has links to local addresses and one that
has links to web addresses. In short, I want to cut down on my workload and
file space. It can get confusing, too.
For instance, here is a snippet of the type of code I have to use now:
$WebPath = "http://www.foo.com";
if($HTTP_HOST == "192.168.1.1") $WebPath =
"http://192.168.1.1/clients/foo";
As I read more, I think I can accomplish what I want through IP Aliasing.
Mike
------------------------------
From: "David Murray" <[EMAIL PROTECTED]>
Subject: Re: eth0: tulip.o: init_module: device or resource busy
Date: Tue, 31 Aug 1999 12:59:02 GMT
Check to make certain that /lib/modules/2.2.5-15/net/tulip.o actually exists
for one.. you can also try using a command:
depmod -a
which sometimes helps.. Also, If you have recompiled your kernel, make
certain to do a
make modules
make modules_install
This will setup dependencies and all that stuff for your network module that
may not have been setup before. Also, use the diag.exe (MS-DOS) program
that comes with your card to check that it works and verify it is setup
correctly.
One last thing.. check /etc/conf.modules and make certain you don't have too
much garbage in there related to the tulip.o or the eth0.
--DavidM
Robert Inskeep wrote in message <[EMAIL PROTECTED]>...
>I try 'modprobe tulip.o' and I receive the message
>'init_module: device or resource busy'.
>
>
>Additionally - when booting...in 'messages'
>Aug 31 06:15:39 localhost insmod: /lib/modules/2.2.5-15/net/tulip.o:
>init_module: Device or resource busy
>Aug 31 06:15:39 localhost ifup: Delaying eth0 initialization.
>Aug 31 06:15:39 localhost network: Bringing up interface eth0 failed
>
>What does this mean? Where do I go from here?
>
>Robert Inskeep
------------------------------
From: John-Paul Allen <[EMAIL PROTECTED]>
Subject: diamond homefree phoneline
Date: Tue, 31 Aug 1999 18:57:33 GMT
Hello. I have a 3com EhterLink XL PCI card(hooked up to the cable
modem) and a diamond homefree phoneline network card in my computer.
The 3com card is working just fine, but I cannot get the diamond card to
work. When linux loads, I see it initialize eth0(the 3com card) and
then it says 'delaying eth1 initialization'. I've tried loading the
pcnet32 module, but it says Device or resource busy. I have a feeling I
already know the answer to this, but is there a driver that I can use to
get the diamond homefree card to work? I'd rather just get some regular
network cards and run cat5, but that is not a possibility right now.
Thanks for any help.
------------------------------
From: "gonZo theGreaT" <[EMAIL PROTECTED]>
Subject: How to deny FTP-access to a whole domain
Date: Tue, 31 Aug 1999 18:41:26 GMT
Short question:
I'm using the wu-ftpd based FTP-daemon (BeroFTPd). Now I want to be able to
deny the FTP-access to several hosts. This works great by adding a file like
deny 209.209.1.11 /msgs/deny.msg
to my 'ftpaccess' in /etc.
But what shall I use to ban a whole domain ?
I tried
deny 209.209.*.* /msgs/deny.msg
I tried
deny 209.209. /msgs/deny.msg
I also tried to use an external deny file, put the denied hosts (domains) in
this, and added this file intead of the denied host (domain) to my
'ftpacces':
deny /*/deniedhosts /msgs/deny.msg
But this also only works for single descripted hosts, not for domains in any
way. :(
Thanx for any help I can get !
Gonzo
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Quick question 3C509B
Date: Tue, 31 Aug 1999 18:40:50 GMT
that is a good question that i would also like an answer to.
i know that 3com groups their nics into families with regard to windoze
drivers, so i would assume (i know what happens when you assume) that
the differences are not sufficient to warrant different drivers.
In article <7qh38l$k2a$[EMAIL PROTECTED]>,
"CNT" <[EMAIL PROTECTED]> wrote:
> Hi,
>
> Will Linux RedHat 6.0 support the 3Com 3C509B NIC. I only found
information
> regarding a 3C509, or IS it the same. I tought they are different...?
>
> Thanks in advance,
>
> Marty
>
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: s <[EMAIL PROTECTED]>
Crossposted-To: linux.samba
Subject: Re: Can't mount Windows drives
Date: Tue, 31 Aug 1999 16:21:44 GMT
please tell how u mounted linux from the win98 side thanks
[EMAIL PROTECTED]
mark wrote:
>
> Easiest is to ignore the other two responses you got ; )
>
> When you are at the smb: \> 'prompt', if you want to mount the share, simply
> type 'mount /mnt/win'. smbmount exits, you're back at your command line,
> and your drive is mounted at /mnt/win. Easy, eh?
>
> Hiawatha Bray <[EMAIL PROTECTED]> wrote in message
> news:7qcp4c$[EMAIL PROTECTED]...
> > I got my Samba working and I can mount my Linux drives on my Windows
> > machine. Now...how do I do it the other way around, mounting Windows
> drives
> > on my Linux box? I'm using RH 6 and Samba 2.0.3, withWin 98 on the
> Windows
> > machine.
> >
> > I created a /mnt/win mountpoint and then tried the smbmount command.
> Here's
> > what I got...
> >
> > [root@linux watha]# smbmount file://mypc/c /mnt/win
> > Added interface ip=10.0.2.16 bcast=10.255.255.255 nmask=255.0.0.0
> > Server time is Sun Aug 29 21:42:48 1999
> > Timezone is UTC-4.0
> > security=share
> > smb: \>
> >
> > I don't even know what this means...is this the correct response? Thanks.
> >
> >
------------------------------
From: "Jan Geertsma" <[EMAIL PROTECTED]>
Subject: Re: What domain do I use?
Date: Tue, 31 Aug 1999 18:09:25 +0200
Robert Inskeep <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]...
> people. I have only a small network. One win98, One NT and trying One
> Linux. I had not set the system up with a domain. I set it up with
> workgroups. What do I use for the domain in the adapter setup?
any crap you want ... Jan
> --
> Robert Inskeep
> [EMAIL PROTECTED]
------------------------------
From: "Torsten Pinkert" <[EMAIL PROTECTED]>
Subject: Re: Force D-Link 530-TX to use 100 MBit / Half Duplex
Date: Tue, 31 Aug 1999 15:30:39 +0200
Cliff <[EMAIL PROTECTED]> schrieb in im Newsbeitrag:
[EMAIL PROTECTED]
> alias eth0 via-rhine
> options via-rhine full_duplex=0 debug=0
My card still "detects" it's capabilities from the remote side and obviously
decides to use full duplex.
If I boot my WindowsNT first (with forced 100MBit/half duplex) and my linux
box after that, everything works fine and ifconfig eth0 showes 0 dropped
packets.
If I boot my linux box first and NT after that I get lots of dropped packets
resulting in about 50k/sec throughput, which is horrible for 2 Pentium2 PCs
with 100 MBits Network...
------------------------------
From: Sender <[EMAIL PROTECTED]>
Subject: Error message: tcp/telnet:unknown service
Date: Sun, 29 Aug 1999 23:01:15 -0400
Hello all:
Has anyone seen this error message before?
telnet: tcp/telnet: uknown service
I receive this message *only* as an ordinary user on a RH 6.0 machine.
Telnet works fine as root.
The machine can accept telnet connections from another computer.
All the other services, such as www and ftp, work fine for the user
account.
There is nothing changed, or misconfigured (as far as I can tell), in
/etc/services or inetd.conf.
Any help / suggestions would be greatly appreciated.
TIA
Sender
fairfiel -AT- dundee dot Net
------------------------------
From: Andrew Williams <[EMAIL PROTECTED]>
Subject: Re: samba passwd encrypt question
Date: Tue, 31 Aug 1999 18:10:41 +0200
Reply-To: [EMAIL PROTECTED]
well: it should work, and it does for me.
My understanding was originally:
clients requiring encryption (win95b, win95c, win98 and winNT) go to the
smbpasswd file for validation, clients that do not require it (WfWg 3.11, win95)
say that they cannot handle it and get validated normally.
Someone then told me this was not true and claimed that win95 + WfWg clients can
also handle encryption if necessary.
This difference is only important if your password files get out of sync, I have
been validating older (and newer) clients against smbpasswd for a year or so.
I suppose you could have a look at section 4.4 on my web-page - it lists some
other potential problems.
best of luck, Andrew Williams
David Murray wrote:
> I am sorry if this has been asked before but I can't find anything on it..
> I have a mixed environment of Windows 95, Windows 98, Windows NT, and Linux.
> (Well, Macs too, but they are not important to this question) I have Samba
> running on Linux. Now.. Windows 98 and NT like encrypted passwords to be
> on. So I turn on password encryption in samba and setup an /etc/smbpasswd
> file and create the users, etc.. works great.. now the 98 and NT systems can
> share files. However, now the 95 workstations can't! They won't ever get
> past the password prompt. So I turn password encryption off and now the
> opposite is true.. Windows 95 works but 98 and NT don't.
> For now I have gone into the registry of my NT and 98 systems and
> enabled plaintext passwords and that resolves it for now. However, I'd like
> Samba to go ahead and use encrypted passwords on NT and 98 and somehow know
> to use plaintext on 95. Is there any way to add encryption to Windows 95?
> There has to be a solution to this problem.
> --DavidM
--
Mielipiteet omiani - Opinions personal, facts suspect, especially on my
http://www.germanynet.de/teilnehmer/101/69082/samba.html
Simple Samba Solutions web page. ICQ 1722461
__________________________________________________________
| Fight Spam! Join EuroCAUCE: http://www.euro.cauce.org/ |
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
------------------------------
From: [EMAIL PROTECTED] (bill davidsen)
Crossposted-To: comp.os.linux.development.system
Subject: NFS caching
Date: 31 Aug 1999 19:48:35 GMT
Has anyone done anything to reduce network traffic between client and
server when the data on the server is known not to change? I'm being
very vague because I don't want to scare off any good solutions.
What I'd like to do is set aside a GB or so of local disk as a cache
area for that data which is known not to change. Unfortunately I need a
"per mount" solution, since there are other mount points which change
normally.
--
bill davidsen <[EMAIL PROTECTED]> CTO, TMR Associates, Inc
"So let it be written, so let it be dumb." Pharaoh Dufus the last...
------------------------------
From: "Pat Crean" <[EMAIL PROTECTED]>
Subject: Re: CPU needed to make use of 100 mbit
Date: Mon, 30 Aug 1999 13:58:07 -0400
David Murray <[EMAIL PROTECTED]> wrote in message
news:mDzy3.178$[EMAIL PROTECTED]...
> I upgraded my Linux server with a 100 mbit network card. However, I've
> noticed my Windows 95 clients are only getting about a 300% boost in file
> copy performance. Now the server is a 90 Mhz pentium but the workstations
> are 350 mhz or better. (Server isn't used very heavily) but I was
wondering
> why I didn't see closer to a 10 fold increase.. I theorize that the 90 mhz
> chip simply isn't fast enough to be able to fully utilize 100 mbits.. is
> this correct or are there other things I haven't thought of?
> --DavidM
>
>
Most likely suspect is your hard drive --- most drives (especially those
likely to be found in Pentium-90 systems) will only sustain 30-40 mbits/sec,
and you've got to add seek time and rotational latency on top of that.
------------------------------
From: Francisco =?iso-8859-1?Q?Jos=E9?= Toledano =?iso-8859-1?Q?Alcal=E1?=
Subject: ppp and samba
Date: Mon, 30 Aug 1999 22:01:46 +0200
I've got a network class C (192.168.0.0/255.255.255.0).
I've got a LInux box as Firewall, masquerade server and mgetty server
with one modem.
I've got a volume share with samba server on linux box.
I've got a remote client with Windows.
In Window's cliente, I'd created one ppp conection to the linux server.
The conection start ok. pppd on server have a proxyarp option. I can do
ping to windows client and windows client to linux server too.
The client can't browse or connect to share volume on samba server.
=BFAny idea about it?
Thank you very much and best regards from C=F3rdoba(Spain).
------------------------------
From: <[EMAIL PROTECTED]>
Subject: Re: Network card - newbie
Date: Tue, 31 Aug 1999 16:31:02 GMT
Paul Fredlein wrote:
>
> Hi,
>
> I've just bought Ret Hat 6 & a D-Link DFE-530TX 10/100 PCI Adapter.
>
> The problem, as you have guessed, is that the card is not listed. What
> do I do? I've downloaded a .c file for the card but don't know what to
> do with it.
You need to download the via_rhine.c file, compile it, and use that as
your driver. See
http://cesdis.gsfc.nasa.gov/linux/drivers/index.html#fast. The resulting
.o file should be in the /lib/modules/[kernel version]/net directory.
Contact your local LUG (Linux User Group). I'm not an expert on this, but
I was tinkering with the same problem earlier (though I had to stop
because of other reasons).
>
> Windoze tell me that it's io is 0xb800 but all the books suggest it
> should be on 0x300 or 0x310.
>
> I suppose I have to re-compile the kernel but don't know where to start.
No. NIC drivers can be used as modules, i.e. no need for recompiling the
kernel. Use linuxconf to tell the OS to use the module.
> Most of the books tell you all about somebody's wonderful windowing
> thing and not what you really need to know. Bill Ball's "Using Linux"
> seems OK but nothing about my basic problem.
>
>
> Eventually I want to run the Linux box as a file server for my NT box
> and my Mac. Is AppleTalk automatically included? (if I ever get it to
> see my card).
>
> BTW, why does Linux list hundreds of monitors when important stuff like
> network cards only get a couple listed and then not much documentation
> on what to do if your card's not there.
Monitors are easier to deal with and configure than NICs.
>
> Thanks,
>
> Paul
E-mail [EMAIL PROTECTED] if you need more help (remove
"nospam.")
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: "Scott MacDonald" <[EMAIL PROTECTED]>
Subject: neighbor table overflow? Error?
Date: Tue, 31 Aug 1999 15:34:04 -0500
Has anyone had this error before and know what it is? I'm getting it on one
of our RedHat 6 servers. Thanks,
Scott
------------------------------
From: Howard Mann <[EMAIL PROTECTED]>
Subject: Re: eth0 and modules
Date: Tue, 31 Aug 1999 15:31:03 GMT
Brian wrote:
> does any know how to get an ethernet card to work in RH 5.2?
>
> --
> Brian Savacool
> [EMAIL PROTECTED]
>
You may find this useful:
http://www.xmission.com/~howardm/ethernet.html
Cheers,
Howard Mann.
================== Posted via CNET Linux Help ==================
http://www.searchlinux.com
------------------------------
From: Tom Eastep <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.misc
Subject: Re: networking slows down
Date: Tue, 31 Aug 1999 20:41:44 +0000
Janet wrote:
>
> Tom Eastep <[EMAIL PROTECTED]> writes:
>
> > Janet wrote:
> > >
> > > Hi,
> > >
> > > I have been recently been experiencing a problem with my networking
> > > slowing down. If I ping my other machine, the ping time is normally less
> > > than 1 ms. However, occasionally (it has happened 2 or 3 times in the
> > > last week), it becomes a lot slower, sometimes taking up to 30 ms.
> > > However, if I just restart networking (using the network startup script in
> > > /etc/rc.d/init.d), it goes back to the sub-1 time. Any ideas?
> > >
> > > Janet
> >
> > I saw a similar problem on a DC21143-based NIC when used with the Tulip
> > driver (in my case, the ping time was 1-2 seconds). Switching to the
> > De4x5 driver seemed to avoid the problem.
>
> I have been using the same Linksys card with the Tulip driver for a year;
> it has just recently started having problems. Also, this only happens
> when the box has been up for a few days. Do you think it could still be a
> driver issue?
Those are exactly the symptoms that I saw...
-Tom
--
Tom Eastep \ Opinions expressed here
[EMAIL PROTECTED] \ are my own and not
Shoreline, Washington USA \ those of my employer
Work: [EMAIL PROTECTED] \________________________
------------------------------
From: [EMAIL PROTECTED] (David C.)
Subject: Re: linux box vs switched hub
Date: 31 Aug 1999 12:20:52 -0400
"Sagolsem C" <[EMAIL PROTECTED]> writes:
>
> What Switching hub are you guys talking about? I thought a switching
> hub is slightly more inteligent than the normal hub for the fact that
> it can maintain a table of the IP addresses of the host connected on
> its port. Therefore it can directly switch an incoming port trafic to
> the specific destination port rather than broadcasting the packets to
> all ports like a normal hub does. Moreovet when a communication takes
> place between two ports other free ports also can also communicate
> with other free ports simeltaneously thus increasing throughput.
Almost right. It automatically learns the Ethernet MAC addresses of
attached hosts and switches packets based on the destination MAC
addresses. While there are layer-3 switches that work based on the
contents of the IP header, they are much more expensive and complicated
devices.
> I also think a 10/100 switching hub can switch packets between 10baseT
> and 100baseT hosts.
Yes.
I think your confusion is in the terminology.
There are 10/100 layer-2 switches, which work exactly as you've
described. These are also known as bridges.
There are also dumb hubs that auto-detect the speed of attached hosts
and auto-configure for 10M or 100M operation. They are no different
than "normal" hubs, excepts that they can handle two different speeds.
These are sometimes referred to as "auto-sensing", "auto-switching", or
"switching" hubs, but they are not switches by any stretch of the
imagination.
-- David
------------------------------
** 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
******************************
