Linux-Networking Digest #29, Volume #10 Thu, 28 Jan 99 05:13:41 EST
Contents:
Re: 2.2.0 All network ports in use. (John Lellis)
Re: realaudio heavy CPU loading (Stephen E. Halpin)
Can't Telnet To Linux 5.2 from MS workstation (Joe Fialkowski)
Road Runner ("Trotter")
Re: DHCP acting funny? (Luca Filipozzi)
Re: SMTP problem, please help ("Marco Magagnini")
Re: PAP problem Please Help!! (Clifford Kite)
Nis & automount (Franck JEDRZEJEWSKI)
linux box as a router (Guido Dolci)
Re: 2.2.0 All network ports in use. (Juergen Heinzl)
Re: SAMBA help (David Kirkpatrick)
Accton EN1207D-TX (Dion Verbeke)
Re: Home network: What hardware on client and server? (Matthew Vanecek)
nlist-command gives problems ("Freek")
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (John Lellis)
Crossposted-To: comp.os.linux.setup
Subject: Re: 2.2.0 All network ports in use.
Date: 27 Jan 1999 14:22:07 -0600
John Lellis ([EMAIL PROTECTED]) wrote:
: I tried upgrading from kernel 2.0.36 to kernel 2.2.0 today. Base
: distribution was a Slackware 3.4 system.
: Everything seemed to run O.K. at first. Then I noticed that I couldn't
: telnet into the box without getting the message "All network ports in
: use."
: I'm guessing that the ttyxx devices are no longer used for telnetd and
: that I need to create new devices in /dev (or else upgrade my telnetd).
: But, I haven't been able to find anything in the 2.2.0 Documentation to
: suggest exactly what it is I need to do.
I may have just answered my own question. In Documentation/devices.txt,
I find the following reference under major number 4 (char):
Older versions of the Linux kernel used this major
number for BSD PTY devices. As of Linux 2.1.115, this
is no longer supported. Use major numbers 2 and 3.
which would suggest that I need to change my ttyxx device definitions in
/dev as I suspected. This would seem to preclude dual-booting 2.0.36 and
2.2.0 on the same system. Or am I missing something?
: Apologies in advance if this is a FAQ, but I haven't been following the
: 2.1.x kernels and so may have missed the structure changes which have
: been going on. Any assistance will be appreciated. A FAQ for upgrading
: to the 2.[1,2].x kernels would be ideal. Does such a thing exist?
--
John Lellis ([EMAIL PROTECTED])
--
... Our continuing mission: To seek out knowledge of C, to explore
strange UNIX commands, and to boldly code where no one has man page 4.
------------------------------
From: [EMAIL PROTECTED] (Stephen E. Halpin)
Crossposted-To: comp.os.linux.misc,comp.os.linux.setup,comp.os.linux.hardware
Subject: Re: realaudio heavy CPU loading
Date: Thu, 28 Jan 1999 07:42:38 GMT
On 27 Jan 1999 17:51:32 GMT, [EMAIL PROTECTED] (Eric Potter) wrote:
>[Posted and mailed]
>
>Kelvin Leung enlightened this group thus:
>> Hello,
>>
>> I got my RH 5.2 running with Ensoniq PCI sound card. I have the alsa
>> sound driver works fine. Both X11amp and Realaudio can produce sound
>> out. But the RealAudio heavily load my CPU (K6-166 overclocked to 200).
>> The audio is interrupted everytime I change to another window , or
>> browsing within Netscape. I got the CPU load monitor fire up and it
>> seems the CPU is fully occupied! I don't think it's the CPU problem 'cos
>> I don't have problem with X11amp, also the Realaudio (not G2) in Windows
>> in the same machine. Is that a fact or just my system doing weird!?
>>
>> Kelvin
>
>The sound driver in the 2.2.0 kernel performs much better. The problem with
>interruptions when switching between windows has been pretty much eliminated.
This falls in the IIRC catagory, but the same discussion occured some
time back in one of the FreeBSD newsgroups, where a new audio driver
was required to run the Linux version of Real Audio. Apparently
Real Audio uses a hook in the sound driver which wasnt fully functional
to determine how much data had really been emitted by the sound card
so it could synchronize the video motion with the sound track. This
seems to cause Real Audio to draw all the CPU you can give it. The
FreeBSD port of RealAudio 3.0 (no video) with the older drivers took
very little CPU time. Supposidly sound driver updates will address
the issue, so it would not be unlikely that some Linux sound drivers
may have similar issues.
>--
> * ^ \ ___@
> *^ / \ \ | \
> / \/ \ \__| \
>/ / ^ \ \
> / \ \ Eric Potter
> / ^ ^ \ \
>
-Steve
------------------------------
From: Joe Fialkowski <[EMAIL PROTECTED]>
Subject: Can't Telnet To Linux 5.2 from MS workstation
Date: 25 Jan 1999 00:06:48 GMT
Has anyone ran into this problem?
I cannot telnet or ftp to my linux 5.2 sever from a MS windows95
workstation. I can however ping the box and see it through smb. Please
Help!
Regards,
Joe Fialkowski
[EMAIL PROTECTED]
------------------------------
From: "Trotter" <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux
Subject: Road Runner
Date: Mon, 25 Jan 1999 00:14:17 GMT
I am trying to log onto Time Warner's Road Runner cable modem service with
Linux. Unfortunatly, Time Warner decided not to support Linux by making a
login program for it. However, they informed me that some people had made
their own linux login clients. If you have a client or know someone who
does, could you please send it to me? Thanks
- Trotter
------------------------------
From: [EMAIL PROTECTED] (Luca Filipozzi)
Subject: Re: DHCP acting funny?
Date: Sun, 24 Jan 1999 16:14:22 -0800
In article <[EMAIL PROTECTED]>, [EMAIL PROTECTED] says...
> I hope this doesn't turn out to be an NT problem :).
>
> I have been trying to set up DHCP server on a Redhat 5.2 machine.
> Eventually I installed dhcp-2.0b1pl8-1 but it didn't help. My dhcp.conf
> set up is the simplest possible for test purposes. I have one IP number
> associated with one MAC address. The first time it worked fine but then
> I changed the IP number in dhcp.conf and the DHCPOFFER is for the new IP
> number but the client requests (DHCPREQUEST) the number I gave it before
> and there is no DHCPACK (or whatever :) ). The clients are NT 4.0 sp3.
> Do I have to do something when I change the dhcp.conf on top of
> rerunning the daemon. How do I persuade the client to accept a new IP
> number?
> Any help very much appreciated,
> Raphael
>
>
I think you've got the Linux stuff just right. It's the clients that are
acting funny...
Try doing one of two things...
1) reboot the NT boxes
OR
2) type "ipconfig /release" to release the current IP number and
"ipconfig /renew" to get a new one
It's the client that is refusing the offer and wants to stick with the
old number.
Hope this helps,
Luca
--
Luca Filipozzi <[EMAIL PROTECTED]>
------------------------------
From: "Marco Magagnini" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.misc
Subject: Re: SMTP problem, please help
Date: Thu, 28 Jan 1999 09:42:16 +0100
I have had a similar problem on my server.
I was able to fix it by finding all pop end imap process and killing it.
After that all was OK.
Hope this help
Marco
Joewono Setiabudi ha scritto nel messaggio
<[EMAIL PROTECTED]>...
>I have a pop mail server (linux slackware) and 5 NT machines (5
>different users).
>The NT machines are using netscape for downloading emails.
>
>This configuration has been fine for months.
>Just recently, 2 of the users can not download their emails anymore,
>whenever they are trying to download them, I got an error message
>in netscape saying that "the mail server responded that
>/usr/spool/mail/username
>has been read" and after that, netscape just keep asking for a password.
>
>This error message is not because the mail file is not there or empty
>and
>it is not because network problem or password problem.
>
>There is no error message either in /var/adm/messages
>(smtp in.pop3d[1071] connect from xxxx
> smtp in.pop3d[1071]: Servicing request for username).
>
>Anybody knows how to fix this problem ?
>
>It seems that something is still locking the mail file.
>
>I tried to delete the mail file but it doesn't help either.
>In fact after deleting the mail file I got an error message saying can't
>
>generate output.
>
>This thing has made me totally confused since it only happens
>to those 2 users, the other 3 users work just fine.
>
>Please help.
>
>If possible, please email me at [EMAIL PROTECTED]
>
>Thanks
>
------------------------------
From: [EMAIL PROTECTED] (Clifford Kite)
Subject: Re: PAP problem Please Help!!
Date: 27 Jan 1999 15:08:19 -0600
Partha Sri ([EMAIL PROTECTED]) wrote:
: Hi All:
: I have been trying to get my PPP connection going for quite sometime
: now. But I am having the most bizarre problem. My ISP uses PAP and
: I dont have to issue a command to invoke PPP. But for some reason after
: establishing the serial connection my pppd does a LCP timeout. And
: I get fcs errors. I even spoke to my ISP guy and he says that his
: terminal server is just waiting and doing nothing after the connect.
: I even tried sending the login and passwd and got the same result. There
: is no noise in my phone connection so fcs errors is what I dont
: understand!! I checked the modem too.. it is fine.
: And btw I have a static IP address. And my ISP guy took the same scripts
: and it worked for him on a slackware kernel 2.0.34 and pppd 2.2.5.
: I use RH5.2, pppd 2.3.5 and my kernel is 2.0.36. But since I dont have
: X going I dont use RH's interface. I use a USR 56K v.90 external modem.
: I have my own set of scripts which are very simple. I have attached my
Good, simple I can handle, obsfucated I can't.
: options file and my log file for your reference. I am sorry about the
: lengthy log file but I am really desperate.. need help please!!
: And another interesting thing is that I am not able to connect to my
: ISP using minicom. After dialing and the hissing noise nothing happens!!
: Technically I shoud get the login prompt. This is happening when I dial
: to my other ISPs too. So, this is not an ISP issue. Something is wrong
: with my m/c.
: Also can anyone recommend a good 56K internal modem which will work with
: linux??
I have no recommendation except to say that an external is probably worth
the extra cost. I, of course, use an internal :X but while it serves me
fairly well, no recommendation.
: Any help will be greatly appreciated.
: Thanks a lot,
: Partha
: Here is my options file:
: ----------------------------------------------------------
: # comments allowed
: +pap
This isn't your current problem but unless you really need for the ISP to
authenticate itself to *you*, you had better remove +pap. Read "man pppd".
: /dev/modem 38400
: name partha
: -detach
: passive
: pap-timeout 120
: debug
: kdebug 25
: crtscts
: modem
: mtu 512
: asyncmap 0
: defaultroute
: 206.158.123.179:0.0.0.0
I would use 206.158.123.179: rather than specifing 0.0.0.0 . You may
need the pppd option ipcp-accept-remote if you choose not to follow
this advice.
: ------------------------------EOF----------------------------
: Here is my log file with prime time debugging:
: --------------------------------------------------------------
.
Selectivly edited..
: Jan 24 11:01:00 agni PAM_pwdb[8381]: (su) session opened for user news
: by (uid=9)
: Jan 24 11:01:00 agni PAM_pwdb[8381]: (su) session closed for user news
: Jan 24 11:16:14 agni pppd[8425]: pppd 2.3.5 started by root, uid 0
: Jan 24 11:16:14 agni pppd[8425]: pppd 2.3.5 started by root, uid 0
: Jan 24 11:16:15 agni chat[8429]: abort on (BUSY)
: Jan 24 11:16:15 agni chat[8429]: abort on (NO CARRIER)
: Jan 24 11:16:15 agni chat[8429]: abort on (ERROR)
: Jan 24 11:16:15 agni chat[8429]: abort on (NO DIALTONE)
: Jan 24 11:16:15 agni chat[8429]: abort on (NO ANSWER)
: Jan 24 11:16:15 agni chat[8429]: send (ATZ^M)
: Jan 24 11:16:34 agni chat[8429]: OK
Here is your problem. Note that the time between when ATZ is sent
to the modem and the time the modem responds with OK is 19 seconds.
This difference should be no more than 1 second, and means that your
Linux setup does not configure the serial port for the same IRQ that the
modem actually uses. Configuration is done in a boot-up script with
setserial, "man setserial".
: Jan 24 11:16:34 agni chat[8429]: -- got it
: Jan 24 11:16:34 agni chat[8429]: send (ATDT 687-2241^M)
: Jan 24 11:16:34 agni chat[8429]: expect (CONNECT)
: Jan 24 11:16:34 agni chat[8429]: ^M
: Jan 24 11:17:04 agni chat[8429]: ATDT 687-2241^M
: Jan 24 11:17:04 agni chat[8429]: CONNECT
: Jan 24 11:17:04 agni chat[8429]: -- got it
: Jan 24 11:17:04 agni chat[8429]: send (^M)
: Jan 24 11:17:04 agni pppd[8425]: Serial connection established.
: Jan 24 11:17:04 agni pppd[8425]: Serial connection established.
: Jan 24 11:17:14 agni pppd[8425]: Using interface ppp0
: Jan 24 11:17:14 agni pppd[8425]: Using interface ppp0
: Jan 24 11:17:14 agni pppd[8425]: Connect: ppp0 <--> /dev/modem
: Jan 24 11:17:14 agni pppd[8425]: Connect: ppp0 <--> /dev/modem
This doubling is strange, but it could be echoing caused by a modem
setting or a side-effect of using the wrong IRQ. Check the modem settings
with AT&V and shut off echoing if it is present (ATF1 ?).
: Jan 24 11:17:14 agni pppd[8425]: sent [LCP ConfReq id=0x1 <asyncmap 0x0>
: <magic 0x59bfc0a7> <pcomp> <accomp>]
: Jan 24 11:17:41 agni last message repeated 9 times
: Jan 24 11:17:44 agni pppd[8425]: LCP: timeout sending Config-Requests
: ----------------------------------------------------------------------------
This means that the ppp negotiations never really got started due to the
misconfigured IRQ..
--
Clifford Kite <[EMAIL PROTECTED]> Not a guru. (tm)
/* The signal-to-noise ratio is too low in many [news] groups to make
* them good candidates for archiving.
* --- Mike Moraes, Answers to FAQs about Usenet */
------------------------------
From: Franck JEDRZEJEWSKI <[EMAIL PROTECTED]>
Subject: Nis & automount
Date: Thu, 28 Jan 1999 09:42:15 +0100
I would like to know how to configure the automount with NIS (YP) ?
I have a NIS SUN server, and I have installed PCs under Linux Redhat.
I want to log on with my PC via the NIS server, with a home directory
automatically mount from the NIS server.
Apparently, the NIS work. I can ypcat, etc.
The autofs daemon is on. I can read the "autofs status" which list the
reference mount point,
but when i try to log on an account which is referenced on the NIS
server, I have a loggin but i have
no automount of the home directory.
How does it work ? Which files must be configured ? How automount starts
from autofs ??
Please reply to
[EMAIL PROTECTED]
[EMAIL PROTECTED]
Thanks,
------------------------------
From: Guido Dolci <[EMAIL PROTECTED]>
Subject: linux box as a router
Date: Wed, 27 Jan 1999 21:56:04 +0100
I need to set up a linux box as a router beetween two different
networks.
For example, I have a 10.20.30.0 nework and a 20.30.40.0 network and I
need that servers on first network can talk to servers of the second one
directly.
How can I set up a linux box for it?
Thanks
Guido
------------------------------
From: [EMAIL PROTECTED] (Juergen Heinzl)
Crossposted-To: comp.os.linux.setup
Subject: Re: 2.2.0 All network ports in use.
Date: Wed, 27 Jan 1999 21:19:34 GMT
In article <78nr2k$fkf$[EMAIL PROTECTED]>, John Lellis wrote:
>I tried upgrading from kernel 2.0.36 to kernel 2.2.0 today. Base
>distribution was a Slackware 3.4 system.
>
>Everything seemed to run O.K. at first. Then I noticed that I couldn't
>telnet into the box without getting the message "All network ports in
>use."
>
>I'm guessing that the ttyxx devices are no longer used for telnetd and
>that I need to create new devices in /dev (or else upgrade my telnetd).
>But, I haven't been able to find anything in the 2.2.0 Documentation to
>suggest exactly what it is I need to do.
See /usr/src/linux/Documentation/devices.txt ...
>Apologies in advance if this is a FAQ, but I haven't been following the
... not yet 8) ... take a look at the pseudo-tty slaves and I have not
found a MAKEDEV script yet, but you can write a tiny script pretty
fast (or adjust MAKEDEV).
According to the docs you glibc-2.1 is necessary to use the Unix98
style devices, so you might not change that right now.
Cheers,
Juergen
--
\ Real name : J�rgen Heinzl \ no flames /
\ EMail Private : [EMAIL PROTECTED] \ send money instead /
\ Phone Private : +44 181-332 0750 \ /
------------------------------
From: David Kirkpatrick <[EMAIL PROTECTED]>
Subject: Re: SAMBA help
Date: Wed, 27 Jan 1999 15:47:23 +0000
Reply-To: [EMAIL PROTECTED]
Check SMB password. See man for smbpasswd. Also review
encryption in smb.conf. MS and Linux must match on encryption.
Pete Stoloff wrote:
>
> I recently installed Samba-2 on my Linux box running RedHat5.2 I am
> able to connect to the Windows-98 and NT4 machines on the network.
> That is I can GET files from the Windows machines to Linux via
> smbclient. However, when I try to browse the Linux machines from
> Windows/NT, I get the message "\\LinuxBox not accessable. The network
> path was not found," when I click on the icon for the Linux box which
> does appear in Network Neighborhood. When I ping the Windows/NT
> machines from Linux this seems to work. I can even Telnet to the
> Linux box from Windows/NT. In addition, I edited the smb.config files
> to announce the Linux box to the IPs of the individual Windows
> machines, added passwords to what I think are the appropriate files.
> (I'm not sure which password files I need to edit and how.)
>
> Would some kind person please provide a "checklist" of the steps
> necessary to establish browsing/ 2-way Samba communication between
> Linux and windows. Thanks in advance.
--
[EMAIL PROTECTED]
[EMAIL PROTECTED]
[EMAIL PROTECTED]
------------------------------
From: Dion Verbeke <[EMAIL PROTECTED]>
Subject: Accton EN1207D-TX
Date: Wed, 27 Jan 1999 21:19:25 +0100
Hi there,
I just received a PCI network card (EN1207D-TX) from Accton. All is
working correctly under WINNT (as a test case) the diagnostic test came
through.
But...
I can't seem to get the network card running under linux. I tried the
link on their homepage, but it simply
did not work (I use kernel 2.2.0, they used kernel 2.0.34).
Which driver-set should I use on kernel 2.2.0 ? The link accton refers
to is outdated.
"mpx5030.c:v0.13 10/27/97 Donald Becker
http://cesdis.gsfc.nasa.gov/linux/drivers/mpx5030.html\n";
I'm with my hands in my hair, can someone help me out ?
Donald , please help ! :-0
Kind regards,
Dion
------------------------------
From: Matthew Vanecek <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.hardware,comp.os.linux.x,comp.os.linux.setup
Subject: Re: Home network: What hardware on client and server?
Date: Wed, 27 Jan 1999 15:20:58 -0600
Chris Bitmead wrote:
>
> I'm building a home network based on a server permanently connected to
> the internet, and a client workstation. The server will be a web
> server/email receiver that must always stay up. I'm not sure which good
> bits to put on the client and what on the server. The LAN will probably
> be 100MBS ethernet.
>
> The server can have a crappy graphics card, and maybe no monitor at all.
> The client will have a 21" monitor and a graphics card with lots of
> memory.
>
> Option 1: Should I plan to run everything off of the server and have the
> client as almost just an X Server? (Well the client would probably I
> guess have a swap disk and small local partition?). Will the performance
> of this suck? In other words should I build a really BEEFY server, with
> SCSI disks etc and a weak client?
>
Run your Internet servers (ftp, http, sendmail, etc) on the router and
firewall everything. If you only have the one IP address, also
masquerade the other boxes. Your workstation should be the beefy
machine; use your older computer for the Internet connection.
> Option 2: The middle option. Run all my apps on the client, but mount
> all my data via NFS from the server?
>
You could, but I wouldn't keep my data on the server. What if it gets
hacked? (it *could* happen)
> Option 3: Should I consider the server to be basicly just a router and
> internet server and put all my real data on the client? The performance
> may be better but it wouldn't allow me to access my data from other
> places on the internet if I need to. (Well I could, but that would mean
> leaving my client workstation on all the time too which seems silly.)
>
Well, if you have several Linux boxes, you could keep all your X apps on
one box, and run them over the net. Each box would have to have it's
own server, though, specific to the video card. If only one box +
internet access box, just keep all your X apps on your Workstation.
Just be aware that the machine connected to the net is vulnerable to
attack, and properly firewall your network, and you can pretty much do
what you want. If you are going to use 100M connections, then the
question of performance is moot; not much can move that fast, even SCSI
(well, SCSI cheap enough for us to own, anyhow! ;) ).
Each box needs to have at least enough on it to boot the machine and get
on the network. From there, you can do what you want.
As for accessing your workstation remotely, you can ssh to your server,
and from thence to the workstation. Or you could use IP port forwarding
(ipportfw or ipchains), which forwards requests on a specific port to
the server, to the workstation. Also, I rarely turn any of my machines
off. I turn monitors off, but my machines run constantly. Most people
I know rarely turn their workstations off. At work, I only turn my
machine off on Friday evenings (well, aside from rebooting several times
a day, because it's Windows)
> My Gut says go with option 2. Can anyone elaborate on the performance
> implications?
>
> But according to which options, where should I bother with SCSI disks?
> If the server only has one disk and a small number of users connected
> (mainly just one) then I guess SCSI is probably overkill. And in this
> case being just a file server it shouldn't be much in need of a fast
> swap partition on a separate disk. So I'm wondering if SCSI is overkill.
>
SCSI should be used in any multi-tasking OS. This includes Linux, of
course. No, it's not overkill. IDE has come a long ways, though, but is
still not as good as SCSI. Use SCSI wherever you can, if you can afford
to. For me, my Linux workstation is all scsi, and my server and, umm,
storage machine? are both IDE, as is my NT box. I would be happier,
though, if I could have made them all scsi, but such is life (and my
pocketbook!)
In short, SCSI is much better for your system than not-scsi.
> The client would have a bare OS and SWAP on the local drive, and most fs
> mounted on NFS so I'm guessing that SCSI would also be not too useful
> here.
>
Not true. IDE works best with OS's working in real mode, such as DOS.
Linux, and Unix in general, as well as NT (and maybe Windows, even
though it's not an 'operating system'), would benefit from scsi.
Although, as I said above, IDE has come a long way, and IDE makers
*claim* that it's (almost) as good as SCSI. (HAH!)
I would go as much scsi as you could afford to.
--
Matthew Vanecek
Studies in Business Computers at the University of North Texas
http://www.unt.edu/bcis
*****************************************************************
Visit my Website at http://people.unt.edu/~mev0003
*****************************************************************
For 93 million miles, there is nothing between the sun and my shadow
except me. I'm always getting in the way of something...
------------------------------
From: "Freek" <[EMAIL PROTECTED]>
Subject: nlist-command gives problems
Date: Thu, 28 Jan 1999 10:55:50 +0100
Hi,
can anyone please tell me why I can't ftp (from an ipmasqueraded client)
with sites who use the nlist command instead of the list -l command. When I
ftp to this sites from our linuxserver,it works.
Thanks
------------------------------
** 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
******************************
