Linux-Networking Digest #407, Volume #11 Fri, 4 Jun 99 16:13:49 EDT
Contents:
recovering mail from qmail ? (Frank Lenaerts)
How many ethernet cards can I install under Linux? (roxu)
Cuantas tarjetas de red puedo instalar en Linux? (roxu)
Linux and ADSL with GTE (Joe Halpin)
Re: Setting up NE 2000 compliant card (Green Screen)
Re: 3Com 3CXEM556 and pppd: working but sloooooow... (David Hinds)
does ip masquerading log ip translations? ("lee berry")
Re: multiple 3com ethernet-card problem (Fulko van Westrenen)
IP Masq works with Linux box, but not WIN NT box ("Dan Drozd")
Re: Linux: ICMP Redirect, IP Source Routing unterdruecken (Gert Doering)
Re: unix ftp API ??? (Rich Mulvey)
NIS/yp problem with automount (Jonathan Foweraker)
Re: Howto Change Apache Layout? ([EMAIL PROTECTED])
Re: pump, dhcpcd, and redhat 6.0 (Francois Magnan)
Re: Setting up NE 2000 compliant card (Green Screen)
Re: Can I deny ordinary user to telnet? (Green Screen)
----------------------------------------------------------------------------
From: Frank Lenaerts <[EMAIL PROTECTED]>
Subject: recovering mail from qmail ?
Date: Fri, 04 Jun 1999 07:26:29 +0200
Hi,
I made a terrible mistake yesterday. While changing my DNS
configuration, I also changed the qmail 'me' control file to reflect the
new hostname, but, I forgot to change my fetchmailrc, which still used
the old e-mail addresses for each user. Now, while fetchmail was
downloading all mail from the different POP3 accounts, it forwarded them
to qmail but qmail could not handle the old e-mail addresses, so, the
mails stayed in the queue but did not get processed. After a while, I
saw that qmail processed all mail but all mail was still in the queue,
presumably because it could not deliver the mail and it will try again
later on or so.
Looking through the queue directory, I can see a lot of files
(numbered). With qmail-qread, I can also see the from and to addresses
(the to addresses are definitely wrong).
The question now is: how can I recover this mail from the queue so that
it gets delivered to the right e-mail address of each user? Should I
create an alias to forward mail sent to user1@localhost to
[EMAIL PROTECTED] or so? Will qmail ever try to deliver this
mail again, or did it give up? OR Can I get the numbered files out of
the queue manually by just copying them? How then, can they be read and
merged into the user Mailbox (mbox format) so that he can read his mail?
Thanks a lot,
Frank
P.S.: The example given above, about user1@localhost -->
[EMAIL PROTECTED], is real. With sendmail, I could use the
user login name e.g. user1 as e-mail address, as well as
user1@localhost, as well as [EMAIL PROTECTED], while with
qmail, I can only use [EMAIL PROTECTED] (i.e. fully qualified
names)
------------------------------
From: roxu <[EMAIL PROTECTED]>
Subject: How many ethernet cards can I install under Linux?
Date: Fri, 04 Jun 1999 17:42:21 GMT
Hello!
I have to install an internal router in my job. We have 6 network
segments routed by a Netware 3.11 and a linux box.
I wonder if i can install
the 6 ethernet cards in one computer with linux. This will make things more
easy for me.
I have read Ethernet and Net-3 and i have't seen any
limitation. I think also that if i can add ip aliases and there are several
logic interfaces, it may be also a lot of phisical ones.
Well, the question
is. If the hardware can, (i have 3 isa slots and 3 PCI), can Linux?
Thanks
in advance
--
Saludos desde Asturias
Spain
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: roxu <[EMAIL PROTECTED]>
Subject: Cuantas tarjetas de red puedo instalar en Linux?
Date: Fri, 04 Jun 1999 17:46:30 GMT
Hola a todos.
En el trabajo tenemos 6 segmentos de red. Actualmente estamos encaminandolos
con un Netware 3.11 y un Linux. Quiero simplificar esto (si es posible) con
dos linux o con uno solo si puedo instalarle las 6 tarjetas.
Mi pregunta es. Soportar�a Linux 6 tarjetas de red. Hay alguna limitaci�n a
esto?
Una suposici�n que hago es que si puedo hacer IP alias de forma que tengo
varios dispositivos l�gicos, no deber�a haber problema en tenerlos f�sicos.
El pc tiene 3 ranuras libres ISA y otras 3 PCI.
He revisado el Ethernet y el Net-3 Howto, pero no he encontrado ninguna
limitaci�n en este sentido.
Gracias por adelantado
--
Saludos desde Asturias
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Joe Halpin <[EMAIL PROTECTED]>
Subject: Linux and ADSL with GTE
Date: Fri, 04 Jun 1999 14:15:08 -0500
I'm attempting to help out some friends who have a small business.
They'd like to install a server that would allow more than one person in
the office to access the internet at the same time. I've suggested Linux
to run the server.
The problem is that they would like to get an ADSL line in for the
connection to the ISP, and their phone company is GTE. This is a problem
because the GTE support staff says they absolutely refuse to support an
ADSL line that's being used by a Linux box.
I couldn't get any kind of reason for this out of them, but I did see
some stuff on the web about GTE and Microsoft getting into bed together,
which probably explains it.
Could anyone here recommend another approach that would give equivalent
cost/performance? I'd like to prevent the "buy NT, and everything that's
needed to do anything useful with it" syndrome if I could. They're not
exactly made out of money.
Thanks
Joe
--
I didn't want to be here, where the future is in store
but my name is on the mailbox, and my key fits in the door
- Bob Bennett (the musician, not the lawyer)
------------------------------
From: Green Screen <[EMAIL PROTECTED]>
Subject: Re: Setting up NE 2000 compliant card
Date: 4 Jun 1999 06:41:59 GMT
Dunn One <[EMAIL PROTECTED]> wrote:
: I have an SMC Ezcard 10(PCI). How do I get this to work? I heard that I can
: use the NE 2000 module. Is there any program in Linux that I can use to do
: this? Please help. Thanks
use the ne.o module... load it with modprobe. at minimum, this driver
needs you to specify the io address... first thing i would try is 0x300...
then maybe 0x280... syntax is 'modprobe ne io=0x300'
Lint^^
------------------------------
From: [EMAIL PROTECTED] (David Hinds)
Crossposted-To: comp.os.linux.setup,alt.periphs.pcmcia,comp.os.linux.hardware
Subject: Re: 3Com 3CXEM556 and pppd: working but sloooooow...
Date: 4 Jun 1999 18:24:49 GMT
Rick Tait ([EMAIL PROTECTED]) wrote:
:
: So I got my 3Com 3CXEM556 working on my SuSE box. Kernel 2.2.5 with
: pcmcia-cs-3.0.11 modules. Both my eth0 and ttyS2 come up OK. There
: was a few problems with ttyS2 working on te default irq (3) which it
: seems it shares with the eth0 interface, so I changed to the irq to 0
: (setserial /dev/ttyS2 irq 0). That allowed me to use minicom on the
: modem, and of course ppp.
It's a bug in the kernel's serial device driver.
It can be fixed, if you want to rebuild your kernel (and PCMCIA), by
editing linux/drivers/char/serial.c and changing each use of
IRQ_T(info) to IRQ_T(state).
-- Dave Hinds
------------------------------
From: "lee berry" <[EMAIL PROTECTED]>
Subject: does ip masquerading log ip translations?
Date: Fri, 4 Jun 1999 16:57:34 +1000
I'm developing a server which runs on a closed network and will serve
requests for a different server (outside), trapped by IP masquerading. But I
need to know the IP that the client was trying to reach. Is there any way to
find this information? Does IP masquerading log translations?
lee
------------------------------
From: [EMAIL PROTECTED] (Fulko van Westrenen)
Subject: Re: multiple 3com ethernet-card problem
Date: 4 Jun 1999 07:03:44 GMT
In article <[EMAIL PROTECTED]>,
Fulko van Westrenen wrote:
>
>joaster:> ping 192.168.2.1
>PING 192.168.2.1 (192.168.2.1): 56 data bytes
>ping: sendto: Operation not permitted
>ping: wrote 192.168.2.1 64 chars, ret=-1
>........
Fixed it: Firewall problem.
Stupid!!!!!!!
--
Fulko van Westrenen
email: [EMAIL PROTECTED] (whatever the header may say)
------------------------------
From: "Dan Drozd" <[EMAIL PROTECTED]>
Subject: IP Masq works with Linux box, but not WIN NT box
Date: Fri, 4 Jun 1999 14:27:42 -0500
I have small 3 computer network. One is a SuSE Linux 6.0 (kernel 2.2.3)
running as a small web-server and doing IP Masquerading for the other two
computers. Everything had worked perfectly for three/four months. BTW, it
connects to an ADSL line. The other two computers are dual boot. One with
Windows 98 and RedHat 5.1 (kernel 2.0.35, I think), the other with WinNT SP4
and SuSE Linux 6.0 (kernel 2.2.5). This morning when I went started the NT
computer, it was not able to load web pages or retrieve my email, but I am
able to ping both internal and external hosts. Nslookup fails when it
cannot identify the DNS host name. I cannot just use IP addresses to bring
up pages, so it seems that the problem is not strictly DNS related. Also I
am unable to telnet or ftp. I can fix all of this and make everything work
on both windows boxes by unplugging the masquerading box from the hub and
assigning my "real" public IP address to either of the windows computers.
Then everything works fine.
Now for the interesting part. While none of these services work in
Windows(either NT or 98), everything works flawlessly(using masquerading) in
Linux, both SuSE and RedHat. None of the settings have been changed since
yesterday when everything was working and the IP addresses, netmasks,
gateways, dns addresses, etc are the same in both windows and linux.
I am not sure if this is at all relevant, but I can also use all
services through the computer which is generally the masquerading/ web
server box, so long as I, obviously, disconnect the Windows machines from
the hub or change their IP addresses back to private 192.168.1.* addresses.
Any help would be greatly appreciated.
thanks
Dan
------------------------------
Crossposted-To: de.comm.internet.routing
From: [EMAIL PROTECTED] (Gert Doering)
Subject: Re: Linux: ICMP Redirect, IP Source Routing unterdruecken
Date: Fri, 4 Jun 1999 19:46:53 GMT
[EMAIL PROTECTED] (Detlef Bosau) writes:
>> >Ich stimme mit Dir ueberein, dass Routing ausgedient hat.
>> Jetzt wuerde mich dann doch mal Deine Alternative zu "Routing"
>> interessieren.
>Ist jetzt ziemlich off topic. Aber ich meine schon seit langem,
>dass die Alternative zu einem gerouteten, paketierten, Datenverkehr
>in einem Datenverkehr besteht, in dem Peers, die verbindungsorientierten
>Datenverkehr moechten (und das ist IMHO doch der haeufigere Fall)
>direkt ueber eine virtuelle Verbindung miteinander kommunizieren.
>Wie in einem PVC bei FR oder ATM.
>Hier wuerden keine paketweisen Routingentscheidungen fallen,
>auch die Paketinspektion waere deutlich einfacher als bei IP Packet
>Switching. Ich denke schon, linked source routing Verfahren wie
>bei ATM oder Frame Relay, bei dem Routing eigentlich ein
>irrefuehrendes Wort ist, weil eben keine Wegewahl_entscheidung_
>getroffen wird sondern bei Verbindungsaufbau durch Signalisierung
>verabredete Verbindungstabellen abgegriffen werden, wuerden die
>Kommunikation erheblich erleichtern.
Letztlich ist das doch genauso "Routing". Ich "route" halt nicht anhand
von IP-Adressen, sondern anhand von FR- oder ATM-Zieladressen, die vorher
(beim VC-Verbindungsaufbau) ausgehandelt wurden.
Der Vorteil davon ist doch nur, dass die "Routingtabelle" schoen klein
ist, und die Geraete damit mit wenig Hardware-Aufwand schnell darin suchen
koennen.
Macht "Netflow-Switching" auch so, ist eine nette Idee, skaliert leider
ueberhaupt nicht, wenn man Netze wie das *Internet* betrachtet, wo mehr
als nur "ein paar 1000" Flows / SVCs parallel ueber einen groesseren
Knoten laufen.
>Es gaebe dann schlicht kein TCP mehr. Es erscheint mir persoenlich
>seit langem widersinnig, verbindungsorientierte Leitungen auf
>Layer 2 kuenstlich auf einem Layer 3 in Pakete zu zerteilen und
>auf Layer 4 wieder Verbindungen zu simulieren.
Hmmm. Also einer von uns hat das mit dem Layering jetzt nicht voellig
verstanden - ich habe keinen Dipl.-Inf., also vermutlich ich.
- was macht ATM/FR, wenn eine Zelle verloren geht? Dumm schauen, weil
man dafuer eine Verbindungsssicherung *braucht*.
- was macht man bei so einem Protokoll mit Daten, die zu gross fuer
eine einzelne Zelle sind - man muss "fragmentieren" und die
Datenpakete am Ende wieder zu einem Stream zusammensetzen.
Ob Du das jetzt in Deiner privaten Terminologie Layer 2 oder Layer 4
nennst, ist doch relativ irrelevant - solange "ganz unten" etwas liegt,
was Pakete / Zellen / Datagrams verschickt (d.h. letztlich, mehrere
parallele und unabhaengige Verbindungen ueber eine Leitungen multiplext),
brauche ich "oben" etwas, was die Verbindung aufbaut, abbaut, und
sichert. Und das ist halt ganz traditionell Layer 4...
Von der Notwendigkeit fuer "Verbindungen" mit minmalem Overhead, z.B. fuer
DNS-Abfragen, mal abgesehen. Eine neue VC fuer eine DNS-Anfrage zu
schalten, grenzt an Wahnsinn (und einfach mal alle VCs, die man zum
DNS-Server hat, offenzuhalten, auch).
>Ein bedarfsmaessiger Verbindungsaufbau zwischen den Datenstationen
>mit einem NBMA mit linked source route switching, wie ich
>das eigentlich immer nenne, ich finde immer kein vernuenftigeres
>Wort dafuer, auch in der Literatur habe ich dazu nichts gefunden,
>halte ich persoenlich fuer organischer, fuer verwaltbarer (siehe z.B.
>QoS), leichter fehlerkorrigierbar (zwischen den Knotenrechnern
>wuerde ein Synchroncode mit Fehlerkorrektur laufen) als
>die bisherigen Protokollstapelungen, in denen sich einfach
>viele Dinge wiederholen.
Fehlerkorrektur zwischen Knotenrechnern ist eine nette Idee, aber
relativ sinnlos - wenn ein Knotenrechner absemmelt, und der hat gerade
irgendwelche Daten im RAM, sind diese weg. Ergo designed man den
Protokollstapel so, dass ein "best effort"-Transport voellig ausreicht,
und faehrt gut damit.
gert
--
[EMAIL PROTECTED]
18 24 61 B 17 17 4
------------------------------
From: [EMAIL PROTECTED] (Rich Mulvey)
Crossposted-To:
comp.unix.programmer,comp.os.linux.development.apps,it.comp.linux.development
Subject: Re: unix ftp API ???
Reply-To: [EMAIL PROTECTED]
Date: Thu, 03 Jun 1999 03:35:20 GMT
On Wed, 02 Jun 1999 14:17:20 +0200, Max Liccardo <[EMAIL PROTECTED]> wrote:
>Hi folks,
>I'm looking for a free unix ftp API. I have to develop an application
>for sending and retriving files via ftp.
>Any links are apreciated.
Sure. Use Perl's Net::FTP. You can't get any quicker or easier to
use.
- Rich
------------------------------
From: Jonathan Foweraker <[EMAIL PROTECTED]>
Subject: NIS/yp problem with automount
Date: Fri, 04 Jun 1999 13:38:56 -0600
Hi all,
I have a problem - similar to one described above - but using automount,
so the suggested /etc/fstab
solution won't work.
I have a Sun Sparc network which runs NFS and yp. I've recently moved a
home filespace to a Linux 5.2
machine. This filespace is in exports with the (rw) option.
I can cd /home/username without any problem. This is set-up on the yp
server using auto.home.
My problem is creating/modifying files on that machine.
If I log-on locally I can use the file space with no problem. If I set
the permissions to rwxrwxrwx then
I can modify the filespace locally and when it is mounted using
automount. If I change the permissions
to a more logical setting rwxr-xr-x, then I get "Permission Denied'
messages. Further, if I do create
files they are nobody.nogroup owned.
Oh - yp is running smoothly, and on the remote machine (which is hosting
the home directory) I can
ypcat passwd without any problems.
It seems that somewhere the uid/guid of the user who mounts the remote
partition is being masked.
Any solutions/suggestions?
Thanks,
Jonathan Foweraker.
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: Howto Change Apache Layout?
Date: Fri, 04 Jun 1999 18:49:46 GMT
In article <JgL53.958$[EMAIL PROTECTED]>,
"NOspam Mackraz" <[EMAIL PROTECTED]> wrote:
> I'm looking for:
> - pointer to information on how linux starts up daemons (and the role
> of rc.d)
The Linux rc structure is SYSV-ish, which is to say overly confusing and
in general complete crap. That is, of course, only my opinion :)
The startup script you found in /etc/rc.d/init.d is one of possibly
several scripts. I would do an `ls -lR > before.ls` in /etc/rc.d before
and `ls -lR > after.ls` after you add the package, then do a `diff
before.ls after.ls` to see which scripts were added. That's usually the
best surefire way to know what a package adds [without having to read
through possibly hundreds of lines of code].
> - suggestions on how to run an httpd I built from sources
Simple.
cd <location-where-httpd-is>
./httpd -d <path-to-server-root>
The server root [in case you plan on *really* mangling the setup :)], is
one level above the "conf" subdirectory. So, if you put your
configuration files in /web/ServerRoot/conf [for example], then your
server root is /web/ServerRoot.
> - perhaps an APACI "layout" configuration that matches redhat 5.2
> - general guidance
Since apache has never seemed to have "gotten it together" on how a
webserver directory hierarchy should be laid out, I took it upon myself
to come up with a layout that was very easily replicated among my
production machines [regardless of platform]. It varies quite a bit
from the "standard" layout [whatever that happens to be this release],
but it has worked wonderfully for me.
/web/ - webserver root [this is what I tar up for replication]
/web/ServerRoot/ - the server root
/web/ServerRoot/bin/ - location of httpd
/web/ServerRoot/conf/ - location of configuration file[s]
/web/ServerRoot/modules/ - location of modules [shared objects]
/web/ServerRoot/logs/ - location of logfiles
/web/DocumentRoot/ - the document root
/web/DocumentRoot/vhost1/ - virtual host document root
/web/DocumentRoot/vhost2/ - virtual host document root
/web/scripts/ - log rotation scripts, etc..
I also make everything owned by user web in group web [80:80], and give
this user sudo access to run [root-owned] start and stop scripts for the
webserver [since httpd binds to a priveledged port].
That's just one suggested layout, but if you're planning on setting up
an environment that's to be used by several different people who have
differing levels of access and expertise, I highly recommend it.
--
-Bill Clark
Systems Architect
ISP Channel
http://locale.ispchannel.com/
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
Subject: Re: pump, dhcpcd, and redhat 6.0
From: [EMAIL PROTECTED] (Francois Magnan)
Date: Fri, 04 Jun 1999 19:51:31 GMT
Try to remove the cache file "/etc/dhcpc/dhcpcd-eth?.cache" before
calling "dhcpcd". This worked for me.
Francois
On 06/02/99, Matthew Levine wrote:
>(snip)
>I'm trying to go back to dhcpcd but I keep getting timeouts. Is
there
>any particular reason why dhcpcd would fail with the 2.2 kernel or
could
>it be that it is entirely unrelated to the kernel version?
>(snip)
--
______________________________________________________
Francois Magnan
Departement de Mathematique & Statistiques
Universite de Montreal
email: [EMAIL PROTECTED] (MIME, NeXTMail Ok!)
------------------------------
From: Green Screen <[EMAIL PROTECTED]>
Subject: Re: Setting up NE 2000 compliant card
Date: 4 Jun 1999 06:58:03 GMT
Green Screen <[EMAIL PROTECTED]> wrote:
: Dunn One <[EMAIL PROTECTED]> wrote:
: : I have an SMC Ezcard 10(PCI). How do I get this to work? I heard that I can
: : use the NE 2000 module. Is there any program in Linux that I can use to do
: : this? Please help. Thanks
: use the ne.o module... load it with modprobe. at minimum, this driver
: needs you to specify the io address... first thing i would try is 0x300...
: then maybe 0x280... syntax is 'modprobe ne io=0x300'
: Lint^^
actually... heh... just re-read your post... it's a PCI card, just use the
ne2k-pci module... 'modprobe ne2k-pci', after you've compiled and
installed it of course...
Lint^^
------------------------------
From: Green Screen <[EMAIL PROTECTED]>
Subject: Re: Can I deny ordinary user to telnet?
Date: 4 Jun 1999 07:09:50 GMT
Natta <[EMAIL PROTECTED]> wrote:
: Can I deny ordinary user telnet to server?
sure... there are a few ways to do it... check out /etc/login.access, or
try using /bin/false as the default shell for the user (make sure you add
/bin/false to /etc/shells)
Lint^^
------------------------------
** 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
******************************
