Linux-Networking Digest #432, Volume #11 Sun, 6 Jun 99 17:13:32 EDT
Contents:
Networking problems caused by upgrade to Red Hat 6.0 (John Pfeifer)
Re: Mars_Nwe & real netware servers (norman elliott)
3c509 module question (Christian Aasland)
Re: MSERVER help needed! (norman elliott)
Re: IP masq works, but can't read news (Mike)
Re: Bandwidth friendly mail notifier ("Cliff")
Re: Redhat 6.0, kernel 2.2.5, IBM token ring == no system (Scott G. Parkerson)
Re: help - we're going berzerk! IP addresses for peers?! (norman elliott)
Re: About two lan card...... (norman elliott)
Re: DNS setup for NT/Wins and Unix mixed network (Marc Kandel)
Re: TCP/IP checksum formula (Lewis Perin)
PPP REconnecting strange ([EMAIL PROTECTED])
Complex DNS/MAIL setup (andreas palsson)
DNS setup for NT/Wins and Unix mixed network ("Mindspring News")
Re: ftp lag help (David Efflandt)
Re: DHCP problem (Marc Kandel)
Re: FTP login problem (David Efflandt)
Re: Forwarding of broadcasts possible? (David Efflandt)
----------------------------------------------------------------------------
From: John Pfeifer <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux,alt.linux,alt.os.linux
Subject: Networking problems caused by upgrade to Red Hat 6.0
Date: Sun, 06 Jun 1999 20:10:52 GMT
I recently upgraded from Redhat 5.2 to Redhat 6.0 and have experienced
problems trying to configure my ethernet card for the @home network. I
have an SMC EtherEZ (8416) ISA card that was functioning properly under
RH 5.2. After upgrading, my connection no longer works. I have tried
going through the steps I used to configure my connection under RH 5.2.
pnpdump > intel.conf
modify IRQ and BASE (IRQ is 10 and BASE is 0x0240)
uncomment necessary lines
isapnp intel.conf
Executing the isapnp command gives me the following errors:
Fatal resource conflict allocating IRQ10 (see /proc/interrupts)
Fatal error occured executing requrest 'IRQ 10' --- further action
aborted
I checked the /proc/interrupts file to make sure that IRQ 10 is not
being used
/proc/ioports tells me that the address 0x0240 is the correct address
How do I fix this problem? Any help would be greatly appreciated.
Thanks,
John
------------------------------
From: norman elliott <[EMAIL PROTECTED]>
Subject: Re: Mars_Nwe & real netware servers
Date: Sun, 06 Jun 1999 20:57:09 +0000
george m hoffman md wrote:
> this question has come up a lot in the last week.
> the most liekly problem is that the ipx frame types are not the same for
> the real netware servers and for the mars server.
>
> 'ifconfig' will tell you which frame types your linux box is running.
>
> nwserve.conf specifies which frame types mars_nwe will run. Some
> unmodified config files default to ethernet_ii.
>
> net.cfg on the client boxes (presumably DOS) will specifiy which frame
> types the clients will use. If not specified, often defaults to 802.2.
> win95 ipx services use 'autoconfigure' which should detect all ipx frame
> types.
>
> I bet your clients aren't using frame type which mars is using.
> I suggest changing your mars server to the primary frame type of your
> existing networks (usually 802.2) unless you want to reconfigure all
> your clients. most people think the frame types 'should' be ethernet_ii
> to co-exist with IP better.
>
> --
> George M Hoffman MD
> Anesthesiology & Critical Care
> Medical College of Wisconsin
> Children's Hospital of Wisconsin
> 414-266-3388 (voice)
> 414-266-3563 (fax)
> <[EMAIL PROTECTED]>
The default frame type for Netware 3.xx servers is 802.3
for Netware 4.xx it is 802.2
Netware 5 defaupt is IP not IPX.
best wishes,
norm
------------------------------
From: Christian Aasland <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux.questions,comp.os.linux.setup,linux.redhat.install
Subject: 3c509 module question
Date: Sun, 06 Jun 1999 19:35:30 GMT
Got a question for you experts:
I have three network cards in my linux box: 3c905b, 3c900, and
3c509. The first two are PCI and use the Boomerang (3c95x) driver, the
last one uses the 3c509 driver.
I installed RedHat 5.2 and modified lilo to include the 'append eth0
= ether,0,0,0 eth1=ether0,0,0 eth2=0,0,0' (that may not be the correct
syntax, I'm retyping that from memory). All three cards are working
fine.
Then I recompiled the kernel, upgrading to 2.3.3 and including the
3c95x and 3c509 drivers (and everything else I need to run ipchains/ip
masquerading). I did the 'make bzImage; make dep; make clean; make
modules; make modules_install' but now the 3c509 card isn't detected.
I changed lilo so I can boot to either configuration, and all the
cards there are still functioning, so I'm assuming there's not a
hardware problem (bios = no pnp, etc). I had this working on a prior
install.
Any ideas what I should do?
--
Christian Aasland
[EMAIL PROTECTED]
AIM:caasland
------------------------------
From: norman elliott <[EMAIL PROTECTED]>
Subject: Re: MSERVER help needed!
Date: Sun, 06 Jun 1999 21:13:23 +0000
Ken McCord wrote:
> I'm trying to use the MSERVER software to set up a linux box to act as a
> dial-out modem server (the box is currently used as a dial-in server using
> MGETTY). I've installed the MSERVER software and configured the
> /etc/mserver.conf file correctly (as far as I can tell). Once I reboot the
> box and try to 'telnet localhost 5800', I get:
>
> Trying 127.0.0.1...
> Connected to localhost.
> Escape character is '^]'.
> DEVICE INACCESSIBLE
> Connection closed by foreign host.
>
> Can anyone shed some light on this??
>
> Thanks,
>
> Ken McCord
Hi,
I know nothing about mserver. However the 127.0.0.1 is your loopback device
and you
need a dial out connection. Normally I would use ppp and this would give a
ppp0 interface.
This would be to my ISP. If this is what you want then you either need to set
up mserver
to make the connection or use it with ppp to get the connection.
Hope this helps,
best wishes, norm
------------------------------
From: Mike <[EMAIL PROTECTED]>
Crossposted-To: news.software.nntp
Subject: Re: IP masq works, but can't read news
Date: Sun, 06 Jun 1999 19:57:42 +0000
Barrett Richardson wrote:
> Try this.
>
> Telnet to port 119 again
> after the response type this
>
> mode reader
OK. Here's the response to that. By the way, there is no local echoing
of my keystrokes, if that provides any debugging help.
200 wnb1 InterNetNews NNRP server INN 2.2 21-Jan-1999 (posting ok).
>
> group news.software.nntp
Again, another response:
211 572 57752 58447 news.software.nntp
>
> xover
Another response, but here's where it craps out:
224 data follows
and then lots of nothing!
> Best I remember there were significant difficulties with doing a
> application level proxy with NNTP (which is kind of off base because
> the IP masq is not application level). If the above experiment does
> not yield an article it is likely an excercise in futility trying to
> make it work. If it does, it would be prudent to experiment with
> some different readers.
Well, I did experiment with booting up Linux on the box in question and
running: Netscape, tin, and KDE newsreader. They all do roughly the same
thing: they all find the number of messages, but never begin downloading
the headers or message bodies.
> Not a tremendous amount of help but something to gander at.
I greatly appreciate any and all help. Thank you.
--
Mike Wright
------------------------------
From: "Cliff" <[EMAIL PROTECTED]>
Subject: Re: Bandwidth friendly mail notifier
Date: Sun, 06 Jun 1999 20:17:32 GMT
Look for fetchmail. On RedHat there is an rpm package for it. It should
do what you want.
--
-Cliff
Views expressed are my own and not necessarily those of my employer
Concordia Net, Inc. When replying via email please use; cwheat at concordia
dot net not
root@localhost
Chris Severn wrote in message <7jag45$k4q$[EMAIL PROTECTED]>...
>I'm chasing a client/server which can do mail notification with minimum use
of bandwidth.
>
>I would like a mail client to access a mail server over a wireless link
which doesn't like lots of packets being sent over the link. The client
would have to be Windows95/98/NT, and the server preferably Linux (Windows
NT if absolutely necessary).
>
[snip]
>
>Thanks.
>
>Chris Severn.
>--
>Delete the 'x' to remove the spamblock.
>Except spammers, for whom my email address is abuse@localhost
------------------------------
From: [EMAIL PROTECTED] (Scott G. Parkerson)
Subject: Re: Redhat 6.0, kernel 2.2.5, IBM token ring == no system
Date: 6 Jun 1999 19:36:18 GMT
Reply-To: [EMAIL PROTECTED]
On Thu, 03 Jun 1999 13:06:59 -0400,
Amelia A. Lewis <[EMAIL PROTECTED]> wrote:
>PCI drivers ... but I wish the Turbo cards worked (are they the
>"pinegrove" code?).
The Turbo ISA cards will work. You just have to use LANAID or LANAIDC to
tell them to look like an Auto 16/4 compatible card. Eventually, the
enhanced support for Turbo adapters to work in their native mode will
be rolled into the ibmtr.c driver -- Paul Norton is purportedly working
on it as my nimble fingers type this followup.
>'Zackly. Apparently, from the notes at linuxtr.net, this is primarily a
>problem at IBM, due to the nature of the internal network.
Yeah. There's some errant packet or two floating around that's a weird
length that decides to hose everything.
>2.2.8 is supposed to have a bad networking bug in it, isn't it?
Umm... maybe. :/ My Linux box at work is more of a test-hobby-sidework thing.
Hence, it's probably always not running the recommended kernel. But, 2.2.8
is the first kernel where the bad slab magic patch appeared, for better
or worse.
>> // Scott Parkerson. Device Driver Development, IBM Network Hardware Div.,
>
>Hmmm ... *laugh* So is IBM's network division actually providing
>information to the nice folks at linuxtr.net?
Sort of. :)
> Seems like they ought to
>... give away the drivers, sell the hardware.
*grimace* Sisyphus, meet rock. Rock, meet Sisyphus. Maybe I'll detail my
account of convincing management to get on with the business of releasing
specifications semi-publicly with no NDA some day. Maybe not. :) At any
rate, I think the results have been good. And all it took was the information
to get into the right hands. (Thanks, Peter, Paul, and Mike! -- no folk
band pun intended)
Glad to be of some help,
scott
--
// Scott Parkerson. Device Driver Development, IBM Network Hardware Div.,
// Research Triangle Park, NC, USA. #include <std_disclaimer.h>
// ObTMBG: "I don't want the world... I just want your half."
// Notice: Please remove YOUR_SPLEEN from my e-mail address before replying.
------------------------------
From: norman elliott <[EMAIL PROTECTED]>
Subject: Re: help - we're going berzerk! IP addresses for peers?!
Date: Sun, 06 Jun 1999 20:52:00 +0000
rocket girl wrote:
> Somehow, both of you guys skipped right over the answer to my question. Is
> a hosts file of any kind REQUIRED or not? You are suggesting that I set up
> a WINS server - which would require IP addresses to be assigned to each
> client. B-z-z-z-z-z!!!! That is exactly what I am trying to avoid - using
> a hosts or lmhosts file. All we want to do is emulate a workgroup in
> windows - where all we have to do is specify a workgroup name and a machine
> name.
>
> Thanks again.
>
> 9wands <[EMAIL PROTECTED]> wrote in message <[EMAIL PROTECTED]>...
> >rocket girl wrote:
> >>
> >> Weve RTFM and still dont see the Linux box in NetworkNeighborhood.
> >>
> >> We have a simple Windows network - where machines can see each others
> shares
> >> via the workgroup (on a simple hub - no domain, no nothing but a few
> win95
> >> peers without IP addresses). Now we've added a Linux box... to be
> nothing
> >> more than another resource (don't really want validation, just a public
> >> share would suffice). But... we can only get the machines to ping when
> >> there is a hosts file on the win95ers, and we're not convinced that this
> is
> >> required. Either way, they still dont see ol' Linux in the Neighborhood.
> >>
> >> Is it absolutely required that the Win95 box has a lmhosts file with the
> >> Linux's IP? Why can't this thing resolve machine names on the workgroup
> >> without knowing explicit IP addresses?
> >>
> >> Please point us to the HOWTO file that explains what network foundation
> that
> >> smbd REQUIRES. The SMB howto suggests a IP config with the hosts
> files...
> >> sucky for us!
> >>
> >> Thanks !
> >
> >If you configure your Samba server for "user" level security and tell it
> >to function as a WINS server, then set your "Access level" in Windoze to
> >"User" and give it the name of your Linux box, Windoze will recognize it
> >as a "Domain Controller" and log the workstations into it allowing each
> >user to use their respective /home/* directory on the Linux box. I
> >haven't gotten much beyond this as my server acts primarily as a
> >fileserver/central point for backup, rather than a medium for
> >communications between the workstations.
> >
> >During the intial setup testing, you will probably want to disable
> >encrypted passwords in Windows. The Samba docs in /usr/doc/Samba* will
> >tell you how to edit the registry to do this. Then, after it's all
> >working read the docs about how to enable encrypted passwords AND CHANGE
> >IT BACK! While you may not feel the need for domain-style
> >authentication, I am sure you DON'T want your users impersonating each
> >other.
> >
> >Regards
> >
> >--
> >Beware the fury of a patient man.
> >
> > - John Dryden
Please read the previous posts again. the answers are there , honest
best wishes,
norm
------------------------------
From: norman elliott <[EMAIL PROTECTED]>
Subject: Re: About two lan card......
Date: Sun, 06 Jun 1999 20:43:53 +0000
Sang wrote:
> There are two lan cards in my computer and I try to setup them.
> But something happens that I counld get these card infomation by command
> "ifconfig".
> I couldn't ping to anyother address but my senond card address.Before(Just
> one lan card in my
> computer) that I could ping to anywhere.
> Could somebody tell me what happens and how to solve this situation??
> Please give me some opinion or methers to solve.
> Thanks very much!!
>
> Linux -- Slackware (kernel 2.0.36)
> And I have edit below files:
> /etc/lilo.conf # detect two cards both of the same time
> (append="ether=irq0,io0,eth0 ether=irq1,io1,eth1")
the line above is wrong.
you have to put the real irq, io address for each card.
they will NOT be irq0 or irq1
they will NOT be io0 or io1
the irq will have to be one which is not already in use. It could
perhaps be 3,5,6, 9 ,10 or 11. Each card will have to be different.
the io address will be somewhere in the range 240 to 3ff ( hex )
If you type the command
cat /proc/interrupts
you will see which are in use.
if you type the command
cat /proc/ioports
you will see which are in use.
You will need to set up each card to use irq and io addresses which are
not already in use.
then you can set up the append correctly.
You may have to do more than this. I am sorry I cannot help more
but I am no Linux expert.
hope this helps
norm
>
> /etc/rc.d/rc.inet1 # configure two cards (ifconfig .....) add to
> routing table ( /sbin/route add .......)
------------------------------
From: Marc Kandel <[EMAIL PROTECTED]>
Subject: Re: DNS setup for NT/Wins and Unix mixed network
Date: Sun, 06 Jun 1999 16:23:23 -0400
Jim,
I'm not sure of any answer but O'Reilly have never steered me wrong on anything.
Marc
Mindspring News wrote:
> We have about 30 Unix systems of various varieties and release levels. Also
> over 100 NT systems.
>
> Currently using NT as DNS; which also gets the Wins names. Not using DHCP
> except for dial-in.
>
> I am not the "official" administrator for this network, but have done pure
> Unix admin & network admin in a past life. We started having
> problems as the current admin guys took over and started rearranging the
> network. E.g. my AIX system could not receive a telnet or rlogin request.
> I deduced that AIX was getting the request and then doing some kind of
> reverse lookup to the nameserver and the nameserver was not responding
> properly (timing out, etc) and was likely going off network somehow. This
> went on for a week.
> Finally I set up a Unix system as a nameserver and the problems went away -
> sort of (I think) proving my theory. I tried to convince them to get rid of
> the NT DNS as a nameserver for Unix systems.
> Problem is the Unix DNS doesn't resolve the NT systems or Wins.
>
> So, the question is:
>
> Is there a way for the Unix DNS to go to the NT DNS and resolve names that
> are not in the Unix named.host and named.rev files. Can I get it to go to
> another Nameserver (on the same domain) and resolve names that it doesn't
> resolve? Or can I just have the Unix DNS be a "caching" DNS and get all of
> it's names from the NT DNS.
>
> I know, I know - we should just fix the problem with the NT DNS, but like I
> said, it's been a week or more and still not fixed - and work has to be
> done.
>
> I've been looking thru the O'Reilly DNS/Bind book and have some things to
> try but not sure if it'll work or not yet.
>
> Thanks,
>
> Jim Webster
------------------------------
From: Lewis Perin <[EMAIL PROTECTED]>
Crossposted-To:
comp.os.ms-windows.networking.misc,comp.os.ms-windows.networking.tcp-ip,comp.os.ms-windows.networking.win95,comp.os.ms-windows.networking.windows,comp.os.ms-windows.nt.admin.networking,comp.os.ms-windows.programmer.networks
Subject: Re: TCP/IP checksum formula
Date: 06 Jun 1999 15:54:51 -0400
> Mygorg wrote in message <7jdrnd$55v$[EMAIL PROTECTED]>...
> >Hi all
> >
> >I'd like to faind any info on TCP/IP checksum formula (TCP chksm and IP
> >chksm).
> >If someone know anything about this formula (or how checksum is precisely
> >calculated).
> >THX in advance.
>
> Take a look at RFC 1071. (www.faqs.org is one place where the RFCs can be
> accessed form)
There are also any number of places from which you can lift working
code. Linux is an obvious place to look. You can also use the
function computeIPChecksum in ntbgps.c in billgPC, my free Win32 bootp
client which you can download from <www.panix.com/~perin/>.
Cheers, Lew
------------------------------
From: [EMAIL PROTECTED]
Subject: PPP REconnecting strange
Date: Sun, 06 Jun 1999 17:51:44 GMT
Hi,
Can someone help me please with a strange behavior
I noticed
with pppd:
Usually /var/log/messages shows me the following
after
dropping down my PPP connection:
Jun 6 17:36:23 linux pppd[2911]: Terminating on
signal 2.
Jun 6 17:36:23 linux pppd[2911]: sent [LCP
TermReq id=0x2]
Jun 6 17:36:24 linux pppd[2911]: rcvd [LCP
TermAck id=0x2]
Jun 6 17:36:24 linux pppd[2911]: Connection
terminated.
Jun 6 17:36:24 linux pppd[2911]: Exit.
Jun 6 17:37:56 linux kernel: PPP: ppp line
discipline successfully unregistered
You should notice the time gap betwen the two last
log lines: 92 seconds!!!
This seems to be the reason why sometimes the
connection establishment
is cancelled, and this only because I reconnected
immediately after
disconnecting (i.e. without waiting up to 92
seconds...).
In this case I got the following log snip:
Jun 6 17:33:44 linux pppd[2537]: Terminating on
signal 2.
Jun 6 17:33:44 linux pppd[2537]: sent [LCP
TermReq id=0x2]
Jun 6 17:33:44 linux pppd[2537]: rcvd [LCP
TermAck id=0x2]
Jun 6 17:33:44 linux pppd[2537]: Connection
terminated.
Jun 6 17:33:45 linux pppd[2537]: Exit.
^^^^^^^^
Jun 6 17:34:22 linux pppd[2696]: pppd 2.2.0
started by i, uid 500
Jun 6 17:34:25 linux chat[2713]: timeout set to
60 seconds
Jun 6 17:34:25 linux chat[2713]: abort on (NO
CARRIER)
Jun 6 17:34:25 linux chat[2713]: abort on (BUSY)
Jun 6 17:34:25 linux chat[2713]: abort on (NO
DIALTONE)
Jun 6 17:34:25 linux chat[2713]: abort on (ERROR)
Jun 6 17:34:25 linux chat[2713]: send (+++ATZ^M)
Jun 6 17:34:25 linux chat[2713]: expect (OK)
Jun 6 17:34:26 linux chat[2713]: ATZ^M^M
Jun 6 17:34:26 linux chat[2713]: OK -- got it
Jun 6 17:34:26 linux chat[2713]: send
(ATDT010110191501^M)
Jun 6 17:34:26 linux chat[2713]: expect (CONNECT)
Jun 6 17:34:26 linux chat[2713]: ^M
Jun 6 17:34:50 linux kernel: PPP: ppp line
discipline successfully unregistered
^^^^^^^^
Jun 6 17:34:55 linux chat[2713]:
ATDT010110191501^M^M
Jun 6 17:34:55 linux chat[2713]: CONNECT -- got
it
Jun 6 17:34:55 linux chat[2713]: send (^M)
Jun 6 17:34:55 linux pppd[2696]: Serial
connection established.
Jun 6 17:35:02 linux kernel: PPP: version 2.2.0
(dynamic channel allocation)
Jun 6 17:35:02 linux kernel: PPP Dynamic channel
allocation code copyright 1995 Caldera, Inc.
Jun 6 17:35:02 linux kernel: PPP line discipline
registered.
Jun 6 17:35:02 linux kernel: registered device
ppp0
Jun 6 17:35:02 linux pppd[2696]: Using interface
ppp0
Jun 6 17:35:02 linux pppd[2696]: Connect: ppp0
<--> /dev/modem
Jun 6 17:35:02 linux pppd[2696]: sent [LCP
ConfReq id=0x1 <mru 1500> <magic 0x29ab97b1>
<pcomp> <accomp>]
Jun 6 17:35:05 linux pppd[2696]: sent [LCP
ConfReq id=0x1 <mru 1500> <magic 0x29ab97b1>
<pcomp> <accomp>]
Jun 6 17:35:06 linux pppd[2696]: Hangup (SIGHUP)
Jun 6 17:35:06 linux pppd[2696]: Modem hangup
Jun 6 17:35:06 linux pppd[2696]: Connection
terminated.
Jun 6 17:35:06 linux pppd[2696]: Exit.
- Does someone understand why things got wrong in
this case?
- Do I always have to wait the time until
"ppp line discipline successfully unregistered"
before
safely reconnecting PPP?
- Is there a way to make the kernel react more
rapidly
i.e. unregister more quickly?
Thanks for help,
Daniel.
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: andreas palsson <[EMAIL PROTECTED]>
Subject: Complex DNS/MAIL setup
Date: Sun, 06 Jun 1999 21:47:41 +0200
Hello.
I have been assigned the task to setup a combined DNS/MAIL-server at our
company.
They gave me two 486's and told me that they should have an identical
installation _but_ one of them should work as backup if the other
doesn't respond.
In other words;
If "DNS1" at ip XYZ.XYZ.XYZ.1 somehow crashes or does not respond then
"DNS2" at ip XYZ.XYZ.XYZ.2 should become "DNS1" at ip XYZ.XYZ.XYZ.1.
I've heard that this can be solved, but how? Is that up to programming
the router (Cisco router) or can this be done by configuring the 486's?
One other problem, is that these servers also act as mailservers. But if
DNS1 goes down, and DNS2 recieves mail and then suddenly DNS1 comes
back, then how I transfer the recieved mail on DNS2 to DNS1?
Any ideas, hints and references to information is very appreciated, and
if further explainations are needed, then don't hesitate to ask.
Sincerely...
Andreas
------------------------------
From: "Mindspring News" <[EMAIL PROTECTED]>
Subject: DNS setup for NT/Wins and Unix mixed network
Date: Sun, 6 Jun 1999 15:19:31 -0400
We have about 30 Unix systems of various varieties and release levels. Also
over 100 NT systems.
Currently using NT as DNS; which also gets the Wins names. Not using DHCP
except for dial-in.
I am not the "official" administrator for this network, but have done pure
Unix admin & network admin in a past life. We started having
problems as the current admin guys took over and started rearranging the
network. E.g. my AIX system could not receive a telnet or rlogin request.
I deduced that AIX was getting the request and then doing some kind of
reverse lookup to the nameserver and the nameserver was not responding
properly (timing out, etc) and was likely going off network somehow. This
went on for a week.
Finally I set up a Unix system as a nameserver and the problems went away -
sort of (I think) proving my theory. I tried to convince them to get rid of
the NT DNS as a nameserver for Unix systems.
Problem is the Unix DNS doesn't resolve the NT systems or Wins.
So, the question is:
Is there a way for the Unix DNS to go to the NT DNS and resolve names that
are not in the Unix named.host and named.rev files. Can I get it to go to
another Nameserver (on the same domain) and resolve names that it doesn't
resolve? Or can I just have the Unix DNS be a "caching" DNS and get all of
it's names from the NT DNS.
I know, I know - we should just fix the problem with the NT DNS, but like I
said, it's been a week or more and still not fixed - and work has to be
done.
I've been looking thru the O'Reilly DNS/Bind book and have some things to
try but not sure if it'll work or not yet.
Thanks,
Jim Webster
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Subject: Re: ftp lag help
Reply-To: [EMAIL PROTECTED]
Date: Sun, 06 Jun 1999 16:14:14 -0400
Sounds like a DNS problem. Make sure that you are NOT running 'named',
'gated' or 'routed' unless you know what they do and have them configured
properly. Linux tries everything in its power to identify any machine
connecting to it, so delays for misconfiguration are common.
If it is a local dialin or LAN, your problem is that the remote IP and
name are missing from your /etc/hosts (or DNS).
On Thu, 03 Jun 1999 15:28:04 -0400, matt <[EMAIL PROTECTED]> wrote:
>Hello, Im having a little bit of a problem ftping.. getting files is
>fine, but when I ftp to anyplace and try to put a file I get this
>halting lag which ends up pinging me out. Not just upload lag or
>anything, but the file starts uploading after a few seconds, and i just
>lagg until im just about dropped connection, then killing the ftp
>process i still remain lagged until i disconnect and redial up. This
>also happened before when someone tried to ftp a file from my box..
>
>I have no idea what could be causing that so I dont really know where to
>look, Any help or ideas would be great.
>
>Thanks
>[EMAIL PROTECTED]
>
--
David Efflandt [EMAIL PROTECTED]
http://www.xnet.com/~efflandt/
------------------------------
From: Marc Kandel <[EMAIL PROTECTED]>
Subject: Re: DHCP problem
Date: Sun, 06 Jun 1999 16:35:06 -0400
Heikki,
Unfortunately I can only confirm your post being seen.... It made it to the
US just fine =)
Good luck!!!!
Marc
Heikki Levanto wrote:
> Oh ye gurus,
> I have got stuck with this problem and need some fresh eyes to look at it
> and to see the obvious.
>
> I am trying to set up a dhcpd on my RedHat 5.2 machine. The server starts up
> without problems, but when my laptop connects, I get the following in
> /var/log/messages:
> Jun 3 18:01:43 bug dhcpd: DHCPDISCOVER from 00:00:e8:5e:a7:32 via eth1
> Jun 3 18:01:44 bug dhcpd: DHCPOFFER on 10.0.0.100 to 00:00:e8:5e:a7:32 via eth1
> Jun 3 18:01:44 bug dhcpd: sendpkt: Network is unreachable
>
> My dhcpd.conf is like this:
>
> option domain-name "lsd.dk";
> option domain-name-servers 10.0.0.1;
> option routers 10.0.0.1;
> subnet 10.0.0.0 netmask 255.0.0.0 {
> range 10.0.0.100 10.0.0.200;
> # host lab {
> # hardware ethernet 00:00:e8:5e:a7:32;
> # fixed-address lab.lsd.dk;
> # }
> }
>
> I can happily ping other machines through eth1:
>
> PING 10.0.0.2 (10.0.0.2): 56 data bytes
> 64 bytes from 10.0.0.2: icmp_seq=0 ttl=128 time=1.0 ms
> 64 bytes from 10.0.0.2: icmp_seq=1 ttl=128 time=0.7 ms
> 64 bytes from 10.0.0.2: icmp_seq=2 ttl=128 time=0.7 ms
>
> Ifconfig says:
>
> eth1 Link encap:Ethernet HWaddr 00:20:18:29:2E:8E
> inet addr:10.0.0.1 Bcast:10.255.255.255 Mask:255.0.0.0
> UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
> RX packets:80800 errors:0 dropped:0 overruns:0 frame:0
> TX packets:93252 errors:0 dropped:0 overruns:0 carrier:0
> collisions:168
> Interrupt:9 Base address:0xff00
>
> I am running an almost plain-vanilla RedHat 5.2, with two net cards (cable
> modem and local net), ip masq, etc. Pentium 133 (if I remember right), 64MB
> RAM, enough free disk:
>
> Linux version 2.0.36 ([EMAIL PROTECTED])
> (gcc version 2.7.2.3) #7 Thu Apr 8 16:46:51 CEST 1999
>
> I have made one change in /etc/rc.d/init.d/dhcpd: I start dhcpd only on
> eth1.
>
> Any ideas would be appreciated, even just a confirmation that this message
> has reached "the world", as my news system is not quite done yet...
>
> Thank you in advance
>
> Heikki Levanto
>
>
>
> --
> Heikki Levanto LSD - Levanto Software Development <[EMAIL PROTECTED]>
>
> --
> Heikki Levanto LSD - Levanto Software Development <[EMAIL PROTECTED]>
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Subject: Re: FTP login problem
Reply-To: [EMAIL PROTECTED]
Date: Sun, 06 Jun 1999 16:31:42 -0400
On Wed, 2 Jun 1999 17:00:29 -0700, bob cent <[EMAIL PROTECTED]> wrote:
>Hi,
>
>When I FTP to my RH Linux server with Netscape (command line syntax =
>ftp://username:[EMAIL PROTECTED]), the FTP root directory is the root of
>the disk rather than the user's home directory. Is this to be expected? Is
>there a "conf" file that I need to change so my users will see their home
>directories instead? Thanks.
>
>Bob.
I see what you mean. Maybe that is because apache (or is it Netscape) now
typically redirects to '/' if no dir is specified.
Try using a full path:
ftp://username:[EMAIL PROTECTED]/home/username/
'man 5 ftpaccess' may shed some light on getting it to chroot to the users
dir by default.
--
David Efflandt [EMAIL PROTECTED]
http://www.xnet.com/~efflandt/
------------------------------
From: [EMAIL PROTECTED] (David Efflandt)
Subject: Re: Forwarding of broadcasts possible?
Reply-To: [EMAIL PROTECTED]
Date: Sun, 06 Jun 1999 16:03:51 -0400
On a similar note, how would I get a dialin machine on a Linux box to see
smb broadcasts from the local LAN (like I can with a Cisco router using
the remote IP as ip-helper-address)?
Subnet 192.169.6.0
SMB broadcasts 192.168.6.255
Modem remote 192.168.6.21, local 192.168.6.1 (same as eth0 and proxyarp to
it). The remote only sees traffic to its own IP and does not see smb
broadcasts unless specifically plugged into /etc/lmhosts (which I am
doing now with ip-up.local, assuming remote machine name is same as login
name).
--
David Efflandt [EMAIL PROTECTED]
http://www.xnet.com/~efflandt/
------------------------------
** 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
******************************
