Linux-Networking Digest #66, Volume #12 Sat, 31 Jul 99 02:14:37 EDT
Contents:
DHCPD Question (NAdams5755)
Re: SMC1211tx PCI nic does not work !! HELP PLEASE. ("John W. Lemons III")
Re: Squid + News = Problem (Alex Harrington)
Re: How do I generate RST on close()? (Stephen Satchell)
Re: Services shall only react to one IP (Bernd Eckenfels)
Re: SIOCSFIFLAGS: Resource temporarily unavailable - Ethernet card problem
([EMAIL PROTECTED])
Re: HELOOOWWW (George Dau)
Re: having problems with leafnode (I think) (Vidar Andresen)
Re: ping has incorrect IP address? ("Andrey Smirnov")
Re: ipportfw/ADSL problems (Greg Weeks)
Re: Hardware Q: 10Mb DSL Router to 100Mb network (Greg Leblanc)
Re: PAM authenticating to Novell NDS or NT Domain? (Pascal Fleer)
Re: ppp - ping OK but no telnet/ftp/netscape - answer. (Michael Hayes)
sendmail configuration ? (tykq)
Re: Reverse Proxy & Load Balance + Redundancy (stanislav shalunov)
Re: Help: ipchains and FTP problem (ip_masq_ftp IS installed) (Bill Steiner)
Re: OOPS! Security Question (Chris Trudeau)
----------------------------------------------------------------------------
From: [EMAIL PROTECTED] (NAdams5755)
Subject: DHCPD Question
Date: 31 Jul 1999 03:24:24 GMT
Evening,
I have a DHCP server running at work. All hardware is set up correctly. I'm
running RH 5.2 on a P200 with an Intel 10/100 NIC.
I am DHCP'ing Windows 9x clients on a LAN.
All the infomation (that I can see) is correct in my /etc/dhcpd.conf file. I
could be wrong and probably am.
My question is about setting up IP lease times. When ever I connect to my ISP,
the 'Lease Obtained' and 'Lease Expired' times & dates are Tue Jan 1, 1980,
12:00 AM. My clients are getting a 1 hour lease time when I set my min & max
lease times to 0. The lease time starts when the client obtains an IP address
and renews after an hour.
How do I edit my dhcpd.conf to get the 1-1-1980 dates for the clients? Is
there a command line switch?
More info is available if needed.
Thanks,
-Nick Adams
------------------------------
From: "John W. Lemons III" <[EMAIL PROTECTED]>
Subject: Re: SMC1211tx PCI nic does not work !! HELP PLEASE.
Date: Fri, 30 Jul 1999 22:50:17 -0500
I just installed that same card under redhat 6.0 about a week ago. I
followed the instructions in the Ethernet HOWTO at the Linux Documentation
Project page. I also used the same source code you used, from SMC.
Everything worked great. What exactly are you trying that isn't working?
John N <[EMAIL PROTECTED]> wrote in message
news:7ntor2$6q6$[EMAIL PROTECTED]...
> Need help please,
> I very "green" in linux, being a DOS/WINDOWS believer for a long
> time... my SMC1211tx PCI nic card does not work. I've downloaded the
source
> code for this driver from smc for this nic card, but I do not know how to
> make it all work.
>
> Can you please HELP thanks a mil...
> John Nunez
>
>
>
>
------------------------------
From: Alex Harrington <[EMAIL PROTECTED]>
Subject: Re: Squid + News = Problem
Date: Thu, 29 Jul 1999 21:12:49 +0100
Just a thought but I thought that proxy servers were usually limited to
HTTP FTP (Sqiud can also do ICMP) but not NNTP or POP3 or SMTP.
If Squid is suppost to be able to proxy these protools, ignore this!
News is transported using NNTP. If Squid can not proxy NNTP, your
requests will never get further than Squid!
If this is the case, you need to run a local newsserver or use a
different method of sharing the net connection such as IPCHAINS etc..
Hope this Helps
Alex ([EMAIL PROTECTED])
[EMAIL PROTECTED] wrote:
>
> I have the same problem and I didn't find the answer yet.
> I don't want to install a news server on my server.
> I just want to access an external news server from my internal network.
> I would appreciate any help.
>
> Ovidiu Dressler
>
> In article <[EMAIL PROTECTED]>,
> [EMAIL PROTECTED] wrote:
> > That's not working, but thanks.
> > I can reading news if I disable the Proxy in my browser.
> >
> > Matthias
> >
> > "Gero H. Marten" schrieb:
> >
> > > Hello Matthias,
> > >
> > > how do you read news? Online or do you have a local newsserver, e.g.
> > > INN?
> > > Just an idea: with INN, put the address of the providers newsserver
> > > in your /etc/hosts.
> > > --
> > > Gero H. Marten
> > > <http://www.provi.de/gmarten/index.html>
> > > --
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.
------------------------------
Subject: Re: How do I generate RST on close()?
From: [EMAIL PROTECTED] (Stephen Satchell)
Date: Fri, 30 Jul 1999 22:10:01 GMT
Responses below, with appropriate snippage:
[EMAIL PROTECTED] (Malware) wrote in
<[EMAIL PROTECTED]>:
>Hi Stephen,
>
>> 2) Server issues a setsockopt(2) specifying a linger time of zero.
>> (This per Stevens.)
>
>Probably you cope best with closing the socket without lingering. There
>will be no RST then but the socket on the client will disappear within
>2*MSL, while on the server it does disappear right after receiving the
>last ACK from the client. The 2*MSL will not be a serious problem to the
>client other than if it does need to reuse the same source port or does
>create a establish amount of connections within a short time.
My server has the option of doing the close with and without the
SO_LINGER thing. The server reacts identically to both cases. (It gets
worse when I try to use shutdown(2), by the way...my server gets into an,
er, interesting state.)
In any case, this is in a testing environment and the two-minute wait for
the server to wake up and smell the coffee is killing me. Or, rather,
it's killing my customer.
>>[snip]
>
>I assume the server does not send a RST because there is no data lost
>during the close. Instead it does go half trough the normal
>three-way-handshake.
One hundred percent correct. Indeed, data transfer has usually stopped
by the time the server is told to kill the link to the commercial
equipment. There is no way for my server to *not* eat data such that a
reset would be sent.
>From tcp_close() of kernel 2.3.10:
>
>...
> /* As outlined in draft-ietf-tcpimpl-prob-03.txt, section
> * 3.10, we send a RST here because data was lost. To
> * witness the awful effects of the old behavior of always
> * doing a FIN, run an older 2.1.x kernel or 2.0.x, start
> * a bulk GET in an FTP client, suspend the process, wait
> * for the client to advertise a zero window, then kill -9
> * the FTP client, wheee... Note: timeout is always zero
> * in such a case.
> */
> if(data_was_unread != 0) {
> /* Unread data was tossed, zap the connection. */
> tcp_set_state(sk, TCP_CLOSE);
> tcp_send_active_reset(sk);
> } else if (tcp_close_state(sk,1)) {
> /* We FIN if the application ate all the data before
> * zapping the connection.
> */
> tcp_send_fin(sk);
> }
Now for the real question: would adding the test
if ((sk->linger) && (!sk->lingertime)) {
tcp_set_state(sk, TCP_CLOSE);
tcp_send_active_reset(sk);
} else
just before the code quoted above be the appropriate hack to the kernel
code? Comments?
------------------------------
From: Bernd Eckenfels <[EMAIL PROTECTED]>
Subject: Re: Services shall only react to one IP
Date: 30 Jul 1999 21:24:27 GMT
Frederik Hermans <[EMAIL PROTECTED]> wrote:
> Is it possible to force Apache, Samba, Wu-FTPd, sendmail, pop3d and ircd
> to react only to one IP? If so, how? ;-)
Depends on the service. Services you start from inetd can be bound to one ip
if you use xinetd or a patched inetd or some other super servers (g2s is
cool).
Some Services like ircd and FTP-Servers and apache support virtual homing an
can therefore be configured to bind to a specific ip.
If a program does not have a config option to do so, u can always use a one
liner in the source. bind is used to do this.
Greetings
Bernd
------------------------------
From: [EMAIL PROTECTED]
Subject: Re: SIOCSFIFLAGS: Resource temporarily unavailable - Ethernet card problem
Date: Fri, 30 Jul 1999 22:19:29 GMT
Similar problem was resolved by turning PnP OFF/DISABLED in Windoz,
and compiling the latest driver for the card.
Other suggestions : If PCI card - is PCI configured in Linux ?
Check IRQ's and address carefully for conflicts.
On Fri, 30 Jul 1999 14:20:49 -0700, Sriram Mudulodu
<[EMAIL PROTECTED]> wrote:
>
>> Hi,
>>
>> I have a netgear 10/100 Mbps PCI card on HP pavilion PC. It seems to
>work
>> fine with Windows.
>>
>> With Linux the card is detected at boot time. When I type 'ifconfig', it
>> does not show the config for eth0. ('ifconfig eth0' however shows it).
>> When I try 'ifconfig eth0 up' the following message is seen :
>>
>> SIOCSFIFLAGS: Resource temporarily unavailable.
>>
>> I tried reinstalling Linux but it doesn't work.
>> (The same card works on another machine with Linux on it)
>
>If you know how to fix, please do e-mail me.
>
>Thanks a lot,
>Sriram
------------------------------
From: [EMAIL PROTECTED] (George Dau)
Subject: Re: HELOOOWWW
Date: Sat, 31 Jul 1999 04:29:45 GMT
"ricK" <[EMAIL PROTECTED]> wrote:
]Can enyone see my posts to this newsgroup, why wont anyone answer my post.
Supply a real e-mail address. Replies to your current address just bounce.
--
,-,_|\ George Dau - Unix (Solaris, DEC Unix, Linux), Oracle, Internet. __
/ * \ Home: [EMAIL PROTECTED] ! Views/opinions above need (00)
\_,--\_/ Work: [EMAIL PROTECTED] ! not be those of MIM or the ( \/ )
v WWW: http://www.pobox.com/~gedau ! Carpentaria Buffalo Club. W--W
------------------------------
From: [EMAIL PROTECTED] (Vidar Andresen)
Subject: Re: having problems with leafnode (I think)
Date: Sat, 31 Jul 1999 00:43:59 GMT
In article <[EMAIL PROTECTED]>,
[EMAIL PROTECTED] (B'ichela) wrote:
[Mailed and posted, it make sense here]
> I am having a problem. Every sinced I switched my hostname to
>my dynamic domain name. Pinkrose.dhis.org. and installed leafnode
>1.10b2 No one on my system can get news out past www.remarq.com! That
>service is who my ISP outsourced its news to. My Uplink never uses
>passwords and the news articals DO get posted but then they don't seem
>to go farther than that machine. I cannot seem to find any replies to
>any of my postings or questions.
You are seen.. This may interest you, i split the 'Path: *' line:
Path: juliett.dax.net!newsfeed1.swip.net!swipnet!hermes.visi.com!
news-out.visi.com!feed.newsfeeds.com!newsfeeds.com!
nntp-relay.ihug.net!ihug.co.nz!remarQ60!rQdQ!supernews.com!
remarQ.com!corp.supernews.com!not-for-mail
From: [EMAIL PROTECTED] (B'ichela)
Newsgroups: comp.os.linux.networking
Subject: having problems with leafnode (I think)
Date: Thu, 29 Jul 1999 16:15:08 -0400
Organization: Posted via Supernews, http://www.supernews.com
Lines: 23
Message-ID: <[EMAIL PROTECTED]>
Reply-To: [EMAIL PROTECTED]
Summary: other than www.reqmarQ.com. any other news sites gettting this?
Keywords: leafnode, dynamic Domain Names, news not getting out
X-Complaints-To: [EMAIL PROTECTED]
User-Agent: slrn/0.9.5.7 (UNIX)
Xref: juliett.dax.net comp.os.linux.networking:104171
> If anyone has ideas I sure could use a few. If anything it
>will tell me that the news articals are going somewhere.
It does. You could get yourself another news-server to check if what
you send from one server is getting received by others.
Mvh Vidar Andresen
------------------------------
From: "Andrey Smirnov" <[EMAIL PROTECTED]>
Subject: Re: ping has incorrect IP address?
Date: Fri, 30 Jul 1999 19:28:32 -0700
Hello,
Depending on name resolution method you use, you may need to check your
hosts file on NT or DNS database for name of your Linux box and make sure
you have the right IP.
On NT the hosts file should be in C:\WinNT\system32\drivers\etc (assuming NT
is installed on C:).
Good luck
joe befumo <[EMAIL PROTECTED]> wrote in message
news:X7ro3.924$[EMAIL PROTECTED]...
> Okay -- I'm on pretty thin ice both in terms of Linux and networking, so
> forgive me if this is a simplistic problem, but here goes:
>
> I recently installed Red Hat 6.0 with the intent of replacing my NT
primary
> domain controller. The installation went absolutely flawlessly (that
should
> have raised a red flag, I suppose 8^)
>
> As soon as the machine was up I tried pinging it from my Primary Domain
> Controller, and vice versa & everything was fine. I installed Samba,
> following the instructions in several excellent step-by-step guides
> referenced elsewhere in this conference, as well as Gerald Carter's "Teach
> Yourself Samba in 24 Hours" (in other words I did RTFM).
>
> At this juncture I have mounted two directories from two different NT
boxes,
> and can move files back an forth from the Linux machine. Hey, I'm on a
> roll!
>
> Next I tried to see my exported Linux directories from the NT machines --
no
> good. After a bit of head scratching, I tried pinging again, and guess
> what: now it doesn't work! Obviously, I must have screwed something up
> along the way, but after some poking around I just can't quite find it.
>
> Now, I CAN ping the Linux IP address (192.168.1.123), however, when I try
to
> ping it by name, the message tells me that it's trying to ping
192.168.0.0,
> and then tells me that "Destination specified is invalid"
>
> This sounds like it should be something simple to someone who knows what
> they're doing -- any suggestions?
>
> Thanks.
>
> joe
>
>
------------------------------
Reply-To: [EMAIL PROTECTED]
From: [EMAIL PROTECTED] (Greg Weeks)
Subject: Re: ipportfw/ADSL problems
Date: Fri, 30 Jul 1999 21:07:05 -0500
In article <yMlo3.428$[EMAIL PROTECTED]>,
"Michael Wheeler" <[EMAIL PROTECTED]> writes:
> I've got a RH 5.2 machine that has been running IP masq
> successfully for some time now. I would like to be able to
> ICQ, and thats my problem. I understand I can use ipportfw
> to forward to the internal machine, but my ADSL IP is dynamic,
> and I can't seem to find the correct parameter or ipportfw to put in
> my rc.firewall. The only info I found in the How To was for PPP,
> which I believe is ip-up.
http://www.users.uswest.net/~gdweeks/DSL-howto.html
is what I did. I'm not forwarding for ICQ, just web and mail.
Greg Weeks
--
http://durendal.tzo.com/greg/
------------------------------
From: Greg Leblanc <[EMAIL PROTECTED]>
Subject: Re: Hardware Q: 10Mb DSL Router to 100Mb network
Date: Sat, 31 Jul 1999 01:41:20 GMT
Did you set the Cisco 675 to 10 meg? My I have my 675 plugged into my
Intel 10/100 meg card with a cross over cable, and it's running at 100
mbit, full duplex. The specs that I've seen on Cisco's web site list
the ethernet port on it as as 10/100 port...
Greg
In article <[EMAIL PROTECTED]>,
Brian Smith <[EMAIL PROTECTED]> wrote:
> Tried the same configuration today - the hub won't "fallback" to 10m.
> It didn't work at all with the Cisco 675. I'm going to add anothe nic
> to my liunx box this week to handle the router traffic. The hub will
be
> just for the three 100m cards in the pc's. I think netgear makes an
> inexpensive hub that will work with both 10 and 100 but I don't know
if
> you can mix-and-match cards that are plugged into it, I'm no where
close
> to a networking expert. So, no that my modem is no longer useful (isp
> allows dial-up OR ADSl), the basement area network can't connect to
the
> 'net
>
> Let me know how your setup works.
>
> Brian Smith
>
> steve davidson wrote:
> >
> > I have a 100 Mb network with three clients, all using Netgear 100Mb
net
> > cards connected to a 4-port 100Mb LinkSys ethernet hub. The hub
has uplink
> > capability. I am considering purchasing DSL from a local provider
- they
> > supply a DSL router, but it is only 10Mb capable. What are my
options for
> > adding this device to my network? I am looking for low-cost
options, as
> > this is my home net.
> >
> > Thanks!
> >
> > Steve Davidson
>
--
It's pronounced "sexy" not "scuzzy"!
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
From: Pascal Fleer <[EMAIL PROTECTED]>
Subject: Re: PAM authenticating to Novell NDS or NT Domain?
Date: Sat, 31 Jul 1999 07:17:58 +0200
Reply-To: [EMAIL PROTECTED]
Just realising that SuSE is delivered with on old ncpfs package. Thanks for the tip.
I will test as soon as possible. My Linux users will be happy if novell connection
works again.
Pascal
PS : For those looking for the package : ftp://ftp.gwdg.de/pub/linux/misc/ncpfs
TurkBear wrote:
> My RedHat 5.2 with the 2.3.10 kernel ( yes, I know---but I'm not risk-averse )
> Handles NCP connections with full NDS authentication...It seems that it has done
> so since 2.3 - Have IPX configured in the kernel and, using LinuxConfig let it
> autodetect everything...
>
------------------------------
From: [EMAIL PROTECTED] (Michael Hayes)
Subject: Re: ppp - ping OK but no telnet/ftp/netscape - answer.
Date: 31 Jul 1999 05:32:31 GMT
Mark Bestel ([EMAIL PROTECTED]) wrote:
: Do you have an ethernet card and modem? You might want to make sure that there
: aren't two default routes that are conflicting.
:
: Type /sbin/route as root from a terminal and see if there are two default
: entries. If there are, one will be for ppp0 and one will be for eth0.
:
: Type "/sbin/route dev default dev eth0"
:
: This will remove the default route through your ethernet card and may help with
: your problem.
Thanks for all the help everyone, unfortunately that's not it. Someone (thankfully)
pointed
out to me that ping sends ICMP packets, so I did some investigation and my connection
only
chokes on TCP (didn't check UDP). I believe that the problem lies in the Van Jacobson
compression. I'll be sure to post a followup if something develops along this line.
Thanks
for all the suggestions, though, I tried them to no avail. If I get it working I'll
post
something that might hopefully help some people out.
mike
------------------------------
From: [EMAIL PROTECTED] (tykq)
Subject: sendmail configuration ?
Date: Sat, 31 Jul 1999 03:26:36 GMT
Hi
I ve been experimenting with linux and i ran into a problem !
I have RH 6.0 on one PC and Win3.1 on an old 386 both running tcp/ip.
The ip of linux box is 192.168.1.1 and the 386 is 192.168.1.4 and pop3
is runing on Linux box.
I'm trying to read mail from the 386 using MS internet mail setting
the linuxbox as the pop3 server.
The maillog reads:
ipop3 : port 110 service from 192.168.1.4 login user=myname
host=[192.168.1.4] nmsgs=1/1
ipop3: connection refused while reading user=myname,host=[192.168.1.4]
If i try to send mail to any user, i have
relay=[192.168.1.4] reject ,Relaying denied
Ive tried to use Linuxconf to add a Relay for, by IP to 192.168.1.4
but no result.
Should i add any thing to sendmail.cf . There is a line in this file
which will permit relaying /etc/mail/relay-domains. I dont have this
file
Help
[EMAIL PROTECTED]
------------------------------
From: stanislav shalunov <[EMAIL PROTECTED]>
Crossposted-To:
comp.infosystems.www.servers.unix,microsoft.public.proxy,alt.unix.wizards
Subject: Re: Reverse Proxy & Load Balance + Redundancy
Date: 31 Jul 1999 01:58:48 -0400
Reverse Proxy would probably not help much in your case. I've
designed a no-single-point-of-failure scheme for my company (it should
survive a nuclear attack on any single facility :-) but cannot give
out the details. Since you know it's possible, you might just think
and reproduce it.
Or you might just use a high availability load balancer. Cisco Local
Director, or there are solutions both commercial and free on *BSD.
But wait; why do you need to load-balance Apache for static content?
You can handle like 200 hits per second on very modest hardware
(PII+64MB RAM+some striped disks) on a single machine running FreeBSD
(crank up maxusers or you'll start to be getting TCB-number bound
because of TIME_WAIT state!). That's several millions hits per day,
depending on how long and intensive is your peak period. Completely
free (except PC hardware for which you'll have to pay some $2000 with
all the good SCSI gear you may want).
------------------------------
From: Bill Steiner <[EMAIL PROTECTED]>
Subject: Re: Help: ipchains and FTP problem (ip_masq_ftp IS installed)
Date: Fri, 30 Jul 1999 21:46:27 -0700
Have a look at the ip-masquerade HOWTO. It doew a great job of explaining how
to set things up, including allowing ftp pass through. It also gives guidance
for ipchains.
Good Luck
Sebastian Bargmann wrote:
> Hi,
>
> I'm running RedHat 6 on a small network. I have configured the linux box to
> run as "internet gateway" for a Windows98 system.
>
> When I try to connect to my newly acquired domain through FTP from Win98, I
> can connect just fine, but I can't get the directory listing (it says it
> can't connect to my PPP client-side ip and port). I have tried forcing the
> FTP app to passive mode with no success.
>
> If I connect directly from the linux, then there's no problem.
>
> All other FTP sites I have tried works just fine. Both anonymous login and
> my personal account at my ISP.
>
> Some relevant configurations:
>
> Linux IP: 192.168.0.20
> Win98 IP: 192.168.0.21
>
> rc.local:
> modprobe ip_masq_ftp.o
> ipchains -P forward DENY
> ipchains -A forward -s 192.168.0.0/24 -j MASQ
> echo 1 > /proc/sys/net/ipv4/ip_dynaddr
>
> I have read several postings and documents mentioning something about
> accepting inbound packets. I have tried numerous combinations of ipchains
> .-A input etc., but I can't get it to work. Please note that I'm far from
> being an expert in this area, so any suggestions are more than welcome!
>
> Thanks!
> --
> Best regards
> Sebastian Bargmann
------------------------------
From: Chris Trudeau <[EMAIL PROTECTED]>
Subject: Re: OOPS! Security Question
Date: Sat, 31 Jul 1999 05:20:29 GMT
Chris Trudeau Wrote:
Also the RPC and xauth settings are going to be important. Run IP
chains and only allow the hosts you want to have access...
You DEFINITELZY want to format-install as whoever it is could have
created themselves a "backdoor" which they may or may not have covered
up well...Check history files to see where you were comprmised from.
Howard Mann <[EMAIL PROTECTED]> wrote:
>
> Nathan T. Lager wrote:
> > Ok, for the past 2-3 months ive been running a linux box connected
to a
> > 500k cable modem. its a lot of fun for me concidering i am a
> > PC-Networking student at the moment. Like i said it was up for
about 3
> > months and suddenly someone decided to crack into my system (i
assume).
> > All i know is that suddenly i cant log into my own system, even as
root
> > from the console! from what i can tell somehow my passwd file got
> > deleted. i intend to fix the system and get it back online but for
now
> > (before i get it back up and vulnerable) id like to find out
anything i
> > can about how to make it more secure. all i want it to do (for now)
is
> > serve Telnet, FTP, HTTP, POP3, and SMTP. I run RedHat 5.2 with a
Cable
> > connection to the 'net and a 5 computer network also connected to
this
> > system. I would also like to "Eventually" set the linux box as a
router
> > so that i can route packets from my LAN over the internet but "Hide"
the
> > LAN behind the Linux Box. Thanx in advance for any help.
>
> A few URL's :
>
> Basic security: http://www.xmission.com/~howardm/security.html
>
> Detailed security : https://www.seifried.org/lasg/
>
> LAN setup/security : http://rlz.ne.mediaone.net/linux
>
> IP Masquerading : http://metalab.unc.edu/LDP/HOWTO/mini/IP-
> Masquerade.html
>
> This should get you going :-)
>
> Howard Mann.
>
> ------------------ Posted via SearchLinux ------------------
> http://www.searchlinux.com
>
Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.
------------------------------
** FOR YOUR REFERENCE **
The service address, to which questions about the list itself and requests
to be added to or deleted from it should be directed, is:
Internet: [EMAIL PROTECTED]
You can send mail to the entire list (and comp.os.linux.networking) via:
Internet: [EMAIL PROTECTED]
Linux may be obtained via one of these FTP sites:
ftp.funet.fi pub/Linux
tsx-11.mit.edu pub/linux
sunsite.unc.edu pub/Linux
End of Linux-Networking Digest
******************************
