Linux-Networking Digest #497, Volume #11         Fri, 11 Jun 99 17:14:40 EDT

Contents:
  Linux, NIS and SGI/IRIX ([EMAIL PROTECTED])
  Re: sendmail - virtual host (Raymonds Doetjes)
  Intermittent DNS functionality ("Russell S. DiPesa")
  Beginner's question (Rajendra Chiplunkar)
  samba and smbclient problem ("Carl D. Blake")
  Re: 3C574 PCMCIA NIC connectivity ("Ulf Leichsenring")
  Re: Netscape / proxy q ("Stefo D. Stojanovski")
  Re: SAMBA & Win95 Logon authentication (Raymonds Doetjes)
  Re: Help: Opportunistic Locks (Raymonds Doetjes)
  Re: DNS caching (Raymonds Doetjes)
  Re: diald dying! ("Ian")
  MS SQL ODBC works for Sybase (Raymonds Doetjes)
  Linux-router question - connecting pc to T1 (Tim Kelley)

----------------------------------------------------------------------------

From: [EMAIL PROTECTED]
Subject: Linux, NIS and SGI/IRIX
Date: Fri, 11 Jun 1999 18:25:21 GMT

I need help.

I'm running NIS in a RedHat Linux, Solaris and SGI/Irix environment.
The NIS master and slave servers are Linux.  Linux to linux and linux
to Solaris work flawlessly.  The problem is with IRIX.  I have 2 O2's
running IRIX 6.5 and an Origin running IRIX 6.4.  I have both versions
of IRIX binding to the NIS servers.  YPWHICH and YPMATCH/YPCAT all
work just fine.

Here's my problem.  I can not login with a user account that was created
on the Linux box and can only be authenticated thru NIS.  When I try to
login is get -  'UX: login: ERROR: login incorrect' - as if the account
was invalid.  Now if I remove a user account from the local passwd file
so that NIS has to be consulted it works fine.

I am using shadow accounts on the SGI's but the Sun's are the same way
and work just fine.  I have added the '+' to the /etc/passwd files and
as you can see above they all appear to be checking NIS for
authentication.

Any hints, tips, etc., are extremely appreciated.

Thanks

BTW.  Since there is so little good info on NIS for Linux, at least,
RedHat, I plan on submitting my documentation for posting on the
various sites - when I get everything working.  Anyone that provides
help that contributes to solving this will be included in the credits.




Sent via Deja.com http://www.deja.com/
Share what you know. Learn what you don't.

------------------------------

From: Raymonds Doetjes <[EMAIL PROTECTED]>
Subject: Re: sendmail - virtual host
Date: Fri, 11 Jun 1999 21:56:29 +0200

Visit www.sendmail.org there is a complete discription on that.

Raymond

Patrick wrote:

> can u tell me the steps to provide virtual host mail services
> by just using one pc?
> thank you!
>
> --


------------------------------

From: "Russell S. DiPesa" <[EMAIL PROTECTED]>
Subject: Intermittent DNS functionality
Date: Fri, 11 Jun 1999 15:13:41 -0400

To All,
    For the last couple of months, I have been having intermittent
connection problems with my RH5.2 box.  I have an ISDN connection through
one of the major Bells, and just about once every couple of days, I am
unable to ping the IP of the machine.  After a while, I will be able to ping
the IP of the machine, but I am unable to ping just using the machine name.
At this point, I do not have FTP, WWW, SMTP, or TELNET access.  Then, after
a while longer, all normal modes of access to the machine return for about a
day or a day and a half.  Then, the vicious cycle starts all over.
    Does this sound familiar?  Any ideas, anyone?

Regards,
Russ



------------------------------

From: Rajendra Chiplunkar <[EMAIL PROTECTED]>
Crossposted-To: comp.os.linux
Subject: Beginner's question
Date: Fri, 11 Jun 1999 12:45:21 -0500
Reply-To: [EMAIL PROTECTED]

I just installed Caldera's OPENLinux 2.2 on a Pentium machine. I set up
my external Hayes compatible modem as being connected to ttys1 (COM2). I
did this logged in as root. I can dial out with KPPP. But as a user when
I use KPPP, it shows me the number etc. but says it cannot find the
modem. Do I have to setup the modem for every user as being dev\ttys1.
Thanks
RC


------------------------------

From: "Carl D. Blake" <[EMAIL PROTECTED]>
Crossposted-To: comp.protocols.smb
Subject: samba and smbclient problem
Date: Fri, 11 Jun 1999 20:08:47 GMT

I've got samba almost completely working on a RedHat V6.0 linux system.  I
just can't get smbclient to work.  My network has several Win95 systems and
one WinNT system.  I can browse the linux server from the Win95 systems no
problem (I can't from the WinNT system yet because password encryption is
set to off in samba - that's fine for now, I'll fix it later).  When I run
"smbclient -L imladris" I get the error
    session request to IMLADRIS failed
    session request to *SMBSERVER failed

When I run "smbclient -L imladris -d 5" I get

Client started (version 2.0.3).
resolve_name: Attempting lmhosts lookup for name imladris<0x20>
getlmhostsent: lmhost entry: 127.0.0.1 localhost
getlmhostsent: lmhost entry: imladris imladris
Connecting to 10.0.1.23 at port 139
Sent session request
size=1
smb_com=0x0
smb_rcls=0
smb_reh=0
smb_err=0
smb_flg=0
smb_flg2=0
smb_tid=0
smb_pid=0
smb_uid=0
smb_mid=0
smt_wct=0
smb_bcc=0
session request to IMLADRIS failed
resolve_name: Attempting lmhosts lookup for name imladris<0x20>
getlmhostsent: lmhost entry: 127.0.0.1 localhost
getlmhostsent: lmhost entry: imladris imladris
Connecting to 10.0.1.23 at port 139
Sent session request
size=1
smb_com=0x0
smb_rcls=0
smb_reh=0
smb_err=0
smb_flg=0
smb_flg2=0
smb_tid=0
smb_pid=0
smb_uid=0
smb_mid=0
smt_wct=0
smb_bcc=0
session request to *SMBSERVER failed

For this particular case I have the /etc/hosts and /etc/lmhosts setup to
recognize imladris.  I have another machine called rivendell which are not
entered into the hosts files and this is the message I get when I run
smbclient -L rivendell -d 5

Client started (version 2.0.3).
resolve_name: Attempting lmhosts lookup for name rivendell<0x20>
getlmhostsent: lmhost entry: 127.0.0.1 localhost
getlmhostsent: lmhost entry: imladris imladris
resolve_name: Attempting host lookup for name rivendell<0x20>
resolve_name: Attempting wins lookup for name rivendell<0x20>
resolve_name: WINS server resolution selected and no WINS server present.
resolve_name: Attempting broadcast lookup for name rivendell<0x20>
bind succeeded on port 0
nmb packet from 10.0.1.255(137) header: id=22789 opcode=Query(0) response=No
    header: flags: bcast=Yes rec_avail=No rec_des=Yes trunc=No auth=No
    header: rcode=0 qdcount=1 ancount=0 nscount=0 arcount=0
    question: q_name=RIVENDELL<20> q_type=32 q_class=1
Sending a packet of len 50 to (10.0.1.255) on port 137
nmb packet from 10.0.1.255(137) header: id=22789 opcode=Query(0) response=No
    header: flags: bcast=Yes rec_avail=No rec_des=Yes trunc=No auth=No
    header: rcode=0 qdcount=1 ancount=0 nscount=0 arcount=0
    question: q_name=RIVENDELL<20> q_type=32 q_class=1
Sending a packet of len 50 to (10.0.1.255) on port 137
nmb packet from 10.0.1.255(137) header: id=22789 opcode=Query(0) response=No
    header: flags: bcast=Yes rec_avail=No rec_des=Yes trunc=No auth=No
    header: rcode=0 qdcount=1 ancount=0 nscount=0 arcount=0
    question: q_name=RIVENDELL<20> q_type=32 q_class=1
Sending a packet of len 50 to (10.0.1.255) on port 137
Connection to rivendell failed

It looks like its not getting a response when it sends a broadcast message
on port 137.  What I find so interesting about this is that all of this
works when I connect to the WinNT system (named erebor).  I can run
smbclient -L erebor and connect to the WinNT system with no problem - even
though it is not entered in the hosts files.  I looked at the debug messages
when connecting to erebor and erebor responds to the broadcast message on
port 137.  What is going on here?  How do I get the Win95 systems to respond
to smbclient?





------------------------------

From: "Ulf Leichsenring" <[EMAIL PROTECTED]>
Subject: Re: 3C574 PCMCIA NIC connectivity
Date: Thu, 10 Jun 1999 23:56:04 +0200

Hi Dave

I have a Toshiba Tecra 8000 and a 3c574tx pc-card too and I just had the
same problem. But today I downloaded and installed the actual pcmcia 3.0.12
packet and the card *works* now! Go to ftp://hyper.stanford.edu/pub/pcmcia
and get it yourself.
Please leave me a note if it helped you too.

Ulf


--
Ulf Leichsenring
[EMAIL PROTECTED]



------------------------------

From: "Stefo D. Stojanovski" <[EMAIL PROTECTED]>
Crossposted-To: alt.os.linux,comp.os.linux.setup,comp.os.linux.questions
Subject: Re: Netscape / proxy q
Date: Fri, 11 Jun 1999 16:02:21 -0400

This is a multi-part message in MIME format.
==============7FCD08E67AB1ACBFD8F280AD
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

David Eno wrote:

> I'm trying to run Netscape on a Red Hat 5.2 box that's connected to our NT 4
> network running MS Proxy v 1.
>
> I set the obvious proxy settings in Netscape.  When I try to go to a site,
> the messages '(site) contacted. Waithing for reply' and 'connecting to
> (proxy IP)' flash back and forth repeatedly at the bottom the the screen.
>
> Anyone have any idea what the problem might be?
>
> TIA
> --
> Dave E.

Dave,
    From what I know about MS Proxy, it gives you a URL that points to a script
that automatically configures your proxy settings.  You can input this address
that points in your proxy settings in Netscape.  Hope it sheds some light....

steve

==============7FCD08E67AB1ACBFD8F280AD
Content-Type: text/x-vcard; charset=us-ascii;
 name="steve.vcf"
Content-Transfer-Encoding: 7bit
Content-Description: Card for Stefo D. Stojanovski
Content-Disposition: attachment;
 filename="steve.vcf"

begin:vcard 
n:Stojanovski;Stefo D.
tel;fax:315-434-9782
tel;work:315-434-9787
x-mozilla-html:FALSE
url:www.cadimensions.com
org:CADimensions, Inc.;Technical Services
adr:;;6310 Fly Road;East Syracuse;New York;13057;United States
version:2.1
email;internet:[EMAIL PROTECTED]
title:Technical Applications Representative
fn:Stefo D. Stojanovski
end:vcard

==============7FCD08E67AB1ACBFD8F280AD==


------------------------------

From: Raymonds Doetjes <[EMAIL PROTECTED]>
Subject: Re: SAMBA & Win95 Logon authentication
Date: Fri, 11 Jun 1999 22:13:13 +0200

See mt website for a full featured SambA logon server with user profiles for win
9x machines
http://node102bd.a2000.nl/rdoetjes
The SMB section is the only thing that works on this side yet (lucky for you ;0)

Raymond

James Dingwall wrote:

> Hi,
> I'm trying to set samba up to authenticate win 95 client logons.  I'm
> having troubles since I don't know on which side things are set up
> properly.  At the moment I can get my 95 clients to bring up a message
> from the logon which says "No Domain Controller was available to
> authenticate the password, some net services may be unavailable" (or
> similar)  I'm sure that I've got samba set up properly and I've tinkered
> with the 95 network settings (wins enabled, NT domain logon enabled).
> Just in case it helps I've included the global section of my smb.conf.
> (Other possibility is I need new samba & PDC but I thought that was for NT
> clients)
>
> Many Thanks,
>
> James
>
> (CCed replies via email appreciated if poss:)
>
> # Samba config file created using SWAT
> # from spook.jkd.net.uk (192.168.1.1)
> # Date: 1999/05/26 23:38:41
>
> # Global parameters
>         workgroup = JAMES
>         netbios name = SPOOK
>         server string = SERVER MACHINE
>
>         hosts allow = 192.168.1.0/255.255.255.0 127.0.0.1
>         interfaces = 192.168.1.1/255.255.255.0
>         remote announce = 192.168.1.0
>         keepalive = 30
>
>         preferred master = Yes
>         domain master = Yes
>         wins support = Yes
>         os level = 65
>         lm announce = Yes
>         time server = Yes
>
>         log file = /var/log/log.smb.%M
>         max log size = 50
>         status = Yes
>
>         security = user
>         domain logons = Yes
>         logon script = %U.bat
>         logon path = \\Spook\profiles\%m\%U
>         smb passwd file = /usr/lib/samba/private/smbpasswd
>         guest ok = Yes
>         guest account = nobody
>         invalid users = root,ftp,www
> #       unix realname = Yes
>
>         hide dot files = No
>         follow symlinks = No
>         preserve case = Yes
>         short preserve case = Yes
>         case sensitive = No
>
>         printing = lprng
>         print command = /usr/local/bin/lpr -r -b -m %U -P%p %s
>         lpq command = /usr/local/bin/lpq -P%p
>         lprm command = /usr/local/bin/lprm -P%p %j
>         printer driver location = /usr/lib/samba/printers
>         load printers = No
>         printer driver file = /usr/lib/samba/printers/printers.def
>
> --
> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> |                              James Dingwall                               |
> +                           [EMAIL PROTECTED]                            +
> |                            [EMAIL PROTECTED]                             |
> +                http://www.fen.bris.ac.uk/students/jd5477/                 +
> |             Maintainer: http://www.nottingham.ac.uk/~lqzweb/              |
> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


------------------------------

From: Raymonds Doetjes <[EMAIL PROTECTED]>
Crossposted-To: comp.protocols.smb
Subject: Re: Help: Opportunistic Locks
Date: Fri, 11 Jun 1999 22:25:59 +0200

Oplocks arent on add default, you need to oplocks = yes to a share.
But then anyway, the time stamp is checked every x seconds I meant it was
10 but I could be off. So your client will check every 10 sec if the
timestamp has been changed on the file that it has cached completly (or x
ammount of blocks) if it changes it will warn you or just reload the file.
Depending on the client.

Oplocks is pretty secure. When you don't want oplocks because you need 100%
data integrity. Set oplocks = no to your share.

But your hypothesis is right except your forgot the timestamp checking ;-)

Raymond

Bob Dusek wrote:

> Hello,
>
> I've been doing some reading on opportunistic locks and I think I
> understand them.  First, I will attempt to explain my understanding of
> them, then I'll give you my question:
>
> Opportunistic locks, from what I understand, allow client machines to
> "aggresively" cache files locally.  The documentation I've read says
> that sometimes a client even caches open/close file operations... in
> other words, it doesn't (necessarily) even tell the server it's
> closing a file.  It just closes it, since it's got the file locked,
> w/opportunistic locking.  For the purpose of continuing my explanation,
> I'll call this client "Client Lock".
>
> This local cacheing is permitted by the server, because whenever another
> client tries to access the file, the server knows who has the file
> locked (Client Lock) and it sends a message to Client Lock, saying
> "Hey!  Give me your updates, someone else needs that file!"
>
> It is in this manner that the server "breaks" the lock.
>
> Now, according to this explanation... if I were to open a file on a
> Microsoft SMB client and make changes to the file, saving it
> periodically, and if I were to open the same file on a Linux machine,
> which accesses the file via NFS, and edit it.... then the SMB client
> wouldn't necessarily see the changes that I made to the file via my NFS
> client.  In fact, from what I understand, since the server's lock
> information for the file wouldn't change ('cause NFS doesn't manipulate
> the opportunistic locks) the SMB client WOULD NOT see the changes that
> are made to the file via NFS.
>
> However, when I actually try to test this hypothesis, I get a
> surprise... the smb client sees every change that I make to a file via
> an NFS client.  What gives?  Could I have opportunistic locks turned
> off?  I've looked in our smb.conf file and there is not any "oplocks =
> off" or anything like that.  And, the documentation says that it is
> turned on by default.  Also -- the documentation says that "kernel
> oplocks" is not yet implemented for Linux (which would allow NFS to
> "break" opportunistic locks).
>
> Could someone lend me a hand?  We want all of our users to maintain only
> 1 directory structure, and we want to prevent data corruption (but we
> can't guarantee that our users will never be accessing their files via
> NFS and SMB).  If necessary, we may have to "smbmount" the file system
> from Linux... but, I'd rather hold off on that, if "kernel oplocks" is
> going to be implemented soon....
>
> Thanks,
>
> Bob
>
>
>
> --
>
> Saint Joseph's College --  [EMAIL PROTECTED] | http://www.saintjoe.edu
> -----------------------------------------------------------------------
> Always the dullness of the fool is the whetstone of the wits.
>                 -- William Shakespeare, "As You Like It"


------------------------------

From: Raymonds Doetjes <[EMAIL PROTECTED]>
Subject: Re: DNS caching
Date: Fri, 11 Jun 1999 22:21:14 +0200

Giva that A record an explicit TTL so that the DNS server will check the
zone files evry 10 seconds f.i.

You can just add a TTL like you add a TTL to your SOA record.

Raymond

Gutrot wrote:

> I was wondering if it's possible to have a nameserver not cache an IP
> address.  I have a dynamic IP from my ISP and am having a problem
> dynamically mapping a hostname from my friends domain to it.
>
> It works.. but all the other nameservers on the net are caching the IP
> address... and when I log on again with a different IP, nameservers are
> using the cached one instead.
>
> Is there a flag I can set in named to stop other servers from caching?
>
> Any ideas?
> Thanks,
> -Gut
>
> ------------------------------------------------------------------------
> Build a system even a fool can use, and only a fool will want to use it.
>
> Sent via Deja.com http://www.deja.com/
> Share what you know. Learn what you don't.


------------------------------

From: "Ian" <[EMAIL PROTECTED]>
Subject: Re: diald dying!
Date: Fri, 11 Jun 1999 09:37:52 +1200

Yeah.
I have the same problem on RedHat 5.1

Seems to die about once every few weeks.

To solve it I wrote a little cron job that runs every 5 mins, checks if it's
still there and if not re-starts it.

Ian

Christopher A. Gaul wrote in message <[EMAIL PROTECTED]>...
>My diald daemon keeps dying (disappearing anyway). I'm not sure why.
>Anyone have any idea why? Here is my set-up.
>
>SuSE 6.1
>Kernel 2.2.7 (SuSE) with kernmod update
>ppp version 2.3.5-36
>diald version 0.16-49
>
>I am using my Linux server as a SAMBA domain server for my home network,
>it serves logins, files, printer, NFS and internet access. I use
>masquerading and diald to provide transparent Internet access to the
>network clients. However, I often find the diald daemon often just
>quits. I have no idea why.
>
>Thanks in advance,
>Christopher A. Gaul
>CyberDyne Systems
>
>



------------------------------

From: Raymonds Doetjes <[EMAIL PROTECTED]>
Subject: MS SQL ODBC works for Sybase
Date: Fri, 11 Jun 1999 22:32:27 +0200

I read somewhere, that the MS SQL sql communication protocoll was
(borrowed/stolen/compatible) to the Sybase communcition protocoll.

Now alot of people want to use the Sybase RDBMS on Linux to let people
interact with it over ODBC. Since Sybase doesn't supply ODBC drivers
themselves, (InterSolve does that for them) and you have to buy them I
thought that there should be another way.

When I recalled that message I tried playing arround with SQLSRV32.EXE
odbc drivers on Windows at first it worked beacuse I had a very
old MS SQL ODBC driver in wich you could enter the portnumber of the
MS SQL server. Now in the newer versions of the ODBC driver you can't
define the prtnumber.

SO I checked with iptraf to wich port the MS SQL ODBC driver connectd
and that was port 1433 so I changed my Sybase setup to listen on 1433
and voila it is up and running

So:
- get a SQLSRV32.EXE (dedault in Win98 its PersonalWeb server package)
- Configure your Sybase to listen on to port 1433
and start running your ODBC based applications.

I tried the ODBC driver within Delphi, Access and VB everything I use to
do with MySQL runs on Sybase now. And Sybase can do alot more!!!!

Raymond



------------------------------

From: Tim Kelley <[EMAIL PROTECTED]>
Subject: Linux-router question - connecting pc to T1
Date: Thu, 10 Jun 1999 16:34:31 -0500

Currently  I have a linux box, acting as a firewall and running squid
for proxy sitting between my network and a Cisco 1600 which in turn is
connected to a T1.

I was curious as to who makes hardware which will enable me to connect
the linux firewall directly to the T1?

Would I see any performance gains?

------------------------------


** 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
******************************

Reply via email to