Re: [SLUG] FCNP - Continued - Apology

From: Bill (selinux@home.com)
Date: Mon Sep 03 2001 - 15:30:27 EDT


Please accept my apologies for the length of this. I should be able to trim 90% of
this away on the next go-round.

Re-reading my previous reply I realized some of it could easily
and reasonably be miscontrued as being snotty. That certainly
was not my intention. I am getting a lot of help here for which
I am deeply grateful ... and, unlike some of the 'help' one gets
online, things seem to be getting incrementally better. I
suspect that if "knowing his stuff" was a criminal offense, we
could get at least an indictment against Steve. Please note the
addition of explanatory notes that temper the tone of my
previous posting. My initial choice of wording was not well
considered. I hope that this addendum will smooth the waters. --
Bill

On Sunday 02 September 2001 22:38, you wrote:
> On Sunday 02 September 2001 16:00, you wrote: > First, was I
> that off in the find command? Are my waning > linux skills so
> impaired and diminished? I could have sworn > that "find .
> -name wins.dat -print" is the correct syntax.
>
> > Did I neglect to mention that the command should be run from
> > / > ?
>
> Yeah ... you skipped that part. :-)

Apparently you are human. But I ran it as root anyway ... or are
you telling me I need to have "/" as my pwd when running the
find command? In this case it was irrelevant ... someone on the
Samba development team threw sand in the gears. Keep reading.

I gotta warn you ... I am
> a dangerous computer user. I know just enough to get into
> trouble but not enough to reliably get back out of it.
>
This was intended as self-deprecating humor ... not actually a
warning. I very much appreciate your help to this point and
don't want this to drag on and on. At the current rate, it looks
like a couple more emails should have me running like a server
farm. :-) A TINY server farm ... but up & running!
> I've gone from being a clueless support nuisance to being a
> mis-clued support nightmare. It's a phase ... a transition ...
> and I'll get over it. But, for right now, I'm dangerous!
>
> Anyway, the location should be "/var/lock/samba/wins.dat" >
> for most distributions.
>
> I think I know what happened to the "find" command.
I think I could have puzzled this out with the ls command
earlier but I didn't know that I was looking for a log file. I
thought that wins.dat was some sort of autogenerated reference
file. Once you provided the above path it was a simple matter to
find the newly re-named file think I actually need.
>
> [root@a bill]# ls /var/lock/samba/ brlock.tdb locking.tdb
> ntdrivers.tdb smbd.pid browse.dat messages.tdb printing.tdb
> unexpected.tdb connections.tdb nmbd.pid share_info.tdb
>
Apparently SAMBA changed file names between releases. <rant>
I've noticed this stuff happening with other programs. </rant>

> Well ... it AINT "wins.dat" but is this the file you were
> referring to?

It's contents look like what I think I am looking for and it was
automatically generated ... another characteristic that wins.dat
was supposed to fulfill. If the contents of this file are what
you expected from wins.dat, then I think that part of the riddle
has been solved and the correctness of the preceding steps have
been confirmed.

>
> [root@a bill]# cat /var/lock/samba/browse.dat
> "@HOME" c0001000 "A" "@HOME"
> "A" 40049a03 "Samba Server 2.2.0" "@HOME"
> "CC289192-A" 40412003 "k7-800" "@HOME"
>
> looks funky ... this is identifying the workgroup, the hosting
> Linux machine and my wife's Win98 machine. It has not
> identified the VMware WIn98 install on the hosting Linux
> machine which should be identified as "D" and was installed
> with both host and bridged networking installed.

This non-identification of the guest OS puzzles me. Although I
can use Start | find | computer to locate both of the Win98
machines and use files on the distant machine, Network
Neighborhood is a bust and Samba is recognizing only one of two
copies of Win98 on my network.

Also ... will I need to fire up a separate copy of SAMBA on the
other Linux machine ("B") or are there entries I need to make in
this copy of SAMBA that will allow it to serve for both Linux
boxes?

Although the
> host Linux box has now been up two weeks, the virtual Win98
> has been bounced a few times. (It is considered non-essential
> except as relates to two application programs)
>
> > > We are going to take a huge step back and return to
> > > basics. > Since I
>
> have no intention of taking a trip to the all but > abandoned
> town of Detroit, let's tear everything down and > build it
> back up :)
>
> I dunno ... the Jazz festival is this week! (<grumble> ...
> can't plow the snow off the streets ... or patrol them very
> well, either ... but can blow a zillion dollars on fireworks
> and "name" entertainment. The choice was "bread or circuses"
> ... and all the clowns voted. </grumble>)
>
> > > Do the Windows boxes get their IP Addresses from a DHCP
> > > Server > or are
>
> the addresses configured statically. If the address is >
> provided by a DHCP server, is the DHCP server setting the node
> > type or specifying a scope ID? To check to see if either of
> > those settings are set, run the command on the Windows >
> machines "IPconfig /all" (winipcfg if 9x boxes) and look for >
> "Node Type" or "scope ID". > > Make a note of what the Node
> Type is. It should say either > "Broadcast" (0x1),
> "point-to-point" (0x2), "Hybrid" (0x8), or > "Mixed" (0x4).
> What I expect to see is "Hybrid" (0x8) > because we set up the
> linux box as a WINS server to cut down > on network traffic.
>
> It is so. Hybrid is showing on box Win98 boxes (real and
> virtual). The box for NetBIOS Scope ID is blank in both
> instances, also.
>
> > > If you see a scope ID, this is the source of the problem.
> >
> > However, I do not expect this so let's move on :)
> >
> > > Now let's test your smb.conf file. Run the command
> > > "testparm"
> >
> > and pipe it to more or a file. Correct any errors.
>
> Load smb config files from /etc/smb.conf Processing section
> "[homes]" Global parameter dns proxy found in service section!
> <=== changed9/2/01|21.20
> Processing section "[printers]"
> Processing section
> "[/var/www]" Loaded services file OK. Press enter to see a
> dump of your service definitions
>
> ____________
> [root@a bill]# vi /etc/smb.conf
> [root@a bill]#
>
> dns proxy WAS "yes", NOW "no"
>
> ran smbd restart
> ran nmbd restart
> (not certain which controls what ... so I ran both)

And the logs bear this out with the comment that they were
already running. But .. ya know what? I think I am going to
issue "smbd stop / smbd start", "nmbd stop / nmbd start" just to
put them completely through the full cycle. NOW is not the time
to miss reading some init file or leave something in ram that
should no longer be running.
> ____________
>
> > If you reached this point, we assume that
> >
> > > A) Your TCP/IP Settings on both your Windows and Linux Box
> > > are good. Same subnet, etc. Doublecheck for
>
> typos.
>
> > > B) You do not have a Scope ID screwing things up.
> > > C) The Linux box is the WINS server for this subnet and
> > > the Windows boxes are configured
>
> to use it. (i.e. the windows boxes are set up for Hybrid)
>
> > > D) Your smb.conf file is clean
> > > E) smbd and nmbd are running without errors.
>
> Review your log files.
>
> Both log.smbd and log.nmbd complained they were already
> running; user "nobody" is bugging smbd(IIRC) and both of them
> are looking for a unicode map they can not find.
>
> > > Now let's check to see if the Linux box can see
>
> the Windows > boxes. Use the command nmblookup. Refer to the
> man pages for > the proper syntax. What you want to test is to
> see if the linux box can resolve the NetBIOS name of the
> Windows boxes. Let's focus on this for now. If this comes
> back successfully then we will work from there.
>
> [root@a bill]# nmblookup -R CC289192-A querying CC289192-A on
> 192.168.1.255 192.168.1.100 CC289192-A<00> <== found wife's
> machine
>
> [root@a bill]# nmblookup -R d.genesis.com querying
> d.genesis.com on 192.168.1.255 192.168.1.104 d.genesis.com<00>
> <== found virtual machine
>
> [root@a bill]# nmblookup -R a.genesis.com querying
> a.genesis.com on 192.168.1.255 querying a.genesis.com on
> 192.168.212.255 name_query failed to find name a.genesis.com
> <== Samba host
>
> [root@a bill]# nmblookup -R b.genesis.com querying
> b.genesis.com on 192.168.1.255 querying b.genesis.com on
> 192.168.212.255 name_query failed to find name b.genesis.com
> <== 2nd Linux box [root@a bill]#
>
> Good. It found both Win boxes and did not find either Linux
> box. I think this was the expected behavior ... no? > > > >
> ------ Do not read this if you are afraid of the registry >
> ------- This can be set from the registry by editing the key >
> \Registry\Machine\System\CurrentControlSet\Services\NetBT\Para
>
> >meters NodeType = REG_DWORD 0x00000008
>
> I'm not afraid o' no stinkin' registry :-) but the above tree
> (Win98) appears to actually be:
> \Registry\HKEY_LOCAL_MACHINE\System\CurrentControlSet\Services
>\InetInfo\Parameters
>
> There is, at present, no "NodeType". This leads me to ask
> whether you wish me to modify a (non-existent) NodeType if it
> exists or to create a new one if there is not one presently.
> Specifically, since there is no NetBT branching, is InetInfo a
> proper alternative location?
>
> There was a time when I had a handle on all things MSFT ...
> even taught it ... but I long ago just simply lost interest in
> it and have NOT done the necessary homework to be mucking
> around in the registry. I am competent to follow clear
> directions but I am also competent to recognize when my
> understanding of the directions is unclear.
>
> Oh ... and I very much want to thank you deeply for all the
> effort you are putting into this. The pizza is in the mail.
>
> Bill
>
>
> -- icq # 126373831

-- 
No .sig today. Sorry.



This archive was generated by hypermail 2.1.3 : Fri Aug 01 2014 - 18:41:39 EDT