From: "ROW::EVERHART"@star.zko.dec.com
Date: Thu, 2 Nov 1995 15:37:22 -0500
Message-Id: <95110215372198@star.zko.dec.com>
To: everhart@GCE.MV.COM
Subject: Decuserver jnl 10/95
X-Vms-To: STAR::SMTP%"everhart@gce.mv.com"
Article 7508 of comp.org.decus:
Path: nntpd.lkg.dec.com!crl.dec.com!crl.dec.com!bloom-beacon.mit.edu!spool.mu.edu!howland.reston.ans.net!newsfeed.internetmci.com!in1.uu.net!eisner!mcmahon_b
From: mcmahon_b@eisner.decus.org (Brian McMahon, Info-VAX Refugee)
Newsgroups: comp.org.decus,vmsnet.decus.journal
Subject: DECUServe Journal October 1995
Message-ID: <1995Oct30.232646.11714@eisner>
Date: 30 Oct 95 23:26:45 -0400
Organization: DECUServe
Approved: mcmahon_b@eisner.decus.org
Lines: 1331
Xref: nntpd.lkg.dec.com comp.org.decus:7508 vmsnet.decus.journal:12
The DECUServe Journal
---------------------
October, 1995
From the Editors' Keyboard . . . . . . . . . . . . . 2
What's Inside, an Antidote to Journal Bloat
Folders and PC Notify . . . . . . . . . . . . . . . 3
About Empty Folders and Multi-user PCs
TeamLinks Searching . . . . . . . . . . . . . . . . 4
How to Implement Searching for TeamLinks Clients
Host Routing in DECnet Phase V . . . . . . . . . . . 6
The Saga Continues
RAID Controller Woes . . . . . . . . . . . . . . . . 7
Trials and Travails of SWXCR Configuration
Fibonacci Series . . . . . . . . . . . . . . . . . 15
A Simple Explanation
Internet Firewalls . . . . . . . . . . . . . . . . 16
What To Do, Where To Learn
Domain Name Service and Windows NT . . . . . . . . 20
DNS, Patches, and Docs
WHOIS for Windows . . . . . . . . . . . . . . . . 22
Sources for Winsock Utilities
The DECUServe Journal October, 1995 Page 2
From the Editors' Keyboard
From the Editors' Keyboard
---- --- -------- --------
Most recent issues of the DECUServe Journal have been a little bit
on the hefty side, as you may have noticed (and our news system
certainly did). This time, your devoted editors have adopted "short
and sweet" as their watchwords. The topics in this issue range from
firewalls to Fibonacci series, but none overly long-winded.
The obligatory Iowa weather report: It's snowing outside. Wet,
heavy, slushy stuff. We'd tell you exactly how delighted we are,
but we can't figure out how to do that without violating several of
the DECUServe Canons of Conduct. And now (speaking of not being
overly long-winded), on with the show.
* * * * *
The DECUServe Journal October, 1995 Page 3
Folders and PC Notify
Folders and PC Notify
------- --- -- ------
Abstract:
Two questions briefly stated and answered; the first, concerning the
possibility of empty folders existing in the file cabinet, the
second, about how to handle PC Notify when users aren't necessarily
tied to one PC.
Participants:
Laura Kilinc, Graham Pye, Russ Radkiewicz, Pat Scopelliti.
Conference: ALL-IN-1
Note 1151.0, 24-Aug-1995
Kilinc: Folders disappearing and pc notify entries appearing
------------------------------------------------------------
I have already called DEC support on the issues below and was told
that what is happening is the way things work and that nothing could
be done about it. I am hoping that someone might know some tricks.
- If an empty folder is left over night, it is deleted.
I would like the empty folder to stay. (Is it MR$TIDY
doing this deletion?)
- When using TeamLinks at the PC and IOS at the server,
if a user logs onto a PC it automatically adds an entry
into OA$PC_NOTIFY. I do not want this to be automatically
done. I want the user to have to enable this because we
often log into the PCs of other users.
Note 1151.1, 24-Aug-1995
Scopelliti: Folders are virtual
-------------------------------
Folders exist in the file cabinet simply as keys associated with
documents. If no document has a folder key value "TEST", then the
folder TEST does not exist. There is no "registry" of folder names -
at least not in the user's personal file cabinet.
This is how it's always worked.
The DECUServe Journal October, 1995 Page 4
Folders and PC Notify
Note 1151.2, 5-Sep-1995
Pye:
-----
RE .1 This is IOS's view of the world. Because MailWorks does allow
empty folders, then TeamLinks and the FCS pretend that an empty folder
still exists. When the FCS cache is emptied, the imaginary empty folder
goes away.
Re .0 You can configure TeamLinks so that it won't automatically add
the PC notify entry. A better way is to configure it so that it adds
*and* removes the entry when you disconnect, so that the next PC you
use will re-add the new node name.
Note 1151.3, 8-Sep-1995
Radkiewicz: More on PC notify . . .
-----------------------------------
Re .2, .0
To add a bit more detail to the PC notify setup, from the TeamLinks
Information manager screen select Options|New Mail Notification...
make sure that the last 2 check boxes are checked;
Enable on each connection to mail service, and
Disable when disconnecting from mail service.
When Enable new mail notification is checked, a record is created in
the OA$PC_NOTIFY database when you connect and the record is deleted
when you disconnect.
TeamLinks Searching
--------- ---------
Abstract:
Searching for text in the ALL-IN-1 File Cabinet is currently not
possible from a TeamLinks client. Following are some thoughts on
how to get around that restriction.
Participants:
Russ Radkiewicz, Ruth Saunders, Don Vickers.
Conference: ALL-IN-1
The DECUServe Journal October, 1995 Page 5
TeamLinks Searching
Note 1154.0, 30-Aug-1995
Saunders: Teamlinks and Searching
---------------------------------
We're currently on ALL-IN-1 v3.0 and were getting ready to install
Teamlinks for some of our PC clients. But, I just found out that in
Teamlinks, you can't search for a word or phrase! Choosing Find and
filling in a word in the text field doesn't work. According to
Teamlink's Help, "This field is not available for searches in ALL-IN-1
IOS drawers."
Do you know when this will be fixed? Do you know of any way that this
can be done? We have to be able to search or else Teamlinks won't be
installed. H-e-l-p!
Note 1154.1, 31-Aug-1995
Vickers: It's a feature - here's a workaround
---------------------------------------------
I fear that Digital sees this as a feature rather than as a bug. I
BELIEVE the rationale is to reduce the potential load on an ALL-IN-1
file cabinet server. I suggest that you may wish to call Digital via
your sales person and the CSC to express your views and the business
need for this feature.
In the mean time, there is a very good start of a solution in the
SUBSCRIP application also in the \AIDA sub-directory of the TeamLinks
SDK. This application allows you to submit a script on the ALL-IN-1
data access server (AIDA) passing parameters. You can create a script
that does the text search (BIND/SFILE mumble) and mails back the list
of documents. It's ugly but effective and you could market it as
freeing the client from having to wait at their 'terminal' while the
search is being performed. (';
Some pointers about the TeamLinks SDK are in topic 1143.
Note 1154.2, 6-Sep-1995
Saunders: Invalid general protection fault ??
---------------------------------------------
Thank you! Thank you for your help! I thought I might need some more
of your valuable help once I started to work on the script. But, I've
hit a stumbling block before that!
My son downloaded the tlsdk.exe file from the Internet for me. I then
had to upload to our Vax, then use Reflections to download to my PC.
(We don't have a Internet connection at work yet.) But, when I type
tlsdk at the Dos prompt on the PC, I get "application has violated
system integrity due to an invalid general protection fault." Does
this error ring any bells to you?
The DECUServe Journal October, 1995 Page 6
TeamLinks Searching
Note 1154.3, 6-Sep-1995
Vickers: Sounds like the EXE became non-binary
----------------------------------------------
The violation of system integrity, which is ALMOST a redundant term with
MS-DOS, is almost certainly because the TLSDK.EXE file was not moved in
a BINARY mode on one of the transfers. The file transfer over the
Internet should have been done with a MODE BINARY and your transfer to
your VAX system should have also been done in a BINARY mode which
depends on the protocol and tool you used for that.
Hopefully, it was the second step that was not done in BINARY so you
can repeat that one.
When you get the file on your PC system, don't forget the /D qualifier
so that all the sub-directories get created. I forget to add /D every
now and then and get everything on one HUGE directory. Most
embarrassing.
By the way, I just copied the Internet version of the SDK to:
USR_SCRATCH:[VICKERS]TLSDK.EXE
If you are connecting to DECUServe via Kermit on your VAX system, you
can use Kermit from here to transfer the file.
Unfortunately, the Internet version is lagging behind the version on
CompuServe by a few weeks. I'll bug the Digital folks one more time
about this.
Note 1154.4, 8-Sep-1995
Radkiewicz: Another way to do half the move
-------------------------------------------
Of course, if you have TeamLinks installed and working, and have the
SDK in a VAX directory, you could import it into an ALL-IN-1
file cabinet drawer and export it with TeamLinks. I do this all the
time especially when I start something at work and continue working on
home it from home with this method.
From ALL-IN-1 VT client do the following:
use the RVC option on the DT menu to get the file into a drawer.
From the TeamLinks client:
select the drawer and folder where the file is located, chose the
EXPORT option from the File menu or the export button from the
button bar, you will be asked where to put the file on your local
PC.
I have found this to be a quick and safe, even on a 9600 async modem
connection.
The DECUServe Journal October, 1995 Page 7
Host Routing in DECnet Phase V
Host Routing in DECnet Phase V
---- ------- -- ------ ----- -
Abstract:
The following update is of interest to those of you who are holding
on to DECnet Phase IV because you're not interested in buying a new
router.
Participant: Larry Kilgallen.
Conference: DEC_NETWORKING
Note 1191.0, 12-Sep-1995
Kilgallen: Host-Based Routing for DECnet Phase V
------------------------------------------------
> <<< Note 350.11 by EISNER::KILGALLEN "Larry Kilgallen" >>>
> -< There will be slight delay in the *real* Phase V >-
>
> Since DEC has not committed to a particular release date or version,
> I guess for most of us it just means that Phase V will come a little
> later than for others. Even more so than DECwindows, DECnet Phase
> V will certainly be a case where only those with the true pioneer
> spirit will run it the first chance they get.
>
> A bug so massive as not supporting the major improvement in Phase
> III (routing) on the acknowledged major operating system (VMS) is
> certainly enough to make most of us refrain from trying Phase V.
> If DEC wants to put that much egg on their face (I can imagine the
> headlines in Computerworld now), it is their privilege.
At the Boston LUG meeting today, VMS DECnet/OSI Engineering Manager
said that support for the present Phase IV DECnet implementation will
be dropped "only after host-based routing is available in DECnet/OSI
on both VAX and Alpha".
As is appropriate, DEC continues to say that host-based routing is
not appropriate for all environments. As is even more appropriate
it seems to me they have gotten a message from those who have not
moved to DECnet/OSI (even if it will route over TCP/IP).
RAID Controller Woes
---- ---------- ----
The DECUServe Journal October, 1995 Page 8
RAID Controller Woes
Abstract:
The following is a troubleshooting session in several iterations,
the goal being to set up a RAID-5 array using SWXCR controllers.
The discussion includes lots of interesting things to check when a
StorageWorks setup exhibits nondeterministic behavior.
Participants: Duncan Brown, Arnold De Larisch,
Keith Parris.
Conference: HARDWARE_HELP
Note 1925.0, 29-Aug-1995
Brown: SWXCR Raid controller flakiness?
---------------------------------------
Anyone out there using the 3-channel PCI-based Raid controller (aka
SWXCR)?
I've got a pair of them in a 2100 4/275, but in that case this disks
are just set up as all "JBOD" disks. It works fine.
We recently got a 2100 4/233 system with a pair of these same
controllers, but in this case I'm trying to set the drives up as
6-member RAID 5 sets. And it's giving me fits. After a lot of dinking
around with one controller, I actually got it to successfully create a
RAID 5 set of 6 drives....with one drive marked as "failed". Leaving
that same drive in place, I had it rebuild it, and I now have a
complete RAID 5 set.
On the second controller, I simply can't get it to work. Sometimes you
can see one drive light stop blinking partway through the
initialization process. I let that go one time, and of course that
drive showed failed (now I just reboot if I see that happening.) Other
times they all blink, but at the end it says "failed initialization"
and one or two drives are failed. If it's just one, I try to rebuild
it (either with or without formatting the drive first) and it fails.
I've swapped drives around to no avail. It actually seems to follow
the SLOT more than anything else, though across time the "bad slot"
drifts (what was once failing will work and some other drive will show
as failed.)
This is all from the SWXCRMGR program level, I haven't even booted the
OS on this thing yet! (It is destined to be Digital Unix.)
the drives are all RZ29B-VW drives (the wide-SCSI version), with wide
rackmount BA350 shelves front and back. I have six shelves, one per
SCSI bus. These are the PCI 3-channel Raid controllers, each with 32MB
of cache. I have write-back caching enabled, although until the
batteries get here (they were backordered) that's risky, but they
The DECUServe Journal October, 1995 Page 9
RAID Controller Woes
should arrive before we really have this in production anyway.
SWXCRMGR software is version 3.11 I believe (April 1995 date)
The cabling (done by DEC) all looks fine. The drives all format fine
if I just do that to them. It's almost acting like the SCSI bus isn't
terminated, or like maybe there's a problem with the SWXCRMGR software.
Does any of this ring any bells with anyone? Anyone else successfully
set up this type of configuration? I'm going to call DEC about it, but
I've been trying to gather more facts about exactly how it's failing,
since they're just going to ask that first thing anyway. Of course with
a format time of about an hour and a 6-drive-RAID-5 initialization time
of 4 or 5 hours, it's taking a while to try the various permutations to
isolate the problem!
Note 1925.1, 29-Aug-1995
De Larisch: Check the SHELF_OK jumper on the BA356
--------------------------------------------------
Duncan .... we've been using the PCI SWXCR for about 6 months now ....
Did you enable "fault management" on the controller? If you did, you need to
ensure that the BA356 is jumpered so the SHELF_OK signal is routed to the
controller (W1 is position 2 on J17). This is located on the Backplane of the
BA356 chassis.
Also, I've notice that the mechanical contact on the SWXCR's edge connector is
poor. I've resorted to using the 'bulkhead' BN31K-0E cable for channel 1 (only
connecting channel 0 on the card's edge connector).
Also ... Are you formatting the drives individually? You can select to format
multiple drives concurrently and save a bit of time.
Let me know how things go!
Note 1925.2, 29-Aug-1995
Brown: Update
-------------
Fault management is enabled. I actually opened up one shelf and looked
at that jumper and remember thinking it was correct, but I'm going to
go recheck it now with the "must be in position 2" mantra running
through my head... (Remember, this system is as-built by DEC
themselves, so everything *should* be right...right? OK, OK, stop
laughing...)
> Also, I've notice that the mechanical contact on the SWXCR's edge connector is
> poor. I've resorted to using the 'bulkhead' BN31K-0E cable for channel 1 (only
> connecting channel 0 on the card's edge connector).
In our first machine (the one with 16 JBODs), the drives were all in
internal shelves, so we only used the edge connectors (those are the
The DECUServe Journal October, 1995 Page 10
RAID Controller Woes
ones you are referring to as "edge" right, the ones on the top edge of
the card?) In this case, all shelves are external, so we have two
shelves plugged in to the bulkhead connectors on the back edge of the
card, and the third plugged into the extra bulkhead connector (which is
then plugged into the top edge connector on the card.) The connections
all look OK to me.
> Also ... Are you formatting the drives individually? You can select to format
> multiple drives concurrently and save a bit of time.
Yes, I figured that out pretty quick! Across the numerous attempts at
this, one common scenario is "initialize RAID set, one drive fails,
rebuild RAID set and let it reinitialize the failed drive" which simply
chews up enormous amounts of time, period.
A local StorageWorks guru had never heard of anything like this, so I
called it in to the CSC. The guy there agreed it sounded a lot like a
lack-of-SCSI-termination situation, but unless all my shelves
auto-termination modules are failing, that probably isn't it. He's
stumped too, though he admits I'm the very first person he's seen
running the PCI SWXCR 3-channel controller with all 16-bit drives. So
who knows, maybe I've bleeding-edged my way into some problem they
didn't catch in the design. Or maybe that SHELF_OK jumper is just
wrong, let me go check that now!)
The fate of the CSC call is that they're going to send the local FE out
to have a look. I'm not sure how he can help though, unless he likes
hanging around at 3-4 hours per shot (like I have been doing!) just to
watch each new strategy fail...
I'll keep you updated.
Note 1925.3, 29-Aug-1995
De Larisch: Need further checks of BA356's Personality and Terminator
---------------------------------------------------------------------
To clear up my terminology ... I used "edge" and "bulkhead" connector
to refer to the same item. It's the pair of connectors on the external
"bulkhead" of the card. I found, mechanically, unsound connections based
upon the relative position of the connector to the "bulkhead's edge" (that part
of the CPU chassis which supports the card's bulkhead. This was an issue on the
AlphaServer 1000 4/200 pedestal.
> A local StorageWorks guru had never heard of anything like this, so I
> called it in to the CSC. The guy there agreed it sounded a lot like a
> lack-of-SCSI-termination situation, but unless all my shelves
> auto-termination modules are failing, that probably isn't it. He's
> stumped too, though he admits I'm the very first person he's seen
> running the PCI SWXCR 3-channel controller with all 16-bit drives. So
> who knows, maybe I've bleeding-edged my way into some problem they
> didn't catch in the design. Or maybe that SHELF_OK jumper is just
> wrong, let me go check that now!)
The DECUServe Journal October, 1995 Page 11
RAID Controller Woes
You probably want to also check the "personality module" on the BA356 just for
grins. All switches should be ON. The cable should be plugged into JA1. Nothing
should be plugged into JB1 (Auto terminates). The BA35X-MC jumper board should
be installed in the backplane in the connector behind slot 6.
That's all I can think of for this moment!
Note 1925.4, 29-Aug-1995
Brown: Fun with mechanical engineering
--------------------------------------
> To clear up my terminology ... I used "edge" and "bulkhead" connector
> to refer to the same item. It's the pair of connectors on the external
> "bulkhead" of the card. I found, mechanically, unsound connections
> based upon the relative position of the connector to the "bulkhead's
> edge" (that part of the CPU chassis which supports the card's
> bulkhead. This was an issue on the AlphaServer 1000 4/200 pedestal.
I was just tearing it all apart to look at it and I see what you mean.
It's trivial to get the card in such that there isn't ANY room between
the metal housing on the connector and the hole in the back of the
chassis. (This is on the RM version of the 2100, but the standard
version has a similar issue.) That makes it difficult or impossible to
get the connector on right.
I also had the problem that PCI slot 0, on the EISA slot side, has this
RF shielding stuff that is supposed to be pinched between the back
plate on the card, and the chassis back plate. Well, my controller
(the one that's failing) had the wavy metal shielding stuff riding up
on the back surface of the card! That's easy to do, especially with
the whole backplane mounted sideways like it is: put the card in too low
and then bring it up into position, and bingo the metal's scraping the
back edge of the card. It doesn't *look* like it was actually
contacting anything other than blank card, but just to be sure I made
very sure to put it in just so, so that the shield was pinched properly
on the other side of the back plate.
This was the same time I was discovering that PCI slot zero is where
"SWXCR #2" resides, and PCI slot 1 is where "SWXCR #1" resides...silly
me for thinking they'd be numbered in the same order...
SO now it's off trying to initialize a 6-drive RAID 5 set again...
> You probably want to also check the "personality module" on the BA356
> just for grins. All switches should be ON. The cable should be plugged
> into JA1. Nothing should be plugged into JB1 (Auto terminates). The
> BA35X-MC jumper board should be installed in the backplane in the
> connector behind slot 6.
The personality module is the first thing I checked, since that's where
the termination happens. The switches are all to the drive side (which
I thought was all OFF, but I'll have to look again) and the cable is on
the front connector. And at least on the one shelf I took apart, the
The DECUServe Journal October, 1995 Page 12
RAID Controller Woes
16-bit jumper board is in the proper place, and the SHELF_OK jumper
is...OK!
Details as they become available...
Note 1925.5, 30-Aug-1995
Brown: It works, but does it?
-----------------------------
Of course it failed to properly initialize the RAID set again. it went
to 100% and showed success, but when you view the config, one drive
(a different one of course!) shows "FLD" status.
Your comments about the SHELF_OK jumper and error management made me
think of something though. I had been powering down the machine,
swapping drives, rebooting, and then manually running the rebuild
procedure so I could keep tabs on its process (it reports back to the
screen when you do that, with percent complete and any error messages.)
That had worked fine when I went through all this on the first
controller.
So this time I let it work its standard magic: I left everything
running, and pulled the "failed" drive. I inserted another drive into
the slot. About 30 seconds later, it automatically started rebuilding
it. I left it running, and sure enough when I came in this morning, I
had a fully initialized, no-failures RAID set. Go figure. Just to
make a point, I formatted that "failed" drive and sure enough it had no
errors. Of course, to *really* drive home the point, what I need to do
is let it configure a RAID set with one drive failed...pull that drive
out...and then put it right back in again! I am convinced it would
rebuild it without error.
SO this whole thing may be some obscure problem with the standalone
routine software or something. I must admit I'll remain skeptical of
these drives' integrity until I get the OS booted and shovel some data
over to them!
Note 1925.6, 30-Aug-1995
De Larisch: Check Spin-up Delay parameter & Try to force failed drive Optima
----------------------------------------------------------------------------
Have you tried the "utilities" menu on the standalone diskette? There is an
option to "Make Optimal" or some such option. Given that you have just
"formatted" all the drives ... there is NOTHING of interest on them. Simply use
this option (and of course ignore the dire warnings about using it ...) since
you have NOTHING to loose. Then for grins, I'd run the parity check software
probably on-line (i.e. from the O/S level) before doing anything serious with
it.
> Your comments about the SHELF_OK jumper and error management made me
> think of something though. I had been powering down the machine,
> swapping drives, rebooting, and then manually running the rebuild
The DECUServe Journal October, 1995 Page 13
RAID Controller Woes
> procedure so I could keep tabs on its process (it reports back to the
> screen when you do that, with percent complete and any error messages.)
> That had worked fine when I went through all this on the first
> controller.
Interesting ... I'll bet that if you remove the "failed" drive and re-insert it
that you can make it "optimal" and all will be well. There may be a "spin up"
issue that you are seeing ...how many drives are you allowing to be spun-up at
a time??? You may wish to extend the delay to say 20 seconds (from the default
of 12s). This will give the drive more time to "come ready". Is the "failed"
drive always a "high unit number" drive?
> SO this whole thing may be some obscure problem with the standalone
> routine software or something. I must admit I'll remain skeptical of
> these drives' integrity until I get the OS booted and shovel some data
> over to them!
I suspect that the controller is being a bit more "reserved" than it should.
If this continues to persist ... I'd be happy to forward you name and number
onto the engineering group responsible for the SWXCR series.
Note 1925.7, 30-Aug-1995
Parris: Faster init's using write-back caching temporarily
----------------------------------------------------------
One trick I've heard of to speed those init's is to turn write-back caching
ON while you're doing the init (then turn if back OFF before you actually
use the array, unless you're using the PCI-bus version (the KZPSC) and it has
the optional battery backup for the cache).
Note 1925.8, 30-Aug-1995
Brown: Wrong connector (I know, you already suggested it!)
----------------------------------------------------------
We're actually using write back caching anyway. (The batteries
themselves have been back-ordered, but I hope they arrive before we
actually put this into production and depend on the data...)
The problem has almost certainly been solved...
I'm not sure where I saw a picture of the JA1/JB1 connector and thought
these were done right but they were *not*. The cables were all on JB1
(the one that DISABLES the termination!) Yes, this system was built by
the Digital rackmount systems folks. The cables were all beautifully
dressed and tied to just the right length, I had to cut it all apart
and retie it to get the connectors in on the top (JA1) connector.
One might ask why DEC chose to make the extremely-hard-to-get-at
connector the main one!! I don't have any 8-bit shelves, did those
cable up to a connector that faced towards the front of the shelf? That
would sure explain why the assemblers put it together the way they did
(nobody told them the 16 bit shelves used the hidden connector!)
The DECUServe Journal October, 1995 Page 14
RAID Controller Woes
It was a guy at the CSC who's been working with me on this that got
another similar call, but with someone working with a pedestal shelf.
They figured out that's what the problem was, and he called me to
suggest it might be the problem here.
I have one set of shelves recabled, and it's initializing the RAID set
now. I fully expect it to work, and I'll keep you posted on the
results.
Anyone who's got an order in for a CAB system with 16 bit shelves
(regardless of controller type) better check those cable connections!
If the connector is buried inside the shelf, you're OK. If you can see
it on the front, you've got problems. (Unless of course you really
intended to use the front connector to chain two shelves together...)
Duncan, feeling more terminated now...
Note 1925.9, 30-Aug-1995
De Larisch: Ah ... the old double terminated SCSI Bus!
------------------------------------------------------
> We're actually using write back caching anyway. (The batteries
> themselves have been back-ordered, but I hope they arrive before we
> actually put this into production and depend on the data...)
Interesting story goes with "why this product is backordered" ... basically is
because of our site. We were among the "lucky few" to have ordered the Battery
kits as soon as they hit the price book. The OEM vendor Mylex sent DEC the
WRONG batteries ... Digital SHIPPED the WRONG batteries ... Arnold called the
product manager to inform him of the Oops. Product placed on Backorder ...
Arnold is now waiting for a 'correct' replacement.
The group in Shrewsbury IS working the issue as fast as they can.
> I'm not sure where I saw a picture of the JA1/JB1 connector and thought
> these were done right but they were *not*. The cables were all on JB1
> (the one that DISABLES the termination!) Yes, this system was built by
> the Digital rackmount systems folks. The cables were all beautifully
> dressed and tied to just the right length, I had to cut it all apart
> and retie it to get the connectors in on the top (JA1) connector.
One of the SWXCR engineers warned me about that ... come to think of it.
Since I'm in the process of 'cabling' up a RM AlphaServer 1000 4/233 as we
speak with a BA356 & BA350 and SWXCR (PCI variety) ... I'll keep this issue in
the forefront of my mind. Thanks in advance for saving many of (additional)
hours of hang wringing.
> Duncan, feeling more terminated now...
All I have to say as the Arnold the Terminator ... "I'll be back..."
The DECUServe Journal October, 1995 Page 15
RAID Controller Woes
Note 1925.10, 31-Aug-1995
Brown: End of story
-------------------
Sure enough, my RAID 5 set configured fine the first time out with the
cables on the right connectors! A lot faster, too- I guess it was
trying its darndest before to retry failed transfers until it got it
right...
So I'm all set, all I need is some batteries for my cache!
Fibonacci Series
--------- ------
Abstract:
Somewhat off the usual beaten track, we find this nicely concise
explanation of the Fibonacci series.
Participants:
Henry Carmichael, Dale Hoehne.
Conference: ARTIFICIAL_INTELLIGENCE
Note 25.0, 14-Sep-1995
Hoehne: Fibonacci Series
------------------------
Looking for a simple explanation of Fibonacci's series.
I found a reference in this forum but not being an AI or
math guru ... it's somewhat hard for "this" layperson to
understand .. Any help would be appreciated .... Any
pointers to internet sites that may have additional info
would also be appreciated ... TIA
Note 25.1, 14-Sep-1995
Carmichael: Fibonacci Series: does this help?
---------------------------------------------
Fibonacci series: a general term for a series where each term n is
equal to the sum of the two preceding terms. In mathematical terms,
n(x) = n(x-2) + n(x-1) [first two terms are special]
To illustrate, usually, the first term is 0, and the "seed" (my name
for it) is arbitrary, but greater than 0. So we have:
1st term: n(1) = 0
The DECUServe Journal October, 1995 Page 16
Fibonacci Series
2nd term: n(2) = 1 = 1st plus seed
3rd term: n(3) = 1 = n(1) + n(2)
n(4) = 2 = n(2) + n(3)
n(5) = 3 = n(3) + n(4)
n(6) = 5 = n(4) + n(5)
...
Some may argue that my first term is really the "zeroth" term, and that
you don't need it. You can pick your own "seed" to generate your own
series (I suppose it could be any real number, but I have only seen
integers used).
So, if you use 3 as the "seed", the series would go:
0, 3, 3, 6, 9, 15, ...
Fibonacci series occur many places in nature - the example that comes
to mind is the number of seeds in successive circles of a sunflower.
Note 25.2, 15-Sep-1995
Hoehne: Thanks
--------------
Thanks Henry,
Your explanation is simple enough ... even "I" can understand :-)
Internet Firewalls
-------- ---------
Abstract:
Sure, Internet security is a big issue; yes, firewalls seem to be
all the rage; but where do you go to find out more about them (and
possibly avoid problems with a vendor)? The following discussion
contains numerous pointers, along with experience and the wisdom
born of it.
Participants: Don Boelling, Bruce Bowler, John Briggs,
Linwood Ferguson, Matt Holdrege, J.M. Ivler, Milton Lopez, Jack
Stevens.
Conference: SECURITY
The DECUServe Journal October, 1995 Page 17
Internet Firewalls
Note 312.0, 26-Aug-1995
Ivler: Internet Firewalls
-------------------------
As the Internet Specialist, one of my many tasks was to implement
security prior to placing the systems on the net (next week we
start on-line). I didn't have a great deal of say about most of
the stuff that happened and was presented with a prechosen vendor
the day I started...
It was the best firewall I have ever seen. Once connected to the
Internet it guaranteed, based on everything I saw, to not allow
any single person access to our internal system.
The firewall software, Firewall-1. The hardware, a SUN Netra.
There was just one little problem... you see, there was just one
ethernet port on the Netra. While this did ensure security, it
also was a bit of a disadvantage for those that actually wanted to
communicate with a firewall between the inner-net and the
Internet.
I have since brought in another vendor (actually, I opened the
bidding up to all vendors, and was deeply disappointed in DEC who
had no real product to offer, and then shucked and jived me on to
a VAR who had no real idea on what to offer... - but that's a
business-whatever conference issue...).
The fun of this whole thing was when the original vendor told me
that I "better take the firewall they provided, or they would sue
me". The sad part is that this was a Sunsoft VAR, and while I was
willing to fight for what was right, just how many jerks like this
are out there doing Internet Security and Firewalls?
Note 312.1, 26-Aug-1995
Holdrege: caveat emptor
-----------------------
There is a plethora of crooks out to make a buck off the Internet.
Firewalls being somewhat arcane, are a good area for these people to
practice the art of the scam.
Your only recourse is to get a good understanding of how firewalls
work. But once you have that, you can build your own firewall for
peanuts.
Note 312.2, 28-Aug-1995
Bowler: Can you recommend a "starting point"?
---------------------------------------------
The DECUServe Journal October, 1995 Page 18
Internet Firewalls
Note 312.3, 28-Aug-1995
Ferguson: Some starting points
------------------------------
Hire a better consultant (there are a couple floating around here and
the one we hired was great but apparently overly busy so I'll let him
chime in if he wants more work).
There's a newsgroup at comp.security.firewalls.
We ended up with a package from Trusted Information Systems, which as I
understand is available in a DIY kit over FTP, or they will sell all
setup (though I do not know what is included in that). That's just
the firewall software not a system to put it on. We got ours on a PC
running unix BSD. Our only problem is we can't do http through it in
the old version we are running, but their newer versions do. So far it
has been completely solid -- only crash was when a bad video card took
it down. We also run a news server on it. I think it was a 486/66.
For the most part we treat it like a black box. That's good and bad.
Good, because we have been able to ignore it due to the fact it just
sits there and runs; bad because our system manager really does not
understand it well enough to do things like upgrade it. Then again,
not being able to run http probably cuts our network traffic way, way
down. Lots of people who want to pay on the internet won't play in
character cell mode. :-)
Note 312.4, 28-Aug-1995
Briggs: Firewall resources
--------------------------
The bible: Firewalls and Internet Security (Repelling the Wily Hacker)
William R. Cheswick and Steven M. Bellovin, Addison-Wesley
ISBN 0-201-63357-4
It is not possible to praise this book enough. It is
precisely the resource you need.
Mailing list: Send to majordomo@greatcircle.com a message containing
"subscribe firewalls" or "subscribe firewalls-digest"
FTP: ftp.greatcircle.com, directory pub/firewalls.
Check out pub/firewalls/FAQ
coast.cs.purdue.edu
newsgroups: comp.admin.policy, comp.protocols.tcp-ip, comp.sources.unix,
comp.sources.misc, comp.unix.admin, comp.unix.wizards,
comp.security.announce, misc.security, alt.security.
We are currently using a home-grown firewall using the TIS toolkit
mentioned by Linwood in .-1
The DECUServe Journal October, 1995 Page 19
Internet Firewalls
Note 312.5, 28-Aug-1995
Holdrege: basics
----------------
As Linwood mentioned, comp.security.firewalls is the newsgroup to scan
and ask questions in. It is a brand new group, so it doesn't have its
own FAQ yet, but there are others floating around.
But the best advice I can give is to start with the basics. Read
Comer's Internetworking with TCP/IP series to fully understand how the
Internet works at the middle layers. Once you have that knowledge, you
will be able to better appraise the various firewall offerings.
Note 312.6, 29-Aug-1995
Stevens: Haven't Had That Problem
---------------------------------
We are using Firewall-1, and are very happy with it and our vendor
(Internet Security Corporation). The hardware we are using is a Sun
SPARCstation, and it can easily be configured with up to four Ethernet
ports.
So I don't understand who is doing what to you.
One of the reasons we did not build our own firewall is that I am very
busy with doing my job and did not want another one doing firewall
research and development.
Note 312.7, 31-Aug-1995
Ivler: We have a firewall...
----------------------------
Okay,
We went with Raptor. It works great, it is GUI based (so our
"firewall admin" doesn't have to understand the innards), and it
used dual-honed DNS to keep us "in the black" from external
probes, but in the clear internally.
It also was an easy setup and install (except for the DNS part).
One nice feature was the use of S/Key one-time passwords for FTP
and Telnet access to the system from the outside.
Note 312.8, 1-Sep-1995
Lopez: And then there's BorderWare
----------------------------------
Some comments about DEC's referral to VARs in .0 are hard to understand. From
trade media announcements (and direct contact with both vendors), I know that
DEC is re-selling the BorderWare Fireall system, a from- the-ground-up,
dont-mess-with-Unix type system that runs on a PC and also provides a split
The DECUServe Journal October, 1995 Page 20
Internet Firewalls
DNS, FTP and WWW servers. We have it. It was a cinch to install, it's sitting
here in my office and nobody else knows it, particularly our users and all
applications running on internal systems. It's inexpensive, and fairly well
supported. See http://www.border.com/ if interested.
Note 312.9, 26-Sep-1995
Boelling: I have a DEC SEAL firewall!
-------------------------------------
I just purchased the full DEC firewall system. Consists of three Alpha
workstations. Works fine. Two are stripped down DEC OSF/1 operating systems
to do the gateway and applications gateway work. The other is a full blown
workstation used for mailgate. I guess I don't understand the comment that
DEC didn't have anything to offer. I got one. It works great. Two ethernet
ports on gate. DNS, Usenet, applications proxy's, are all working fine.
Maybe after I learn how this thing operates a little better I will be able
to answer some questions about it.
If you need the DEC order numbers or the contacts for info, let me know and
I will get them to you.
Domain Name Service and Windows NT
------ ---- ------- --- ------- --
Abstract:
A discussion of where to get a domain name server for NT, with
digressions concerning the 3.5.1 patch and documentation resources.
Participants:
Joe Crum, Joseph Gill, Fred Lobmeyer, George Merriman, John Vottero.
Conference: WINDOWS_NT
Note 154.0, 25-Sep-1995
Gill: How to config TCP/IP & DNS on WNT 3.5
-------------------------------------------
I just inherited a Window NT machine. The machine needs to be
configured with TCP/IP using Domain Name Services. Since I'm new to
Windows NT, I had a few basic questions,
o The machine is running WNT Advance Server 3.5. Does 3.5
come with TCP/IP? Does it support DNS?
o I thought I read version 3.5.1 is out. Is it worth upgrading
The DECUServe Journal October, 1995 Page 21
Domain Name Service and Windows NT
to? Is the patch free?
o Since no one can find the documentation for NT box, is there
any sites (FTP, WWW, FAQs, etc) with info on NT?
Note 154.1, 25-Sep-1995
Merriman: Plenty of reading material on WNT
-------------------------------------------
> o The machine is running WNT Advance Server 3.5. Does 3.5
> come with TCP/IP?
Yes.
> Does it support DNS?
If you mean "Can it use TCP/IP DNS for address resolution?",
the answer is yes. WNT Server doe not provide DNS server
functionality out of the box.
> o I thought I read version 3.5.1 is out. Is it worth upgrading
> to? Is the patch free?
The latest version is 3.51. I paid about $70 for the upgrade.
At that price, it is probably worth it.
> o Since no one can find the documentation for NT box, is there
> any sites (FTP, WWW, FAQs, etc) with info on NT?
Spring for the WNT Resource Kit. I think the list price is about
$150, but I paid $99, as I remember. It consists of four printed
volumes and a CD-ROM. Microsoft also offers a Self-paced
training course on NT Server. There are also loads of third party NT
books out these days.
You can also look in ftp.microsoft.com, and www.microsoft.com.
Note 154.2, 26-Sep-1995
Vottero: DNS is in the resource kit
-----------------------------------
DNS server software is included in the Resource Kit. It's also
available from ftp.microsoft.com.
Note 154.3, 27-Sep-1995
Lobmeyer: Latest Version is 3.51 Service Pack 1
-----------------------------------------------
Just an update to the latest version is
WNT 3.51 Build (1057:Service Pack 1)
The DECUServe Journal October, 1995 Page 22
Domain Name Service and Windows NT
You can get the Service Pack at
http:://ftp.microsoft.com/bussys/winnt/winnt-public/fixes/NT351/ussp1/
Note 154.4, 29-Sep-1995
Crum: Resource kit also has the DNS server & HTTP server
--------------------------------------------------------
Also, if you have the NT CD, the manuals are in the SUPPORT\BOOKS
directory in Windows help file format.
WHOIS for Windows
----- --- -------
Abstract:
Several references to WHOIS software and related Winsock stuff.
Participants:
Shawn Allin, Charlie Byrne, Karl Erbland (via DCS).
Conference: INTERNETWORKING
Note 391.0
Allin: Windows version of WHOIS?
--------------------------------
I don't know if this belongs here, or in PC Software, but here goes.
I'm looking for a MS Windows-based freeware/shareware implementation of
WHOIS. Does anyone know of such a beast?
Note 391.1, 25-Sep-1995
Byrne: Internic pointer
-----------------------
re windows WHOIS, don't know of one but you might want to nose
around
http://INTERNIC.NET/rs-internic.html
The DECUServe Journal October, 1995 Page 23
WHOIS for Windows
Note 391.2, 25-Sep-1995
Erbland: Have several WHOIS for Windows
---------------------------------------
I have several. I will post them to the LINX - LUG Information News
Exchange's FTP area or will find the direct pointer for the file(s).
I will also post the answer here when I find out for sure. Give me a
few minutes and I'll return with more info.
Note 391.3, 26-Sep-1995
Erbland: Some WHOIS info...
---------------------------
Two WHOIS zipped files for Windows are:
winwhois.zip -- whois
wsfngr15.zip -- finger/whois
Try most any Windows shareware/PD areas. You will usually find them
under a "winsock" subdirectory for Windows 3 files.
I will place these in the LINX FTP directory tomorrow. They will be
available after 10:00 AM EDT. Using a web browser pointed to:
http://www.sims.net/linx
Follow the directions from the LINX Main Page for FTPing files. Other
FTP instructions are available by e-mailing me at:
karl@ksme.net or karl@linx-pubs.com.
Note 391.4, 26-Sep-1995
Allin: Thanks
-------------
Thanks Karl,
I'll look for them this morning. I actually did look at CICA before
asking here and that was the one utility that I couldn't find.
Note 391.5, 1-Oct-1995
Erbland: WWW site for Winsock applications - TUCOWS
---------------------------------------------------
The following Web link is one I knew about but forgot to mention. It is:
The Ultimate Collection of Winsock Software (TUCOWS)
http://www.tucows.com/
You will absolutely love this Web site.
The DECUServe Journal October, 1995 Page 24
WHOIS for Windows
Every Winsock application you could ever want is here, and more (some
software for other platforms).
Uses a unique bovine rating system! This is DEFINITELY a Five Cower!
Heavy graphics and tables.
You can also visit the TUCOWS link from the LINX FTP / Download area
which is available from the LINX pages:
http://www.sims.net/linx
Let me know what you think of TUCOWS.
About the DECUServe Journal
---------------------------
Publication Information
Topic threads in the DEC Notes conferences on DECUServe are selected
for publication on the basis of strong technical content and/or
interest to a wide audience. They are submitted to the editor from
various sources, including DECUServe Moderators, Executive Committee
members, and other volunteers. Suggestions for inclusion are
enthusiastically solicited. Articles selected for publication are
edited on an OpenVMS VAX system in TPU and then formatted with
Digital Standard Runoff.
What's a DECUServe, Anyway?
DECUServe is an on-line conferencing system which is available 24
hours a day, seven days a week. System backups are done on Fridays
from 7:00am to approximately 9:00am Eastern (Daylight/Standard)
Time. Scheduled downtimes for equipment or software maintenance are
announced in advance. The system can be reached worldwide, and
membership is by individual subscription only (no group or company
accounts). The current annual subscription fee is US$75.00.
On-line subscription information is available in the U.S. by
dialing 1-800-521-8950 and logging in with username INFORMATION.
DECUServe and the INFORMATION account can also be reached on the
Internet via telnet connection to decuserve.decus.org or e-mail to
information@decuserve.decus.org.
Contact Information
-------------------
The editors of the DECUServe Journal are Brian and Sherrie McMahon.
They can be reached by any of the following means:
The DECUServe Journal October, 1995 Page 25
WHOIS for Windows
mcmahon_b@decuserve.decus.org
mcmahon_s@decuserve.decus.org
mcmahonb@decus.org
griffith@decus.org
mcmahon@ac.grin.edu
griffith@ac.grin.edu
+1 515 269 4901 (normal office hours, U.S. Central time)
+1 515 269 4936 FAX