DECUServe Journal April 1995
Brian McMahon, Info-VAX Refugee
The DECUServe Journal
=====================
April, 1995
From the Editor's Keyboard
==========================
Welcome to Yet Another Highly-belated DECUServe Journal. We
apologize for the long delay in getting this April issue out (it is
still April, isn't it?); your devoted editors have been suffering
from severe fragmentation of the calendar all month long.
This issue's mix consists of a few things that are new, relatively
speaking, and a few that are old. We're constantly monitoring the
various technical conferences, making note of interesting topic
streams. Not everything we make note of gets used immediately, so
some very deserving items languish for months on end. Caught up in
the spirit of spring cleaning, we've decided to flush a few of those
buffers. And so it happens, for example, that the last three
articles all come from the ponderously titled
USER_INTERFACES-WINDOWS conference, and all are over a year old.
Right about the time we should be working on next month's issue,
some of us will be at the U.S. DECUS event in Washington, D.C. --
if you happen to stop by the terminal farm at just the right time,
you might see Journal production work in progress. (Follow the
distraught wailing sounds -- just kidding!) We'll try to have the
May issue out in a more timely fashion, possibly with an emphasis on
hot topics at the symposium. Stay tuned....
The DECUServe Journal April, 1995 Page 2
Table of Contents
Table of Contents
=================
CONTENTS
From the Editor's Keyboard . . . . . . . . . . . . . 1
Table of Contents . . . . . . . . . . . . . . . . . 2
MicroVAX Boot Message . . . . . . . . . . . . . . . 3
Trouble Reading 4mm Tape . . . . . . . . . . . . . . 5
SCSI Troubleshooting . . . . . . . . . . . . . . . . 8
CD-ROM Writing for OpenVMS . . . . . . . . . . . . 13
Open Channels to Disk . . . . . . . . . . . . . . 17
Document Management Systems . . . . . . . . . . . 18
All About DECamds . . . . . . . . . . . . . . . . 22
X colormap . . . . . . . . . . . . . . . . . . . . 30
Forcing a Window to the Top . . . . . . . . . . . 31
EVE DECwindows Configs . . . . . . . . . . . . . . 34
About the DECUServe Journal . . . . . . . . . . . 36
Contact Information . . . . . . . . . . . . . . . 36
The DECUServe Journal April, 1995 Page 3
MicroVAX Boot Message
MicroVAX Boot Message
=====================
This issue's lead-off article is taken from the HARDWARE_HELP
conference. The scenario is a particularly dreaded one -- equipment
with no documentation. So what does that funny message at boot time
mean...?
Participants: Arnold De Larisch, Paul Flaherty, Jamie Hanrahan,
Chuck McMichael, Gregg Nelson, Pat Scopelliti.
Note 1824.0, 21-Feb-1995
Flaherty: MicroVAX 3100e Boot Message
-------------------------------------
I've inherited a MicroVAX 3100e, without any hardware documentation, naturally.
The machine seems to run fine, but I get a message at boot time that I am
unable to look up in non-existent documentation...
KA41-D V1.0
F_..E...D...C...B...A...9...8...7...6...5...4_..3_..2_..1..
? C 0000 0000.4001
83 BOOT SYS
-DKA0
OpenVMS &c...
Can anyone interpret the "? C" stuff above for me? Also, if you can give me
the part number for the manual in which you found it, I'd appreciate it.
Note 1824.1, 21-Feb-1995
De Larisch: You want the MicroVAX 3100 Maintenance Manual
---------------------------------------------------------
The manual you want is the MicroVAX 3100 series Maintenance Manual. I'm
at home right now so I don't have it here to give you the exact title
and/or part number.
If someone else doesn't post it by tomorrow, please send me a quick
e-mail message.
P.S. Fair warning ... the book was about $100 back a couple of years
ago!
Note 1824.2, 22-Feb-1995
McMichael:
-----------
>Can anyone interpret the "? C" stuff above for me? Also, if you can give me
>the part number for the manual in which you found it, I'd appreciate it.
The DECUServe Journal April, 1995 Page 4
MicroVAX Boot Message
Try typing SHOW CONFIG or SHOW ERROR at the >>> prompt for more info.
Note 1824.3, 22-Feb-1995
Nelson: C indicates serial controller
-------------------------------------
? C 0000 0000.4001
I can possibly shed a little light. My VAXStation 3100 Owner's Manual
indicates the above message is about the serial line controller.
Note 1824.4, 22-Feb-1995
Flaherty: Thanks for the help so far, folks!
--------------------------------------------
Note 1824.5, 23-Feb-1995
De Larisch: VS 3100 Maintenance Manual EK-285AA-MG
--------------------------------------------------
The book I was referring to .-couple is the VS 3100 Maintenance Manual (part
number EK-285AA-MG). I think this manual set us back about $100 but has paid
for itself time and time again.
There is probably a sister publication for the MV 3100 series, but we don't own
any.
Note 1824.6, 23-Feb-1995
Hanrahan: Does this cover the VS 3100 model 76 also?
----------------------------------------------------
Note 1824.7, 8-Mar-1995
Scopelliti: 3100 Hardware info
------------------------------
Finally got a chance to check out my docs:
? C 0000 0000.4001
The single "?" indicates a minor problem
The "C" does mean the serial line controller
The 0000.4001 means
This line is untested (4)
No error (1)
My guess is you don't have a loopback connector in the port.
>> KA41-D V1.0
The DECUServe Journal April, 1995 Page 5
MicroVAX Boot Message
>> F_..E...D...C...B...A...9...8...7...6...5...4_..3_..2_..1..
^ ^ ^ ^
The underscores simply mean the option is not installed or was not
tested. The options are:
F Reserved for later use (Base video on 3100/38)
E System clock
D Nonvolatile RAM
C Serial line controller
B Memory
A Memory-management unit
9 Floating point unit
8 Interval timer
7 SCSI-A bus controller
6 SCSI-B bus controller
5 Interrupt controller and Ethernet ID ROM
4 DSH32-B communications module asynchronous lines
3 DSH32-B communications module synchronous lines
2 Reserved for later use
1 Ethernet Network Interconnect (NI)
So.. your 3100 doesn't have a video board, either DSH32-B, or the
"Reserved for future use" device ;-}
The usual annoyances are:
? C 0000 0000.4001 No loopback on serial port(s)
? 7 00A0 0000.4001 No devices on internal SCSI-A bus
? 6 00A0 0000.4001 No devices on external SCSI-A bus
? 1 00C0 0000.7004 No loopback connector or no cable on
ThinWire ethernet connector.
? 1 00C0 0011.700E No loopback connector or no cable on
AUI ethernet connector.
Hope this helps...
Note 1824.8, 9-Mar-1995
Flaherty:
----------
Hey, thanks for the followup!!!
Trouble Reading 4mm Tape
========================
Next, also from HARDWARE_HELP, we have a tale of woe involving a 4mm
DAT tape written on a TLZ06 and intended to be read on an HP drive.
Sounds like it should be easy, right? And we all know about those
things that "Should Be Easy", right?
The DECUServe Journal April, 1995 Page 6
Trouble Reading 4mm Tape
Participants: Dale Coy, Malcom Dunnett, Keith Hare, Terry Kennedy.
Note 1834.0, 1-Mar-1995
Hare: Problem reading TLZ06 4mm on HP C1533A
--------------------------------------------
I am attempting to read a 4mm tape with VMS Backup and am having problems.
The tape was written on a TLZ06 under AXP VMS V6.1 on a AXP 2000-500. The
tape was initialized with /media_format=compaction and the backup command
had a /media_format=compaction switch.
I am attempting to read it on an HP C1533A under AXP VMS V6.1 on a DEC
3000-400.
The tape is a maxell HS-4/120s.
Basically, the tape just sits there and rocks back and forth until I get a
"%MOUNT-F-DRVERR, fatal drive error".
I know I can read tapes that are backed up without compaction using this
combination of hardware. Is it possible that the two drives have
implemented compaction differently?
Has anyone tried this combination? Been successful? Have any insights
into what the problem is?
Note 1834.1, 1-Mar-1995
Coy: Could be
-------------
>combination of hardware. Is it possible that the two drives have
>implemented compaction differently?
Yep. There isn't a single "industry standard" in this area. The only
place that there is a standard, is in 8mm drives -- because they are
all proprietary to the same manufacturer.
Note 1834.2, 1-Mar-1995
Kennedy: Compatibility is better than that
------------------------------------------
Eh? There are 3 DAT formats in use today:
DDS (Digital Data Storage) - the original DAT format, no compression
DDS-DC (DDS with Data Compression) - compressed data, 4-8GB on a 90M tape
DDS-2 - (DDS 2 8-) - compressed data, new algorithm, 8-16GB on a 120M tape
The newer drives read all of the old formats (a DDS-2 drive can read DDS-DC
and DDS tapes). Older drives can have problems reading newer media if a) it
is written using a compression method they don't understand, and/or b) it is
written on a tape longer (and thus thinner) than they support.
The DECUServe Journal April, 1995 Page 7
Trouble Reading 4mm Tape
I believe you are having both of those problems - trying to read a 120M tape
written in DDS-2 on a DDS-DC drive.
Note that just because tapes are interchangable at the raw block level
doesn't mean that they can be read on dissimilar OS's - DAT, like 8mm, supports
both fixed-length and variable-length records. Then, of course, there's the
logical data format (VMS BACKUP, tar, IBM standard label, etc.).
Note 1834.3, 1-Mar-1995
Dunnett: HP C1533A is DDS-2
---------------------------
The HP C1533A ( which the original poster said he is trying to read the tape
on ) is a DDS-2 drive, so this shouldn't be the problem.
Note 1834.4, 2-Mar-1995
Kennedy: Oh
-----------
Hmmm. It's always possible DEC did something brain-dead with the firmware
in the drive that wrote it, though that seems unlikely (well, at least the
part about messing with the compression).
Note 1834.5, 2-Mar-1995
Hare: Autosens on DDS 2 Tapes
-----------------------------
The maxell tapes are rated as DDS2. Is it possible that the TLZ06 is
writing the tapes using DDS-1 and the HP C1533A is sensing the tape type
and attempting to read them as DDS-2?
Note 1834.6, 2-Mar-1995
Dunnett: Can't use 120 meter tapes in TLZ06
-------------------------------------------
Bingo.
I just realized the TLZ06 is not a DDS-2 drive, it is DDS-DC.
The Maxell HS4/120S tape you are using is a 120 meter tape. These
are only designed to be used in DDS-2 drives.
The tape has a "media recognition stripe" on it, which the HP drive
reads when the tape is loaded ( you can disable this by turning
dip switch 3 on the bottom of the drive to the ON position ).
I suspect the HP drive is determining this tape is a DDS-2 tape and
so is trying to read it in DDS-2 format. It's possible that if you disable
the media recognition ( as described above ) you will be able to read
The DECUServe Journal April, 1995 Page 8
Trouble Reading 4mm Tape
the tape, but I wouldn't count on it.
Hopefully you can make a new tape of this data, using a 60 or 90
meter tape. I recommend you not use any more 120 meter tapes in the TLZ06,
at best you'll get unreliable results, at worst you'll destroy the tapes or
damage the drive.
Note 1834.7, 3-Mar-1995
Hare: 90 meter tapes work!
--------------------------
After some discussion with some Digital customer support, they also
emphatically said that 120 meter tapes should not be used in TLZ06.
I have a 90 meter tape from the same tape drive and it is restoring on the
HP tape drive just fine.
Well, I know understand 4mm tape drives a lot better.
Thanks for all of the input.
SCSI Troubleshooting
====================
Continuing on our HARDWARE_HELP theme, we move from DAT to SCSI. If
you're like us, the terms "SCSI" and "flaky termination problems"
seem to go together naturally. In that case, you just may find
enlightenment herein. There is a bonus reference (included at no
extra charge) to what looks like a very useful DEC PC BSS at the
end.
Participants: Barton Bruce, Linwood Ferguson, Terry Kennedy, Gerald
Krulewicz, George Merriman, Alan Striegel.
Note 1847.0, 18-Mar-1995
Ferguson: DEC XL 466d2, SCSI CDROM hangs in Windows, Termination?
-----------------------------------------------------------------
I just got a DECpx XL 466d2 from DEC with PCI and on-board SCSI, with a
CDROM and 535 MB SCSI disk.
Everytime, under Windows for Workgroups, that I do large transfers from
the SCSI CDROM the system locks up (large = anything more than a
directory, like copies, running a program, etc.)
The CDROM is a RRD43-AA (Toshiba, apparently), and is last on the
internal cable, and does not have a terminator pack installed, nor is
the Termination jumper ("Shared Terminator Power") installed. The
external SCSI port is not in use, and does not have a terminator on it.
The FE who set this up said all that is correct. The same symptoms
The DECUServe Journal April, 1995 Page 9
SCSI Troubleshooting
occur whether or not the external scsi is enabled in cmos.
Does it sound right?
I have found no problems using it in DOS (6.22).
The SCSI utilities and drivers look a bit old (NCR SCSI Drivers release
1.1), but I cannot find any newer (or any at all) on gatekeeper, and
did not see an FTP site for NCR.
Anyone else using one of these systems and/or have any advice? Tried
DEC, but no one around on Saturday (well, we did the "hardware or
software" dance for a while but it seemed neither were available except
for "premier" customers "who pay $144/year" -- I pay a lot more than
that but I must not pay the right people).
Note 1847.1, 18-Mar-1995
Ferguson: Fixed
---------------
I may have answered my own question after beating my head against it
for a while. Will leave it here in case anyone else is interested or
someone has different advice.
First, there is an NCR ftp site, do not know why it did not work at
first. It's FTP.NCR.COM.
I grabbed the new drivers from NCR and installed the new CDROM.SYS.
Unlike the old, this required a CAM or similar driver, so I installed
it also. That driver, unlike CDROM by itself, gives some info as it
starts, and told me that the NCR SDMS system is using IRQ 15 for SCSI.
I was using it for the nic card. DEC said SCSI was 11 and the manual
said nothing and MFT and QAPLUS said nothing. Sigh.
So with new driver and new IRQ's in hand, reboot -- works fine.
So, though it worked I wanted to know why: Went back to old drivers and
corrected IRQ -- failed again. So the problem is the CDROM driver that
DEC delivered with the DEC PC and DEC (bundled) CDROM. Sigh. Though to
be fair I did install Windows for Workgroups; maybe it would work with
Windows (which I think came with this slightly older PC).
What was really seriously weird about this was that before putting in
the new driver, I had a .ZIP file with it. Both WINZIP and PKUNZIP
refused to extract CDROM.SYS to an empty directory claiming "disk full"
in Winzip and "invalid filename" in PKUNZIP. I used a different PC to
extract it (worked fine) and now, running the new driver, WINZIP can
extract it just fine (same identical file and locations). Almost like
it was in there somehow trying to protect itself from being replaced
(and neither input nor output of the zip was the CD). The zip file
tested good (even when it failed to extract) and everything else in it
came out fine. Expected to see Rod Serling any moment.
The DECUServe Journal April, 1995 Page 10
SCSI Troubleshooting
Note 1847.2, 18-Mar-1995
Kennedy: SCSI termination rules
-------------------------------
Glad to see you got it fixed (from .1) but here's some SCSI hints. The
SCSI bus needs to have exactly two terminators - one on each end of the
bus. If the controller is at one end (no external devices) it needs to be
terminated. Likewise for disks/tapes/CD's/etc. on the internal and exter-
nal buses.
You should look at your SCSI disk and make sure it isn't terminated, and
then enable the terminators (2 10-pin SIP resistor packs) on the CD-ROM, as
long as the CD-ROM remains the last device on the internal bus.
Remember - a terminator at each end - that means if there's an internal
device *and* an external device, no terminators on the controller as it's
no longer at the end of the bus.
Note 1847.3, 18-Mar-1995
Ferguson: Understood the need, having trouble with the implementation
---------------------------------------------------------------------
> Remember - a terminator at each end - that means if there's an internal
>device *and* an external device, no terminators on the controller as it's
>no longer at the end of the bus.
I knew that in principle, it is in application I get confused.
The scsi adapter is a chip on the motherboard. I can find no jumpers
for it regarding termination, nor any obvious place for a connector or
resistor pack. DEC said it was "internally terminated" when he
installed it and "the cables are so short that is all you need".
Since I thought it was working and I usually think of DEC FE's as
board swappers I did not press and assumed whoever put it together did
it right.
The CDROM does not have the resistor pack installed, and is at the end
of the internal cable. No resistor pack came with it either.
Any idea, by disabling the external SCSI connector in CMOS, if I
somehow remove the need to terminate it externally? (what exactly does
it mean to disable it anyway; I assumed it was a continuation of the
same bus) The external connector is one of the mini connectors and I
have no terminators that I can find for those either.
And when there is an internal and external connector like this, where
is the controller: in the middle somehow? I.e. with respect to
terminating the ends?
I'll try talking to DEC on Monday and see if I can find out a bit more.
Even if working, I worry if it is not configured right.
The DECUServe Journal April, 1995 Page 11
SCSI Troubleshooting
Note 1847.4, 19-Mar-1995
Bruce:
-------
> resistor pack. DEC said it was "internally terminated" when he
> installed it and "the cables are so short that is all you need".
It is possible that there is a terminator that can be gated on or off
either by a jumper or dipswitch or by some configuration command.
Some snazzy active terminators may not simply unplug, but can be made to
'disappear', and might be what they have on the mother board.
Note 1847.5, 19-Mar-1995
Merriman: I'm not to sure about internal terminators.
-----------------------------------------------------
When I installed my external TEXEL SCSI CD-ROM, It claimed to
have a switch selected internal terminator as well. I found that
with the terminator turned off, the drive wouldn't work at all.
With the internal terminator on, it worked but was flakey. With
an external terminator plugged into the SCSI cable at the CD-ROM
it has been working fine for two years now.
Note 1847.6, 20-Mar-1995
Krulewicz: PC XL's
------------------
On a DEC PcXL: When you disable the external SCSI bus in CMOS, the external
bus is terminated by the system.
All the XL's that have come through here have had a terminator "plug" at the
end of the internal SCSI bus cable. The hard disk and the CDroms are
connected to other connectors on the internal cable.
Also, in CMOS, the on board SCSI bus must be enabled in the PCI section and
you must assign an IRQ for it.
Good luck.
Note 1847.7, 20-Mar-1995
Ferguson: Where is my SYSGEN CONFIGURE when I need it?
------------------------------------------------------
Do I feel silly. Indeed, when I unwrap all the nice, neat cables I see
that the CDROM was not at the end like I thought, but in the middle,
and indeed there is a terminator on the end.
Maybe that was what my FE meant by "internally terminated". :-)
> Also, in CMOS, the on board SCSI bus must be enabled in the PCI section and
> you must assign an IRQ for it.
The DECUServe Journal April, 1995 Page 12
SCSI Troubleshooting
Now that's interesting. It is setup to "None" and it is working.
Though when the CAM program starts it does display an IRQ.
Interestingly CAM displays 15 and MINICAM (which I switched back to)
showed 95. I went in to change it to 15, and found it only accepts
"11" and "None". I tried 11, and now MINICAM shows 11 when it starts
(reassuring). Everything seems to work either way, but will leave it
at 11 (appearing as 95 is a bit disconcerting).
Also interesting two PCI slots are shown as enabled there, 1 and 2.
Slot 1 is the video card and set with IRQ none which seems reasonable.
Slot 2 is set to Enabled with IRQ 10, but nothing is in that slot.
I assume I should disable slot 2. Did, and it works OK so far.
BTW, I tried calling DEC support. Did a dance around the hardware vs.
software issue (since it originally dealt with drivers), but finally
said "hardware". Now we are stuck on "obligation number". The PC
serial number is not in their database, and they cannot accept
a contract DEC number. Sigh.
Note 1847.8, 21-Mar-1995
Krulewicz: Disable #1 Also.
---------------------------
> Also interesting two PCI slots are shown as enabled there, 1 and 2.
> Slot 1 is the video card and set with IRQ none which seems reasonable.
> Slot 2 is set to Enabled with IRQ 10, but nothing is in that slot.
> I assume I should disable slot 2. Did, and it works OK so far.
You should disable the unused slot and also the slot the video controller is
in (usually number 1.)
There is a MS help file for field service personnel for the new Celebris XL
(really the PCXL with new firmware), that goes into some detail about the
CMOS settings. It can be found on the DEC PC BBS.
Note 1847.9, 21-Mar-1995
Ferguson: What's a DEC PC BBS?
------------------------------
> You should disable the unused slot and also the slot the video controller is
> in (usually number 1.)
I tried. It did boot OK, but on going into windows it went nuts. First
claimed a corrupt swap file, then a trashed logo, then an "error
loading GDI" and exit to DOS. Set back to enabled and everything came
up fine. Could slot 1 not be the first slot, but something internal?
> There is a MS help file for field service personnel for the new Celebris XL
> (really the PCXL with new firmware), that goes into some detail about the
> CMOS settings. It can be found on the DEC PC BBS.
What's a DEC PC BBS? I don't see anything like that in DSNlink or
The DECUServe Journal April, 1995 Page 13
SCSI Troubleshooting
Gatekeeper. Is there yet another place?
By the way, re the SCSI interrupt: finally found someone at DEC
hardware support to talk to me. He said it should be set to "none" not
"IRQ11". I kind of like it working at 11 with the MINICAM showing IRQ
11 instead of IRQ 95 so think I will leave it there for now. He did
not sound real competent, anyway.
Note 1847.10, 21-Mar-1995
Striegel: DECpc BBS 508-496-8800
--------------------------------
There is a Bulletin Board System maintained by Digital's DECpc
Engineering folks found at 508-496-8800. This BBS has lots of the diagnostic/
driver/utility software needed to make best use of Digital's PCs.
CD-ROM Writing for OpenVMS
==========================
Now that our SCSI problems are sorted out, we switch to the VMS
conference, where we find a discussion of options for making one-off
CD-ROM disks. The conversation ranges from general options (what
hardware exists) to specific experiences with InfoServer setups.
Participants: Wayne Bruzek, Dan Esbensen, Larry Kilgallen, Gary
McCready, Gerardo Razumney, Tony Scandora, Pete Sivia, Jeffrey Sue.
Note 2492.0, 9-Mar-1995
Esbensen: Need to Make One-Off CDROMs for OpenVMS
-------------------------------------------------
We need to create some "one off" CDROMs to be used with OpenVMS. Has
anyone had any luck finding a company (other than Digital) that can do
this?
Can the CDROM Publishing System that Digital sells be used to make
OpenVMS CDROMs?
Note 2492.1, 9-Mar-1995
Kilgallen: Just like making a CD-ROM for mass release...
--------------------------------------------------------
What we do is use a CD writer (Sony 900-W in our case) attached to the
Infoserver (bought in pieces, not as a "Publishing System") to write
the CD-ROM.
That is exactly the same process we use to create a master for the
pressing plant.
The DECUServe Journal April, 1995 Page 14
CD-ROM Writing for OpenVMS
Note 2492.2, 10-Mar-1995
Bruzek:
--------
I'm interested too. Is the SONY an expensive device?
Note 2492.3, 10-Mar-1995
Kilgallen: $7000
----------------
That Sony model was $7000 18 months ago.
Note 2492.4, 10-Mar-1995
Esbensen: Any Special formatting insztructions or other tricks?
---------------------------------------------------------------
Do you format the cdrom in a special way or anything like that?
Note 2492.5, 10-Mar-1995
Kilgallen: Native format
------------------------
For VMS we use ODS-2. For Macintosh we use HFS.
Operating systems sometimes have features they support for their own
native formats which they will not support for ISO 9660.
Note 2492.6, 13-Mar-1995
Scandora: $2000?
----------------
Larry Kilgallen remembered $7000 for a SONY device 18 months ago.
I just saw an ad in the back of one of the PC newspapers for a $2000
device. Sorry, I've already recycled the paper, but if you're
interested, have a look. I don't know how hard it might be to hook it
up to a VAX or if there is a way to send physical blocks from an ODS-2
volume to a CD writer on a PC.
Note 2492.7, 13-Mar-1995
Sue: Device errors from recorded CDs
------------------------------------
I'm curious in how this device performs for others out there.
We purchased the Sony CDW drive, are using the InfoServer V3.1
software on an IS1000 (4MB memory), and using Kodak
CDs.
Unfortunately, using backup/image from the recorded CD to
an RZ disk drive causes many (sometimes *lots*) of device
The DECUServe Journal April, 1995 Page 15
CD-ROM Writing for OpenVMS
errors to be logged. These are all of the type "extended sense
data received", and we haven't verified that the files themselves
are bad, but that's kind of tough to verify on 650 MB of data.
The errors are reported if using an RRD42 to read the data. If
we use an RRD43, we don't see the errors... but this could merely
be a masking of the errors by DEC. In any case, if our end users
have RRD42s, they'll see the errors (sometimes in the thousands),
and their confidence in the information on the disk is lowered
considerably.
These errors occur regardless of whether we use 1X or 2X recording,
and even if we use 3M CDs.
I really need to know if anyone has had experiences such as these and
has found a solution. For the time being we've had to resort to
sending out TK tapes to a service to make the 20 or so CD copies of
each and provide them within 6 days.
Note 2492.8, 14-Mar-1995
Esbensen: Can you tell us the name of teh service?
--------------------------------------------------
What service do you use...and what kind of cost is there for each
on-off CD?
We might want to go this route to start with -- waiting for prices to
drop and for technology to get better!!
Note 2492.9, 14-Mar-1995
Sivia: DEC's Media and Replication Service
------------------------------------------
I work with Jeff Sue.... the service is DEC's Media and Replication
Service.
Note 2492.10, 14-Mar-1995
Kilgallen: What disk for recording ?
------------------------------------
> Unfortunately, using backup/image from the recorded CD to
> an RZ disk drive causes many (sometimes *lots*) of device
> errors to be logged. These are all of the type "extended sense
> data received", and we haven't verified that the files themselves
> are bad, but that's kind of tough to verify on 650 MB of data.
What disk are you using on the Infoserver to record the CD. It must be
at least an RZ26, I believe, due to firmware issues.
The DECUServe Journal April, 1995 Page 16
CD-ROM Writing for OpenVMS
Note 2492.11, 16-Mar-1995
McCready: ISO-9660 format CD's on VMS?
--------------------------------------
If I just wanted to use that $2000 PC-based CD-rom writer, which
creates the ISO-9660 format CD's, would my VMS system (post 5.5, I
think...) be able to read them without problem? Our need is to
store flat data files, which are read in by vms fortran programs to
for analysis.
Would there be any real restrictions on what could be done? All I need
are the files to be accessible - security is not an issue (as long as
they are world-readable).
Note 2492.12, 17-Mar-1995
Sue:
-----
> What disk are you using on the Infoserver to record the CD. It must be
> at least an RZ26, I believe, due to firmware issues.
We have RZ28's in a BA350. We create InfoServer Partitions and
RECORD from them (You have to, unless you're using RZ25s, because
an RZ26 is larger than the CD capacity of 650MB).
The service we've gone with is, as Pete said, Digital's service. There
has been an issue over what media they need us to send them (TK85?
TK86?... they don't always know, and for some reason they think they
can't get the tape drives for the media we sent them. This is DEC,
right? In New Hampshire, right? A large site, right? They can't get
the correct hardware? Go figure!).
The usual cost is somewhere around $1,000 for the initial mastering fee,
and then something like $1.25 (give or take) per CD created. Thus, if
you're really cost conscious, and if you will be doing this more than
once a year, you need to be making somewhere around 40 or 50 CDs per
master to meet up with the 3-year cost of purchasing the InfoServer,
software, RZ26, and CDR. Actually, it's been awhile since I looked
into this, and hardware costs for the different components has changed
drastically, so anyone interested should evaluate this for their own
situation.
Apparently there's some question as to whether our hardware setup may
be causing some of the problems. We have the IS1000, an RRD40, the
CDR, and a BA350 on the SCSI chain. DEC thinks we might have a SCSI
cable length issue, so the next attempt would be to include *only* the
CDR and the BA350 on the chain. We're not sure when we'll be able to
run that test as there's lots of project implementation activity at the
moment. If we ever do solve this issue, I'll try to remember to put an
update in this notes topic.
The DECUServe Journal April, 1995 Page 17
CD-ROM Writing for OpenVMS
Note 2492.13, 17-Mar-1995
Razumney: CDCOPY
----------------
> creates the ISO-9660 format CD's, would my VMS system (post 5.5, I
> think...) be able to read them without problem? Our need is to
> store flat data files, which are read in by vms fortran programs to
> for analysis.
You could try using the CDCOPY utility available here in the directory
DECUSHARE0:[VMS93A.TK.CD_ROM]
I have been using it without problems, although I have not tried a great
variety of files -- my need was for one specific application.
Open Channels to Disk
=====================
Rounding out the set of reasonably-recent articles, we have the
following discussion of file-system arcana, again from the VMS
conference. The basic problem can be summarized by the question:
When do you get a channel open to a disk, but no file associated
with the channel?
Participants: Larry Kilgallen, Seth Stern.
Note 2502.0, 23-Mar-1995
Stern: Channels assigned to disks with CCB$L_WIND=0
---------------------------------------------------
What is the meaning of a channel, held by a process, to a disk device
which does not have a WCB or process section associated with it
(CCB$L_WIND = 0)? I am trying to track down a problem where a process runs
out of channels. Looking at the channels with SDA (show process/channel) shows
numerous channels to the user disk, but no associated files.
Note 2502.1, 23-Mar-1995
Kilgallen: $ASSIGN without $DASSGN
----------------------------------
Deep down inside, somebody did a $QIO/IO$_DEACCESS without a $DASSGN.
Or maybe they did a $ASSIGN but never used the results.
How to accomplish this at a higher level depends on your programming
environment. You could try SET WATCH FILE/CLASS=MAJOR to watch the
history of how a program behaves. Maybe there is even some help on
ASSIGN from SET AUDIT/ALARM/ENABLE=(ASSOCIATE), but I am not certain.
The DECUServe Journal April, 1995 Page 18
Open Channels to Disk
Note 2502.2, 24-Mar-1995
Stern: That's the ticket!
-------------------------
Thanks! I had forgotten about SET WATCH FILE .
Note 2502.3, 24-Mar-1995
Kilgallen:
-----------
Be sure to report back on what you found, so we can add it to our
exhaustive statistics on the nature of bugs :-).
Note 2502.4, 30-Mar-1995
Stern: LIB$FIND_FILE caught in the act!
---------------------------------------
The joys of long-distance debugging of other people's code. :-)
The upshot is that there are three cases where you can expect to find a channel
to a file oriented disk with no associated file:
1. While opening the file. The channel is assigned first, then the file
is accessed. This is a very small window in time.
2. While closing a file. The file is deaccessed, then the channel is
deassigned. Again, brief interval.
3. When you want to look around a disk, but don't want to open a file.
This can leave a channel around indefinitely.
The application has a call to lib$find_file . Under most conditions
it failed to generate matching calls to lib$find_file_end . Each mismatched
call resulted in one extra channel to the disk.
A combination of set watch file/class=major and adding user action
routines to the Pascal open and close calls to dump all of the
conditions and channel numbers resulted in enough information to track down
the problem.
Document Management Systems
===========================
The following thread from OFFICE_AUTOMATION is a discussion of
document management systems. It began in '91, woke up again in '93,
and again in '94. Maybe we're about due for another round...?
Participants: Michael Connelly, Dale Coy, Brad Hanson, Gary
McCready, Jean-Francois Mezei, Chris Simon, Patrick Stair, Don
Vickers, Ray Whitmer.
The DECUServe Journal April, 1995 Page 19
Document Management Systems
Note 133.0, 3-Jul-1991
McCready: VMS document manager wanted.
-------------------------------------
A friend of mine is looking for a document management system
that will run under VMS and interface with Wordperfect.
He considered the product from Datability, but it seems that
they are no longer enhancing it, so that product is out.
Can All-in-1 work nicely as a shared document manager? Are
there any products that will work with it?
The application in mind is for a law firm (1500 possible
users).
Note 133.1, 3-Jul-1991
Coy: Worth a serious look
-------------------------
> Can All-in-1 work nicely as a shared document manager? Are
> there any products that will work with it?
Depends partly on the details of your requirements.
With addition of the "Shared File Cabinet" package from DEC ASSETS, it
does a pretty good job. Rumored to be even better in "future release".
With WordPerfect VMS, this would probably make a very nice fit with
your application. At least, it's worth a good look.
And _lots_ of other things work with ALL-IN-1.
Note 133.2, 18-Dec-1993
Stair: VMS & DOS document management?
-------------------------------------
It's been a while since this discussion was originally posted, so
maybe there are document management packages available in
addition to All-in-1.
Specifically, does anyone know of a package that will allow VMS
and Pathworks users to participate together? The VMS users are
all character-cell, as will be many of the PC users. WordPerfect
is the wordprocessing software used.
I've seen some nice PC-based items, and some overkill Motif- &
Windows-requiring packages, but nothing to help track our paper
flow allowing character-cell interfaces.
The DECUServe Journal April, 1995 Page 20
Document Management Systems
Note 133.3, 28-Dec-1993
Connelly: WordPerfect doc mgt
-----------------------------
WordPerfect Office would seem to be something you should look at.
There have been numerous threads on this in several conferences. Take
a look at 349.* in 3RD_PARTY_SOFTWARE
Note 133.4, 28-Dec-1993
Whitmer: WordPerfect offers no real VMS document management
-----------------------------------------------------------
The beta-tested VMS WordPerfect Document management package was never
released. WordPerfect has just acquired the leading PC document
management company. I am aware of no plans to port it to VMS, although
it is reportedly quite portable, if there were enough demand for VMS
solutions again (and it would help if it became easier to port apps to
VMS -- the old days with a large number of developers for a VMS port
are past -- new efforts require better OS interoperability).
WordPerfect Office allows things to be mailed/scheduled on and between
VMS and Pathworks. This is not document management, as I understand it.
You _can_ turn on long document names between VMS and Pathworks, with
WP 5.1. This is quite inferior to real document management, but if it
helps, it is there. If what ALL-IN-1 provides is adequate for you and
in a package you can use, by all means use it!
I wish there were a better solution on the market. I wish the market
justified the release of such a product. As it is, it is a rough
market to be in, often for either the user or the vendor.
Note 133.5, 5-Mar-1994
Hanson: Mosaic?
---------------
Are you looking for hypertext documents???
I'm not clear what type of document management you are looking for..
One that we have moved to is to deploy on-line documents in MOSAIC.
There is a character-cell viewer, Lynx, that allows our "VT" users to
access to documents. There are implementations on Windows and
Macintosh as well.. You aren't limited to text... it tends to support
all multi-media types... but not on the VT!!
Note 133.6, 6-Mar-1994
Stair: I should have said "document tracking"
---------------------------------------------
Perhaps "document tracking" would have a better phrase to use. We
need something that allows the users to pass around a document for
revision and review, and that allows a manager to see who has reviewed
The DECUServe Journal April, 1995 Page 21
Document Management Systems
a document and who is sitting on it.
It doesn't even have to allow fancy annotation of the document, since
we can do that in WordPerfect with redlining and comments. We just
need to be able to track it.
It has to be VT-compatible because approximately 75% of our users are
on VMS with VT terminals.
We do not have All-in-1 at our shop, which is the reason it is not
being considered. (We are not considering it because of price and
other considerations, for instance: we use WordPerfect Office on our
VAXes and PCs.)
Note 133.7, 6-Mar-1994
Mezei: Teamroute
----------------
You could look at Teamroute. It is a document tracking and
authorization and routine system. It works on multiple platforms
(MAILWORKS, ALL-IN-1 etc) and definitely works on VT terminals.
With all the changes, the product's name may have changed or a newer
product may exist, but Digital does have such a product available.
Note 133.8, 6-Mar-1994
Vickers: ForComment from Computer Automation was a nice solution
----------------------------------------------------------------
Access Technology (now part of Computer Automation) had a nice product
for doing the tracking and comment tracking. It was called ForComment
and was a sister product to ForWords.
You might wish to try to contact Computer Automation to determine
whether ForComment still is available and, if so, get an evaluation
copy if it seems close to your needs.
Note 133.9, 7-Mar-1994
Simon: I hate to correct Don, but...
------------------------------------
> -< ForComment from Computer Automation was a nice solution >-
Make that Computer Associates.
The DECUServe Journal April, 1995 Page 22
Document Management Systems
Note 133.10, 16-Mar-1994
Stair: Thanks!
--------------
Thanks, all.
It is still called TeamRoute (with versions for VMS and MS-Windows).
Also, we have one of Access Technology's products, and we'll give them
(whoever "them" are these days) a call.
All About DECamds
=================
Here is another fine example of a thread on DECUServe awakening to
new life years after the original discussion had died out. Things
began with a discussion of using DECamds for system management tasks
in '92 in the DEC_SOFTWARE conference. Much later, when VMScluster
licenses started to include DECamds with no extra charge, the
discussion revived.
Participants: Alan Bruns, Dale Coy, Larry Kilgallen, Fred Lobmeyer,
John Osudar, Stuart Renes, Rollo Ross, Pat Scopelliti, Chris Simon,
Harrison Spain, Dan Wing.
Note 557.0, 24-Nov-1992
Wing: DECamds
-------------
I just received some literature from DEC on the product DECamds, which
appears to take care of all sorts of system management headaches.
Apparently, this is a new product, and DEC is offering free 60-day trials.
Is this an old assets package, or several assets packages thrown together?
Anyone have any experiences with it? Comments?
It sounds like a *VERY* powerful product. Some of the points in the 4-page
letter I got:
Motif interface
Catch Runaway Processes
Change working set sizes
Exit an image or delete a process
Change process priorities
Purge a working set
Suspend/resume a process
Access Hung Systems - ? how?
"Fix Systems in a Flash"
Resolve Lock Contention
The DECUServe Journal April, 1995 Page 23
All About DECamds
Note 557.1, 24-Nov-1992
Simon: Formerly Captain
-----------------------
I don't have personal experience with it, but our LUG had a
presentation on it given by DEC, and several LUG members had tested the
software before it became DECamds. Back then it was apparently called
Captain, but I don't know if it was an ASSET, or if it was something
else.
Note 557.2, 24-Nov-1992
Coy: Sorta new
--------------
See INDUSTRY_NEWS 563.18 (February).
Also VMS 1805.1
I'd sure like to hear from someone with real experience...
Note 557.3, 24-Nov-1992
Kilgallen: Not an ASSET
-----------------------
AMDS is the "Captain" demonstrated at DECUS Symposia a while back.
To the best of my knowledge, it was built by VMS Engineering, but they
had a hard time convincing management to make it a product.
There is a site where I have had access to it, but I have not had the
sort of problem it is aimed at, so I have no experience.
Note 557.4, 25-Nov-1992
Lobmeyer: OK I have DECamds...
------------------------------
OK fire away. I have it. I use it to monitor four nodes of a Cluster.
It only works over and LAN. This can be an extended LAN.
It's setup is a driver on the nodes you are monitoring. The Console is
then run on a node that can display to an Xterminal or Workstation.
This is a DECwindows only product. The information displayed could not
practically be done to a VT style terminal.
DECamds is not cheap and is not marketed for that type of group. It is
used to try and keep a node up at all costs. Thus Digital is targeting
at that type of site.
The driver can be loaded and unloaded on the nodes that are being
watched. Thus upgrades don't require a reboot.
DECamds allows you to track down locks, adjust any of the of the
working set parameters of a current process on the fly. (ie. WSDEF,
The DECUServe Journal April, 1995 Page 24
All About DECamds
WSEXTENT). It also reports back to you if VMS is having any problems.
One problem that it reported to me was that the RESHASHTBL was to
This was causing the HASH table to have to many duplicates in a
particular level. IT was currently set to the max value for v5.4-3 so
for now I am out of luck.
My current display node is running a version of VMS that doesn't play
nicely with VMS v5.4-3 version of the product.
You can adjust the interval that the product checks your nodes at as
you add more nodes so as not to over run the VMS node that is the
Display node. Speaking with the developer the Display node can easily
handle monitoring a 100 nodes. I believe the Display node was one of
the original 3100s. The big limiting factor for the Display node is
memory. The more it has the faster and happier it performs.
If I get the Cluster up to VMS v5.5-2 to weekend I will be able to
report back to you on DECamds and try anything you would like...
Note 557.5, 25-Nov-1992
Wing: great! I have questions
------------------------------
> It's setup is a driver on the nodes you are monitoring. The Console is
> then run on a node that can display to an Xterminal or Workstation.
> This is a DECwindows only product. The information displayed could not
> practically be done to a VT style terminal.
"The Console" isn't OPA0:, so is it just something like a Motif screen that
lets you control that node?
> The driver can be loaded and unloaded on the nodes that are being
> watched. Thus upgrades don't require a reboot.
Upgrades to DECamds, right?
> DECamds allows you to track down locks, adjust any of the of the
> working set parameters of a current process on the fly. (ie. WSDEF,
> WSEXTENT). It also reports back to you if VMS is having any problems.
> One problem that it reported to me was that the RESHASHTBL was to
> small. This was causing the HASH table to have to many duplicates in a
> particular level. IT was currently set to the max value for v5.4-3 so
> for now I am out of luck.
How does DEC place DECamds with DECps, which is supposed to help you tune
your system? Same question for VCS, which (basically) looks at OPA0: on
all your nodes.
> Speaking with the developer the Display node can easily
> handle monitoring a 100 nodes. I believe the Display node was one of
> the original 3100s. The big limiting factor for the Display node is
> memory. The more it has the faster and happier it performs.
The DECUServe Journal April, 1995 Page 25
All About DECamds
Are you using a dedicated 3100 model XX to monitor all your nodes? Can you
connect to the console (OPA0:) of remote nodes using DECamds? Can you CTRL
-P a MicroVAX using DECamds (can you get to the >>> prompt w/o touching
the switch on the MicroVAX)?
Does DECamds record OPCOM events, page file fragmentation warnings, etc.,
like VCS?
-dan, dw...@uh01.colorado.edu
University Hospital, Denver
(we like to be up and running all the time!)
Note 557.6, 27-Nov-1992
Lobmeyer: More Questions?
-------------------------
> "The Console" isn't OPA0:, so is it just something like a Motif screen that
> lets you control that node?
The console can be anything running X. DECamds does not use OPA0: at
all.
> Upgrades to DECamds, right?
Yes specifically the driver which is reporting back to the console.
> How does DEC place DECamds with DECps, which is supposed to help you tune
> your system? Same question for VCS, which (basically) looks at OPA0: on
> all your nodes.
DECamds plays well with both of these. I am using both. It is more of
an additional tool. DECps reports things to you only. VCS reports
back to you what happened on OPA0 and can take action based on text
reporting to OPA0:
DECamds can modify a running VMS system or process. It also tells you
about problems that I don't believe DECps can describe to you because
it is monitoring the health of your system. DECamds will tell you that
your pagefile space is getting low before the message comes to the
console. Thus you can stop some process to free pagefile space before
your system hangs. If your system does hang DECamds can tell you while
it is hung who has all the pagefile space. It allows you to give
people a bigger chunk of memory than they are normally authorized to
use or trim them down on the Fly.
> Are you using a dedicated 3100 model XX to monitor all your nodes? Can you
> connect to the console (OPA0:) of remote nodes using DECamds? Can you CTRL
> -P a MicroVAX using DECamds (can you get to the >>> prompt w/o touching
> the switch on the MicroVAX)?
No I happen to use a 4200 with 16MB of memory. It could use a little
more. I am running DECamds, VCS, and PMDF with DECUS UUCP. It keeps
The DECUServe Journal April, 1995 Page 26
All About DECamds
this little guy pretty busy.
It allows you to crash a node. What happens to the MicroVAX after it
crashes? Does it return to the >>> prompt? If not there are programs
somewhere here on DECUServe that allow you to do remote reboots of
MVIIs.
> Does DECamds record OPCOM events, page file fragmentation warnings, etc.,
> like VCS?
No it is not meant to replace VCS but add to it. If VCS reports this
warning to you DECamds allows you to go in and stop processes so that
the rest of your system can continue without requiring to reboot
hopefully.
DECamds can be used Pro-actively or as a reaction to events from other
software like VCS. In a pro-active mode you have displays that tell you
how much of your memory, CPU and other resources are being chewed up.
If I get my VMScluster running at v5.5-2 I can fire up DECamds again
and I will cut some of the screens out to show you.
> (we like to be up and running all the time!)
Us too! But the issue becomes if your system hangs what does that cost
you as opposed to the price of this software? If you can reboot or
crash your system and take time to figure out what happened then this
software becomes an expensive toy. If keeping the system up at all
costs is the goal then you should look at DECamds.
Note 557.7, 30-Nov-1992
Renes: Our experiences...
-------------------------
We've used it here on a large, 100 vups VMScluster. It can't get you
out of any MWAIT/RWnnn jams, which were our biggest problems. It
does provide a lot of information about processes - almost too much
for a VAXstation 3100 to handle sometimes
Since the remote parts run at driver IPL (8?), you can track down
things on a node that is hung (at lower IPL). Unfortunately, this
hasn't been useful to us.
We found DECamds to be very pricey for what it provides.
Note 557.8, 23-Sep-1994
Spain: AMDS no longer out of reach!
-----------------------------------
I'm surprised that the AMDS thread has not picked up just a bit now
that you can install this kit with only a Vaxcluster license! :-)
I've got it running on my 4000-60 (VMS V6.0 or V6.1 is required) and it
The DECUServe Journal April, 1995 Page 27
All About DECamds
really is an impressive product!
You can install the kit from either the VAXVMS061 or AXPVMS061 CDROM.
Note 557.9, 24-Sep-1994
Osudar: had no idea
-------------------
I suspect that a lot of people are like us -- I had no idea what AMDS
was, until you posted your note and I went back and reread this thread
(which I ignored initially, as soon as I saw that it was going to cost
us $$$). Now I see that I should have installed AMDS when I upgraded
our cluster to VMS V6.1. I'll bet a lot of other people haven't
bothered with AMDS because they haven't taken the time to find out what
it's all about.
Note 557.10, 27-Sep-1994
Renes: Not quite ready on AXP
-----------------------------
Unfortunately, DEC hasn't gotten the display portion of AMDS to work
under AXP yet - those nodes can only run the data collector driver :-(
Note 557.11, 27-Sep-1994
Spain:
-------
True, but if you have a Vax somewhere running V6.1 you should be able to
set your display back to your local workstation and still monitor all
your V6.x VAX/AXP systems :-).
Note 557.12, 28-Sep-1994
Ross: Display only on VAX 5.5?
------------------------------
> True, but if you have a Vax somewhere running V6.1 you should be able to
> set your display back to your local workstation and still monitor all
> your V6.x VAX/AXP systems :-).
Does this really work? Not 30 minutes ago I decided not to try
installing AMDS after reading the following in the $$$README.FIRST file
in [AMDS061] of the AXP 6.1 CD:
| Only DECamds data collection is supported for OpenVMS AXP Version V6.1
| systems. In order to use the DECamds Console application, you must install
| the console application on an OpenVMS VAX Version 5.5 system with at least
| DECwindows Motif V1.1 installed.
I interpreted this as meaning that the display part of AMDS only
works on VMS 5.5. Was I unduly pessimistic? Does it actually work on VAXes
The DECUServe Journal April, 1995 Page 28
All About DECamds
running 6.1?
Note 557.13, 28-Sep-1994
Renes: ...
----------
> True, but if you have a Vax somewhere running V6.1 you should be able to
> set your display back to your local workstation and still monitor all
> your V6.x VAX/AXP systems :-).
yeah, but that defeats the purpose of having the display software
running on a "capable" workstation (i.e. Alpha)!
Is the porting of this software on DEC's Alpha port list?
Note 557.14, 28-Sep-1994
Osudar: 6.1 OK
--------------
The display part works fine on V6.1 -- I finally installed DECamds on
our production cluster Monday night, and I was playing with it a bit
yesterday. So far I've been impressed. (I also thought "Crash system"
as an option in the "Fix" menu was pretty interesting :-)
Note 557.15, 28-Sep-1994
Renes:
-------
You mean on VAX/VMS V6.1?
Note 557.16, 28-Sep-1994
Osudar: VAX only at 6.1
-----------------------
Yes, on VAX/VMS V6.1; there's no display component for the AXP.
Note 557.17, 29-Sep-1994
Spain: AMDS under fire still works
----------------------------------
We were using AMDS this morning on a system that was out of page file
space (we have all been there right?).
Rather than reboot or add an additional page file, we decided to see if
AMDS could get us out of the jam (CPU was 90% kernel mode).
We were impressed that AMDS continued to work even though we could not
get logged into the system and most processes on the system were busy
outswapping :-).
The DECUServe Journal April, 1995 Page 29
All About DECamds
Note 557.18, 29-Sep-1994
Scopelliti: Did it?
-------------------
> Rather than reboot or add an additional page file, we decided to see if
> AMDS could get us out of the jam (CPU was 90% kernel mode).
Note 557.19, 4-Oct-1994
Spain:
-------
Very nicely :-). We killed off the offending processes using AMDS and
the system came back to life :-).
Note 557.20, 2-Dec-1994
Bruns: AMDS Patch Kit Available on DSNLink
------------------------------------------
There's an ECO kit for AMDS available on DSNLink that fixes quite a
number of problems. In particular, the AMDS console now pays attention
to threshold classes (which allow you some latitude for setting
different trigger points for excessive (DIO, BIO, faulting, etc) based
on the type of systems you're monitoring).
Description: DECamds V6.1 ECO 1; OpenVMS VAX V5.5 - V6.1,
OpenVMS AXP V6.1 DECAMDS$01_061
At the time I found out about it (DSNLink_New pointed out a "Latest
ECOs" posting), the article wasn't available, so I had to request the
patch via e-mail rather than INVOKE it.
The patch has made this program very useful here. I've already used it
to solve a lock conflict.
Note 557.21, 5-Dec-1994
Renes: DECamds console for AXP (VMS 6.2)
----------------------------------------
I wonder if this patch kit was incorporated in the "pre-release" of
DECamds client display DEC-AXPVMS-AMDS-V0602--1.PCSI for VMS 6.2 that
was given to me by the CSC?
Note 557.22, 6-Dec-1994
Bruns: Dunno, but you can try this:
-----------------------------------
Don't know about that, but it should be easy to tell if my particular
problem is solved on your copy: Just mess around with the
The DECUServe Journal April, 1995 Page 30
All About DECamds
AMDS$THRESHOLD_DEFS.DAT file: set a really low value for, say, memory
use for a class that matches one or more of the systems you're
monitoring, and see if it generates an entry on the event log display.
The unpatched version ignored class settings entirely; I was getting
event log entries for BIO and DIO 100% of the time on my systems doing
a lot of X.25 traffic, regardless of what I put in
amds.Threshold.HiBIOrate.Class3 or amds.Threshold.HiBIOrate.Class4.
Note 557.23, 7-Dec-1994
Renes: !
--------
thanks, I'll try it.
X colormap
==========
Three short, year-old gems from the USER_INTERFACES-WINDOWS
conference (now that's a mouthful!) round out this belated April
issue. In the first, the topic concerns mysteries of colormaps in
X.
Participants: Charlie Byrne, Matt Holdrege, Erik Husby, Ed Kozam.
Note 206.0, 23-Feb-1994
Byrne: X Toolkit COlormap Entry Error
-------------------------------------
Every once-in-a-while I get
"X Toolkit warning: Cannot allocate colormap entry for "#e7e7e7e7e"
The "#e7e7" part changes.
I usually have some spiffy graphic in my root window, and then I get
this when I bring up Mosaic, RN, or some other x App.
I don't notice anything wrong otherwise.
Note 206.1, 23-Feb-1994
Husby: The colormap is a shared resource
----------------------------------------
The color map maps a pixel to a RGB value. If you have a screen
that can support 256 different colors simultaneously, then your
color map would have 256 entries. A pixel would range in value
from 0 to 255. In the color map would be the various RGB values
for all colors used by the different X applications.
The DECUServe Journal April, 1995 Page 31
X colormap
When an application wants to display a particular color, it
allocates a color map entry. Then it sets the RGB value for that
color in the color map.
Applications that display GIF images typically use a lot of
colors. Motif needs at least 16 colors. 256 colors can be used
up real fast. When that happens, an attempt to allocate a new
color map entry will fail.
The value #6e6e6e6e (or what ever you used) is the RGB value to
be placed in the color map.
Note 206.2, 24-Feb-1994
Holdrege: how to deallocate?
----------------------------
When you kill a GIF, does it deallocate the color map entry?
Note 206.3, 25-Feb-1994
Kozam: Root window also uses color map entries
----------------------------------------------
Remember also that the root window also uses color map entries, so if
you display a pretty GIF as your root, you may not have any entries
left for your applications.
Even after the process that put up the root window disappears, the
color map entries continue to be used and don't get released until you
display a new root.
Note 206.4, 25-Feb-1994
Holdrege: xsetroot
------------------
I've noticed that xsetroot will deallocate color map entries, but I was
wondering how to do it for pop-up windows.
Forcing a Window to the Top
===========================
The next item from USER_INTERFACES-WINDOWS has to do with managing
windows in Motif. More specifically, it has to do with moving a
particular window to the top (foreground, for those of you who, like
us, think of windows in Macintosh terms).
Participants: Mike DePriest, John Vottero, Ray Whitmer.
The DECUServe Journal April, 1995 Page 32
Forcing a Window to the Top
Note 208.0, 9-Mar-1994
DePriest: How can I get a window on top?
----------------------------------------
I'm a fairly new Motif programmer, and I checked the ForWords index to
no avail trying to find pointers for my problem,
How can you force a window to go to the top of the stacking order?
I'm running X11R5 and Motif. I thought I could do this with an Xlib
call (XRaiseWindow) but apparently Motif is intercepting the request and
sending it to the bit bucket.
The kicker to this is that the window I want to raise belongs to another
client currently running. I want to programatically force any client I choose
to the top of the display's stacking order.
Any hints or code fragments will be most appreciated.
Note 208.1, 10-Mar-1994
Whitmer: Questions.
-------------------
I also have not done enough Motif to know the answer to this. A couple
of, probably stupid, questions occur to me:
I assume you do _not_ have your focus model set such that it
automatically raises whatever window contains the pointer device?
How are you getting the ID of the window you wish to lower?
In the past, when I have needed to get the ID of a DECTerm window to
attach another window to it, I used the technique: Use the escape
sequence to modify the title/icon name, and then search through the
window heirarchy to find the window with the name I had set. But this
always failed if the DECTerm was run locally on an XTerminal (even
though the escape sequence still properly modified the title/icon
name), leading me to believe that it is more difficult to get proper
window IDs of windows not local to the current client. Have you tried
changing other features of the window so that you know you are able to
get _any other_ modification to work on the window using the ID you
currently have for it? Obviously, things like this _can_ be done
because remote window managers do it all the time, but I never figured
out how.
Note 208.2, 10-Mar-1994
DePriest: Some progress
-----------------------
> I assume you do _not_ have your focus model set such that it
> automatically raises whatever window contains the pointer device?
That is correct - focus follows the pointer but the windows retain their
The DECUServe Journal April, 1995 Page 33
Forcing a Window to the Top
stacking order unless explicitly given a raise command (by clicking on the
client decorations or a f.raise command.
> How are you getting the ID of the window you wish to lower?
Pretty much the same way you described:
1. Know the name of the client. For my testing I'm using known
clients like xload, xclock, and verifying their window IDs
with xwininfo.
2. Use XQueryTree() on the root window to get the window hierarchy.
3. Walk the tree to find the window with the name I'm after
(using XGetWMName() and XTextPropertyToStringList() to get
the name).
I made some progress after posting .0, using XMapRaised() instead of
XRaiseWindow() but the results are flaky.
The documentation is _so_ helpful ...
Note 208.3, 10-Mar-1994
Vottero: You CAN do it.
-----------------------
I have been able to do this using XRaiseWindow, or XMapWindow if the
window is currently iconified. I think the trick is to get the right
window ID. It's not always the one with the name associated with it.
Try using window ID's which are higher (or lower) in the window tree.
Note 208.4, 10-Mar-1994
DePriest: Solved, sort of.
--------------------------
I finally got it to work using XMapRaised() - under these conditions:
- I used a recursive routine to walk the child list created
by XQueryTree()
- I tested _each_ id for the correct name and if found
executed the XMapRaised() _right then_ instead of
returning the window ID for later raising
- I did _not_ XMapRaised() any more windows in that branch
of the tree.
What I thought would be a good idea, that was to only seek out the
window ID by name, then to the XMapRaised() once at the end, actually
did not work. Nor could I hard-code the ID of a window I wanted raised
and get it on top.
The DECUServe Journal April, 1995 Page 34
Forcing a Window to the Top
I'm willing to write off part of my problems as being attributable to
my marginal C language skills, but I can't understand why the hard-coded
window ID never got raised. Oh well, live and learn.
Thanks for your suggestions and moral support!
Note 208.5, 17-Mar-1994
DePriest: XFlushed with pride
-----------------------------
I got some additional help on this problem and finally resolved it!
I wasn't doing an XFlush() after the XMapRaised() in the cases where I
gave the window ID explicitly instead of using XQueryTree(). Adding
the XFlush() did the trick. It turns out that XQueryTree causes the
server to do its own XFlush-equivalent, so it was just my good luck
that my code was working at all.
EVE DECwindows Configs
======================
The last article of the issue deals with some (non-obvious?) tricks
for configuring EVE under DECwindows.
Participants: Duncan Brown, Brian Tillman.
Note 205.0, 17-Feb-1994
Brown: How can I change DW Eve attributes?
------------------------------------------
Here's an interesting question, that I can't seem to find the answer
for:
When using "DECwindows EVE" (specifically, when using it as a mail
editor, but I believe the solution to this will be more general than
that), I don't like the way it comes up, in terms of font size and
window size. How can I fix this?
Is it something I need to put in my initialization file? Is it a
resource file I need to create and edit somewhere? I can't seem to
track down a clue anywhere in the docs or HELP...
Note 205.1, 17-Feb-1994
Tillman: Try this
-----------------
You want to create the file DECW$USER_DEFAULTS:TPU.DAT. In that file,
place commands like this:
The DECUServe Journal April, 1995 Page 35
EVE DECwindows Configs
Tpu.Tpu$MainWindow.Rows: 36
Tpu.Tpu$MainWindow.X: 347
Tpu.Tpu$MainWindow.Y: 67
*fontSetSelect: 0
You can also add things like
*littleFontSetName: somefontspec
*bigFontSetName: somefontspec
Note 205.2, 21-Feb-1994
Brown: Thank you sir, may I have another?
-----------------------------------------
OK, the TPU.DAT file seems to do the trick. Where the heck did you
find out this information?! I have used Bookreader's SEARCH capability
(now that I know specifically what I'm looking for: TPU.DAT), and can't
find it in any of the obvious or even non-obvious places?
Next question: what are the big and little fonts used for? That may be
part of the answer to my next question: How do I get the font to be
condensed when I go to 132 columns? (With DECTerms, there's a specific
setting for that in the options, how do I do that in my TPU.DAT file?)
I love undocumented features...
Note 205.3, 21-Feb-1994
Tillman: I'm not as good as I appear
------------------------------------
Well, I know that all DECwindows applications that utilize resource
files have default resource files in the DECW$SYSTEM_DEFAULTS
directories. I used DIR to look there and saw TPU.DAT. I cut the
contents of it to the note I posted. I also used DSNlink to search the
DECW-MOTIF database for "eve" and "font". That produces one article
entitled "How To Change The Fonts For DECwindows TPU/EVE". That's
where I got the other information and it's the sum total of what I know
about it.
Note 205.4, 21-Feb-1994
Brown: Sneaky hiding places for .DAT files
------------------------------------------
Aha! My big mistake! I had looked in
SYS$COMMON:[DECW$DEFAULTS.SYSTEM] and .USER], forgetting that
SYS$COMMON:[SYSLIB] (where TPU.DAT resides) was also on the full search
path of DEW$SYSTEM_DEFAULTS. Sheesh. I'm too smart for my own good!
OK, at least I now know all the same clues you do. If I gain any great
insights into the problem, I'll be sure to share them here...
The DECUServe Journal April, 1995 Page 36
About the DECUServe Journal
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.
Where To Find the DECUServe Journal
Issues are posted to the ARTICLES conference on DECUServe and DCS,
as well as to the DECUS World-Wide Web and Gopher server
(http://www.decus.org/decus/pubs/pubs-index.html) and the
vmsnet.decus.journal and comp.org.decus newsgroups. E-mail
subscriptions can be had by sending the command "subscribe
DECUServe-Journal" to mail...@decus.org.
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
infor...@decuserve.decus.org.
Contact Information
===================
The editors of the DECUServe Journal are Brian and Sherrie McMahon.
They can be reached as follows:
mcma...@decuserve.decus.org or mcm...@ac.grin.edu
mcma...@decuserve.decus.org or grif...@ac.grin.edu
+1 515 269 4901 (normal office hours, U.S. Central time)
+1 515 269 4936 FAX
Assistant Editor Curtis Reid is reachable as
re...@decuserve.decus.org.