(last updated: 29-Dec-2016)
The current configuration is slightly changed from the original 'as acquired' condition. I haven't changed any of the modules/cards or their order in the backplane, but I have replaced all 5 of the original RF71 (400MB) hard disks. I now have 2 RF73 (2GB) + 4 RF72 (1GB) hard disks installed. One of the RF72s is in the main BA213 cabinet, the other three are in the R215F expansion cabinet. The sixth is just 'freely' mounted in the spare disk position in the R215F expansion cabinet. I need proper mounting brackets.
| BA213 | A | B | C | D | |
|---|---|---|---|---|---|
| Q22/CD | 1 | M7636-AA : KN210 I/O w/ DSSI (MSCP), SLU, Ethernet | |||
| 2 | M7635-AB : KN210 : RISC 3000 CPU | ||||
| 3 | Clearpoint Research CE-061 : 16MB RAM | ||||
| 4 | M7622-AT : MS650 : 16MB RAM | ||||
| 5 | M7622-AT : MS650 : 16MB RAM | ||||
| 6 | M7740-PA : RQC25 (KLESI-SA) : LESI bus adapter (MSCP) | (empty) | |||
| 7 | M7559-OO : TQK70 : TK70 cartridge tape controller | (empty) | |||
| 8 | (empty) | (empty) | |||
| 9 | (empty) | (empty) | |||
| 10 | (empty) | (empty) | |||
| 11 | (empty) | (empty) | |||
| 12 | M9060-YA : Minimum load board | ||||
20-Feb-2006
Picked up this system along with several uVAXen and PDP-11's in Longmont, Colorado.
08-Mar-2006
The box suffered some damage during shipment. The grill on on the lower rear of the unit, on the right-hand side, was crushed and the plastic ribs were cracked and fractured along a vertical line. This is a tough material, so I was lucky. Nothing was broken loose and lost. Using a screwdriver as a pry bar, and bending each grill rib laterally to shorten it, I was able to reseat each of the ribs, mating the fracture almost perfectly. I don't know what to use to bond the crack, but I don't think I need to. The material is so strong that, now that everything is back in its proper position, I expect it to be just fine as is.
As usual, I vacuumed the exterior, and 'baby wiped' it to clean the dust from the unit. On the rear is a label that reads: "L1103 / Property of Chrysler Capital Corp."
Now if I can only figure out how to get inside! I do have the plastic 'vending machine' key the locks the sliding front panel. There must be some relationship with that front panel and opening the cabinet. Found it! With the control panel cover completely down, there is a slider/drawer left of the power switch that can be partially pulled outward with thumb and forefinger. This allows the entire front of the chassis to lift and be removed.
I then managed to lift (ugh!) the unit to my work surface and laid it on it's back to give more convenient access to the card cage, and other internals. I removed the rolling pedestal from the unit by releasing the two triangular 'cams' in the base.
Once inside I vacuumed the interior, and removed the control panel's captive screws giving access to the hard drive compartment in the upper portion of the cabinet. Again, I vacuumed everything, and removed one drive for identification. I removed all boards from the card cage and vacuumed the backplane. Everything was then re-assembled.
09-Sep-2006
I wanted to try powering up and booting the system today, but when I went to connect a modular AC cord I find an unusual plug socket. It's the standard modular AC plug except that it has an extra extrusion that keys the cord, in addition to the already key-able shape. I wonder what are the special properties of this. Maybe just high amperage. I'll google to see what I can find.
Well I found that it is a normal power cord, but for high amperage uses. I notched out a 18-gauge cable I had and plugged it in.
Powering up the 215F, I heard a little squealing at first, maybe from the cooling fan, maybe from a drive. After a few seconds, it stopped. I then powered the Ds5400 again, and got the same ouput, without the language query.
KN210-A V2.0 Performing normal system tests. 51..50..49..48..47..46..45..44..43..42..41..40..39..38..37..36..35.. 34..33..32..31..30..29..28..27..26..25..24..23..22..21..20..19..18.. 17..16..15..14..13..12..11..10..09..08..07..06..05..04..03.. Tests completed. >>>
It's alive! It's alive! It's alive!
10-Sep-2006
With all of the drives WRITE-PROTECTED, eventually I was able to boot with this command:
>> boot -f rf(0,0,0)vmunix
After a lengthy boot process, I get the # prompt, so I'm in some kind of single-user mode. There is only one mount (/dev/ra0a /) even though there is much more in the /etc/fstab file.
Just for interest's sake, I checked what the system thought the date was. It comes up with:
Thu Sep 12 17:16:51 MDT 1996
Almost 10 years ago!
17-Oct-2006
After a number of false starts and adjustments to my bench setup, I was able to finally get all five disks image to an external Debian Linux box using NFS.
Disks 0, 1, 2, and 4 imaged without trouble, but disk 3 has 8 hard sector errors, and I had to make adjustments to the dd command (and the partition table) to let me step past those bad sectors and get the rest of the image. In the end, I had several chunks of the disk, and I cat'd them together with null-filled blocks as infill at the locations where I wasn't able to read particular sectors.
As a precaution, I also tar'd the drive's file systems and discovered only one file that contains a single bad sector. Turns out that the system has another copy of the same file so I injected that into the tarfile and should be able to successfully reconstruct all of the system's disks and filesystems in the future.
16-Mar-2007
Today I won an eBay auction from eBay seller 'uechi' for 3 RF72 drives at $20 each. These RF72 DSSI drives each have a 1GB capacity, so with them, I should have a system that is able to really be useful as some kind of Internet server. It turns out that the shipping charge added to the eBay checkout was too high, and the seller has offered to include a fourth drive for the same money paid! And, even better, he thinks that he has a #5 Drive ID plug that will allow me to put a third (sixth system-wide) drive in the R215F expansion chassis!
13-Apr-2007
An additional 2 RF72 drives arrived yesterday from eBay seller 'bealingtongroup2'. They were very well packaged, along with a KA650 uVAX III CPU and memory boards.
Unfortunately, the brackets attached to these 2 drives (which was an important reason for buying them) are not the correct type for the BA213/R215F cabinets. I took a chance, so that's okay. I need to find one more set of mounting brackets/slides and I'll have all I need. Maybe I'll come across them someday. I have searched the web and it seems the part is a "RA7X-BA", which is the rail set for a BA213/215 chassis.
With hope that I would be successful, I wanted to at least temporarily mount (without the bracket) one of these drives into the R215F cabinet's empty drive compartment and see if it would be recognized, simply by installing the #5 drive ID plug from the March 16 auction. The DSSI cable and the 10pin control cable are already installed in the cabinet, so it was a simple matter to connect up the new drive. First I had to demount it from the wrong bracket, but it wasn't difficult at all.
I powered the R215F cabinet alone and listened as the drive spun up and chattered awake. It is somewhat louder than the RF71's but all seems normal. I then powered up the Ds5400 and got to the maintenance prompt (>>>). I entered the DUP command to talk to the drive:
>>> SET HOST/DUP/DSSI 5 Starting DUP Server...DSSI Node 5 (R2AIJC) Copyright (C) 1990 Digital Equipment Corporation PRVMON V1.0 D 22-NOV-1991 14:57:25 DKCOPY V1.0 D 22-NOV-1991 14:57:25 DRVEXR V2.0 D 22-NOV-1991 14:57:25 DRVTST V2.0 D 22-NOV-1991 14:57:25 HISTRY V1.1 D 22-NOV-1991 14:57:25 DIRECT V1.0 D 22-NOV-1991 14:57:25 ERASE V2.0 D 22-NOV-1991 14:57:25 VERIFY V1.0 D 22-NOV-1991 14:57:25 DKUTIL V1.0 D 22-NOV-1991 14:57:25 PARAMS V2.0 D 22-NOV-1991 14:57:25 Total of 10 programs.
Task Name? _
So, looks like the drive works! And the #5 plug works!
18-Apr-2007
After checking out all of my new RF72 drives, and finding one bad, I still didn't have a good handle on why when I booted the original installation that it couldn't talk to the Drive #5. I had the plug installed, the DUP/DSSI maintenance routines happily talked to the drive, creating the special files in /dev for it, and, after I created a partition (chpt) and new file system (newfs) on RF72 Drive #5 (as #3 at that time), and copying the boot files over to it, and I could even actually start the boot from it, all as Drive #5. But when the kernel loaded, it would complain that Drive #5 wasn't configured. Okay, okay, after all of this I finally realized that it was the kernel image that wasn't built to accomodate Drive #5. What I needed was a new kernel.
After I looked at some of my ULTRIX documentation, I discovered a section that talks about rebuilding the kernel (Volume 2, Configuration File Maintenance, Chapter 2), for reasons that include adding hardware devices. So, I set out to rebuild a kernel (first time for that!). The instructions are very explicit and included all the information required. Also, I benefited from just needing to modify what already was here, rather than starting from scratch. So I made a copy of the original configuration file, /usr/sys/mips/CONVEXCS, saving it as /usr/sys/mips/CHRYSLER. I then edited it to include a Drive #5. Then I followed the procedure to 'config', 'make clean', 'make depend', and 'make' the new kernel. Worked!
20-May-2007
I've been trying to get the system to connect to the Internet and have had to discover all kinds of things about setting up the IP addresses and so on.
I modified /etc/rc.local to set my address to 192.168.1.19, and hopefully set the static route to my router/gateway, 192.168.1.1. I've not successfully got the gateway part working quite yet.
A little scouting (mabye a lot!) on the web turned up some information that suggested that nameservers are given in /etc/resolv.conf.
So I guess my changes to resolv.conf and rc.local are at least getting the router/gateway and DNS stuff configured enough to do real work. Just for fun, I ftp'd to ftp.netscape.com, navigated their subdirs and downloaded the windows version of Netscape 4.8. Worked fine.
09-Aug-2007
Tonight, while waiting for other systems to execute various processes, I thought I might as well go ahead and swap out the original RF71 drives for the RF72 drives that I bought via ebay.
First, I pulled from the BA213 cabinet, the left-most drive which corrosponds to the Drive #2 plug. This I tagged "Longmont BA213 #2" and removed its mounting rails. In it's place I installed the "bealingtongroup2 #3" drive that I used to store the entire combined original 'as acquired' configuration. This drive was delivered with the appropriate BA213 style rails so I just slide it into place. I continued, putting drives in place as follows:
Longmont RF71 Drive #2 (left-most) => bealingtongroup2 RF72 #3 Longmont RF71 Drive #1 (middle) => bealingtongroup2 RF72 #2 Longmont RF71 Drive #0 (right-most) => bealingtongroup2 RF72 #1
Oops! Something's wrong! Before I swapped drives in the R215F cabinet I wanted to try these out. I attached a terminal and moved the DSSI terminator from the R215F to the BA213. I powered up. The countdown goes just fine, the drives chatter away, but then they fault, one or two at a time, and then they reset and then another will fault. This appears to go on indefinitely. It's like there is a conflict, or maybe the bus isn't really terminated, though I see the terminator 'lit' so it's getting powered, at least.
I think what I should do is disconnect two of the drives, and try just one. Then add another, as I can. This might help to sort out if there is a conflict among the drives. They each have their own Drive ID plug, but perhaps one is hard-programmed to a certain address.
I pulled the cover plate from the drives, and disconnected the 50-pin DSSI cable from drives 1 and 2, leaving 0 connected. I booted the machine. No problems. The system seems fine. I went to MAINT mode (>>>) and did a SET HOST/DUP/DSSI 0 and was able to talk to the drive without any difficulty. It stayed ready, without any faults.
11-Aug-2007
After a couple of reboots due to fsck I was up and running, with six(!) RF72 drives for a total of 6GB! Wahoo! I had to remember to boot with my rebuilt kernel but once I did that, all six drives are recognized
17-Aug-2007
I started the installation of v4.5 from scratch, but it didn't work! There was a complaint about not being able to mount /dev/ra1c to /rootmount, again. I finally figured it out.
Turns out the I really should be using partition 'a', not 'c', for my media_p variable. I halted and restarted the whole process from scratch, as if I were really doing this with only the summarized instructions, and with the change to media_p, it works! So, here is the corrected and simplified summary of steps:
1- boot the system, using pre-existing OS 2- image the CDROM "BASE INSTALL" to a spare drive (say, /dev/ra1c):
# mount aero-6:/mnt/windows/DropZone/Ds5400 /mnt # dd if=/mnt/ultrix_AG-MM16M-RE.img of=/dev/rra1c obs=5123- reboot, using the newly imaged drive as boot volume:
# shutdown -h now >> boot -f rf(1,1)vmunix4- at the installation options menu, select 3) System Management
Enter your choice: 3 #5- edit the install.1 script (that's a 'dot one', not 'dot ell'), using the device used as the CDROM image for BTD=xxxc:
# ed install.1 173d i BTD=ra1c . 193d i ra* ) . w q #6- restart the installation process:
# exit7- proceed, using BASIC or ADVANCED, as desired.
I went ahead and began the ADVANCED intallation process in earnest, using the above steps. I selected install ULTRIX Worksystem Software, with the X11/DECwindows option. When prompted for optional subsets, I selected ALL MANDATORY AND OPTIONAL. I also selected all of the Configuration File Kernel Options. The whole process took about 1 hour 30 minutes, and required about 500MB of disk. This now gives me a baseline for any future, more customized installation, giving me some guidance on time and space.
18-Aug-2007
All day I've been struggling to get the correct configuration on this new v4.5 installation so that my internet connectivity works properly.
At last I hit on a site that discussed setting up name services on various OSs. It had one little hint, that sort of rang a bell from my earlier reading of the documentation. It pointed out that the resolver also uses a file /etc/svc.conf that specifies where to look for name information. Most items are set to 'local' which means they must be named in the local config files (/etc/hosts) to resolve. If instead the items is set to 'local,bind' then it will be searched locally first, but then nameservers will be consulted if not found locally. This is it! When I changed the line for hosts, it instantly started working:
/etc/svc.conf: ... hosts=local,bind networks=local,bind ...
So, to summarize, four files need to be correctly configured (as shown above) to get the system on the internet:
- /etc/rc.local - /etc/hosts - /etc/resolv.conf - /etc/svc.conf
By the way, I tried the Ultrix tool 'netsetup' hoping it would set all this stuff up. It does most of it, but it didn't configure the hosts= line in svc.conf, and that's what was keeping me back. I just now discovered another tool, svcsetup, that might be what I need. Ah well, now I'm smarter.
19-Aug-2007
With this system now running Ultrix v4.5, and properly configured for the Internet, I wanted to get Lynx installed so that I could browse the web, and download files from ftp sites. Based on the success I had with NetBSD-1.5.3 on the VAXstation 3100, I knew that I would need a gzip first, since Ultrix doesn't include one. Most sources are distributed as gzip'd tarfiles, so that was the very first order of business.
I downloaded gzip-1.2.4a.tar from ftp.gnu.org and did a make on it. No problems, but the 'make intall' created a small problem. The default prefix for installation paths is /usr/local, and Ultrix doesn't have that path by default. At first I figured I could let that path be used, and change the user's .profile skeleton to include this path. I tried this, and for some reason the .profile changes weren't being picked up. Admittedly, I don't know much about how this skeleton stuff works, so I decided to take another route. I would just set the --prefix= setting in the ./configure. Something like this:
> ./configure --prefix=/usr
This worked just fine. I rebuilt gzip, and when I su'd to root and did 'make install' it went into Ultrix's normal directories, for the executables and manpages. Very nice.
With this figured out, I downloaded gnu's less and built it. It works like a charm, though I did purposely chose an old version (3.20) to kind of match up with the timeframe of this version of Ultrix. In fact, though I didn't mention this, I had to do the same with gzip. The current gzip wouldn't build, complaining about some kind of redefine in sh. So that's why I'm using such an old gzip.
Okay, now that I have a little experience I went for broke and tried Lynx. Again, I went for the 2.8.1r2 version, since the current 2.28.6 wouldn't configure. But the older one did and I did a build on it. There were some complaints from the cc compiler about not being able to optimize certain files, because the '-Olimit 600' was too low. I stopped the make and manually edited the makefile to include '-Olimit 6000' instead. Seems to have worked.
26-Aug-2007
Over the past week, I've tried to build and install various pieces of software, updated tools for the eventual build of some internet related stuff, like, Apache and Exim, etc. I've had pretty poor results. Many sources won't build.
At first I tried to build with the normal DEC C compiler, but it failed. I've been building mostly as user 'jared' and then su'ing to 'root' to do the installs. While trying to build this as jared, it would fail on insn-attrtab.c with 'Out fo file space' errors. Yesterday, finally tried it as 'root' and I was finally successful building gcc-2.8.1, and in the end it was much smoother than my prior experiences. Here are the details:
gcc-2.8.1: - Build Tools: -- gcc-2.5.8 -- (g)make 3.76.1 -- gzip-1.2.4 -- less-382 --- User -- root - Build Commands: -- setenv PATH /usr/local/bin:$PATH -- ./configure --prefix=/usr --exec-prefix=/usr/local -- make bootstrap -- make install
My current set of add-ons are:
- gcc-2.5.8 (1994) - gcc-2.8.1 (1998) - (g)make-3.76.1 (1997) - gzip-1.2.4 (1993) - less-382 (2002)
01-May-2008
In the past two days I have received two RF73 2GB drives from two different sellers. The sellers were "excess_computer", which packed the drive very, very well, and "digitalhw" with packed the drive very poorly with only a single layer of small bubble-wrap and then a few chunks of styrofoam (the drive was loosely floating around inside the carton because of this poor packing!).
Today, I wanted to test them out. Using the same procedure from last August, I checked each drive first with the DSSI DUP diagnostic tools built into the console.
After about 3 hours 30 minutes, I'm happy to report that it appears to have successfully imaged the whole disk, in spite of the poor packaging!
While the second disk was being imaged, I ran 'strings' on the first drive's image. It seems to have a bunch of VMS related stuff, maybe for an INGRES database system. I found copyrights from "Hughes" with X.25 communications stuff. It might have been from some kind of communications hub or controller, maybe a commercial one. Not smart enought to tell for sure.
Again, with the second drive, I got a complete image, no trouble. This is great! I now can put a total of 5GB into the R213 cabinet, and using another 3GB in the R215F, for a total of 8GB! Quite a bit more storage than the 2GB that I had at first, broken into 400MB RF71 chunks! This will make things much easier for volume and directory organization. I'll probably move the root directory / to the first 2GB drive, and /USR to the second 2GB drive. To be done another day.
Just for kicks, before winding things up for the day, I ran strings on the "excess_computer" drive and find that it is all zero'd out, except for just a very few non-printable bytes in a few sectors scattered in a few widely separated spots on the disk. So, there's no point in saving this image.
02-May-2008
First, I had to add an RF73 drive type entry into the /etc/disktab file. I find it interesting that DEC still hadn't added the RF73 drive type into ULTRIX v4.5 which was released in 1995. As I understand it, the RF73 drive was available in 1992, if not before, so it should be in the later ULTRIX, you would think. Anyway, here is my added entry for the RF73:
#### Added by JRB 2008-05-02 ####
#
# The values for the partition table below are based on the cam_data.c
# entry for this device (RZ73)
#
rf73|RF73|DEC RF73 Winchester:\
:ty=winchester:ns#71:nt#21:nc#2621:\
:pa#32768:ba#8192:fa#1024:\
:pb#131072:bb#4096:fb#1024:\
:pc#3907911:bc#8192:fc#1024:\
:pd#844800:bd#8192:fd#1024:\
:pe#844800:be#8192:fe#1024:\
:pf#2054471:bf#8192:ff#1024:\
:pg#819200:bg#8192:fg#1024:\
:ph#2924871:bh#8192:fh#1024:
Copyright © 2006-2025 Jared Blaser. All rights reserved.