I’ve had a VAXstation 3100 M38 sitting in my flat for years, but it doesn’t quite work. I’ll write up the repair process elsewhere, but as part of troubleshooting I decided to see what’s on some of the ROMS.
There are two M27C1024 chips installed on the main board. These are STMicroelectronics chips, both 1Mbit EPROMs. Reading them on Minipro produced an interesting error:
Maybe this is a DEC variant. Throwing caution to the wind, I re-ran Minipro with the -y switch to ignore the error. This produced a 131,072 byte file for each chip, which is correct for a 1Mbit IC because 131072 bytes is 1 megabit.
The contents of the files were all over the place – nothing stood out as text except for this:
) ns Fn s uie mae) 1Desc 9)taan Dts (hwz) 10Nerlds 3Enis 1)or ) glh rishri) 1 Ptu s 5Es o 3)uo ) anis 1 Sns ) anisCadi) 1 Vam
If you’ve used a VAX before, you might recognise that this looks a little like the language selection menu:
0) Dansk 8) Français (Suisse Romande)
1) Deutsch 9) Italiano
2) Deutsch (Schweiz) 10) Nederlands
3) English 11) Norsk
4) English (British/Irish) 12) Português
5) Español 13) Suomi
6) Français 14) Svenska
7) Français (Canadien) 15) Vlaams
Since this is the second ROM of a pair, let’s look at what the first contains:
0Dak 8)raai(SssRond ) uth Ilio 2)euchScei ) dean ) glh 1 Nsk 4Enis(Bti/Ish 2)orgu ) pal 1 Smi 6Fr a 4)veka 7Fr a (naen 5)las
If we take two bytes from one ROM, and two bytes from the second ROM, we can take a guess as to how the image is laid out. The characters “En” are from the the second ROM, “gl” from the first ROM, “is” from the second ROM, “h ” from the first ROM, and so on.
If it sounds like madness, but the datasheet makes it clear why:
It is ideally suited for microprocessor systems requiring large data or program storage and is organized as 65,536 words of 16 bits.
This handy bit of Python code will reassemble the data by reading two bytes from one file, two bytes from the second and so on:
# Read the content of two files and output an interleaved file
from pathlib import Path
INTERLEAVE = 2
FILE_A = './rom1.bin'
FILE_B = './rom2.bin'
if Path(FILE_A).stat().st_size != Path(FILE_B).stat().st_size:
fileA = open(FILE_A, 'rb')
fileB = open(FILE_B, 'rb')
data = b""
for n in range(Path(FILE_A).stat().st_size):
data = data + fileA.read(2)
data = data + fileB.read(2)
f = open('./out.bin', 'w+b')
How do we know it’s worked? Simple – the content of the binary file is exactly the same as the original KA42B firmware!
Having recently changed jobs and become self-employed, I have a lot more spare time. What better way to spend it than getting an OpenVMS machines running on a Raspberry Pi.
These instructions are necessarily quick and make assumptions that you know what you’re doing. They’re also heavily influenced by this YouTube video, and I’m very grateful for the simple explanation of Linux bridge and tap interfaces.
I started by copying the 32-bit version of Ubuntu 18.04LTS for the Raspberry Pi 3 on to an SD card. Pop that in your Pi and boot up with an Ethernet cable connected and a DHCP server available. There’s no need to attach a keyboard or monitor – just ssh in to the server with username and password, and it’ll prompt you to change your password on first login.
Cloning the simh repository and installing dependencies is straightforward, and takes about four minutes in my Pi 3B+:
apt install make libsdl2-dev libpng-dev libpcap-dev libvdeplug-dev bridge-utils
git clone https://github.com/simh/simh.git
make -j4 vax8600
There are other targets you can build for, but we’re going to use the VAX 8600.
MicroVAX I (KA610)
MicroVAX II (KA630)
VAX 8200 (KA820)
VAX architecture build targets for SimH
Whilst it’s compiling, create a directory to hold your SimH installation, and copy your OpenVMS 7.3 (VAX) ISO image over to it. I’ve used ~/simh, but you can use whatever you like as long as you’re consistent.
Create the file vax8600.ini with the following content. I’ve annotated each line so you can see what it does:
; Set the memory size to 512 megabytes
set cpu 512M
; Use a TCP socket for the console
set console telnet=12344
; Set the CPU idle detection method to VMS to improve performance when OpenVMS isn't doing anything
set cpu idle=vms
; Set the CPU to a model 8650 (https://en.wikipedia.org/wiki/VAX_8000)
set cpu model=8650
; Configure an 1.5 gigabyte RA92 disk (http://bitsavers.informatik.uni-stuttgart.de/pdf/dec/disc/ra90/EK-ORA90-SV-003_RA90_RA92_Service_Jun90.pdf) on disk interface RQ0
set rq0 ra92
; Attach a disk image to interface RQ0 - SimH will create this on boot
attach RQ0 rq0-ra92.dsk
; Configure a CD-ROM drive (RRD40) on disk interface RQ3
set rq3 cdrom
; Attach the installation disk ISO image to interface RQ3
attach RQ3 -r openvms73.iso
; Disable the RP Massbus controller
set rp disable
; Disable the RL11 cartridge disk controller
set rl disable
; Disable the RK611 cartridge disk controller
set hk disable
; Disable the RX211 floppy disk controller
set ry disable
; Disable the ??
set ru disable
; Disable the TS11 magnetic tape controller
set ts disable
; Disable the TUK50 magnetic tape controller
set tq disable
; Disable the DZ11 8-line terminal multiplexer
set dz disable
; Disable the LP11 line printer
set lpt disable
; Enable the Ethernet controller
set xu enable
; Set the MAC address to use for the Ethernet controller
set xu mac=08-00-2b-00-00-0a
; Attach the Ethernet controller to a TAP interface 'vaxa'
attach xu tap:tapvax
When your VAX binary has compiled, copy it from simh/BIN/vax in to your VAX directory.
Setting up networking
We’re ready to boot our emulator, but we need to set up networking first. It’s not much fun to have a standalone VAX, so we’re going to create a TAP interface and a bridge to allow our VAX to access our network.
A TAP interface virtual interface which runs at the Data Link layer (MAC) and allows us to connect the VAX’s Ethernet interface to an interface on our host operating system.
Setting the TAP interface is as easy as running this command:
ip tuntap add mode tap user ubuntu tapvax
ip link set dev tapvax up
If you’re not running under user ‘ubuntu’, change it to your user name.
To access the network, we need to set up a bridge – which is like a switch (or hub, if you’ve been working with Ethernet for as long as I have) and connects many interfaces together. We’re going to create a bridge called ‘br0’, and add the ‘tapvax’ interface to it, and add the Pi’s ‘eth0’ interface.
My preferred way of doing this is to configure Netplan to do it for us. It’s quite straightforward – create /etc/netplan/01-network.yaml with the following:
After rebooting your Pi, you’ll have an interface named ‘br0’ with a DHCP-assigned IP address, and another interface named ‘eth0’ with no IP address. Here’s the first issue – I haven’t yet found a way to configure the TAP interface with NetPlan, so each time you reboot, you’ll have to run the TAP interface setup above.
Booting the emulator
Now comes the exciting part – run ./vax which will start and take a moment or two to create your RQ0 disk. At the simh> prompt, enter ‘boot dua3’. This will boot a minimal OpenVMS image to install the base system. It will ask for the date and time, so enter it – then it’ll search for and configure devices on your system.
Three devices should appear:
Available device DUA0: device type RA92 Available device DUA1: device type RD54 Available device DUA2: device type RD54 Available device DUA3: device type RRD40
DUA0 is your system disk. DUA1 and DUA2 aren’t connected but are available, and DUA3 is your CD-ROM image. Enter ‘YES’ when prompted, and you’ll be dropped at a ‘$’ prompt. Enter the following command:
BACKUP DUA3:VMS073.B/SAVE_SET DUA0:
This will copy over a minimal install of OpenVMS in about 3-4 minutes. To actually boot from this image, press CTRL+E, type ‘boot rq0’ and the emulator will reboot.
After entering the current date and time, you will be prompted to enter a volume label for the system disk. I use the nodename followed by SYS, but it can be anything from 1-12 characters.
Next, you’ll be asked which components you want to install. A reasonable set is:
OpenVMS Help Message
DECnet Phase IV networking
Before installing DECnet Phase IV, you’ll be asked if you want to continue – just select yes.
A word on DECnet
Setting the SCSSYSTEMID parameter needs a bit of background knowledge about DECnet addressing. In short, DECnet addresses comprise an area (1 to 63) and a node (1 to 1023). There is a relationship between a DECnet address and a MAC address, and all DECnet hosts have a MAC address starting AA-00-04-00.
The SCSSYSTEMID is calculated by multiplying the DECnet area by 1024, and then adding the node number. For simplicity, I’ve used addresses 1.1 and 1.2 for my two hosts, which means the SCSSYSTEMIDs will be 1025 and 1026.
After DECnet Phase IV has been configured, you’ll be asked if you want to register any Product Authorization Keys. Select no – we’ll do that later.
Select your timezone and daylight savings settings, and the installation process will finish up by running AUTOGEN and rebooting.
Configuring your system
The first task to do is decompress the OpenVMS libraries. Disk space isn’t an issue, so we might as well run @SYS$UPDATE:LIBDECOMP.COM and decompress all libraries. This will take a while!
At this point, we should install the OpenVMS Hobbyist licences we’ve got. These can either by typed in at the console very slowly, or the script run from a CD-ROM image. We’re going to go with the latter.
To do this, copy the licence script in to a file called vms_paks.com and put this in a directory with no other files. The ‘mkisofs’ command will create an ISO image, and we run it as follows:
mkisofs . > /tmp/licence.iso
To make this image accessible to our OpenVMS machine, we need to quickly break out of the running operating system and type this at the console:
attach rq3 -r /tmp/licence.iso
When you’re back in OpenVMS, mount the image using the following command:
You can shorten OVERRIDE=IDENTIFICATION down to OVER=ID. This parameter tells OpenVMS to mount the image as-is, without trying to mount it according to the volume name.
We could run the script straight from our CD-ROM image by running @DUA3:VMS_PAKS.COM, but this won’t work – the line endings are wrong. To work around this, copy the file to your home directory and set some attributes on it:
COPY DUA3:VMS_PAKS.COM SYS$LOGIN
SET FILE/ATTRIBUTE=(RFM=STM) VMS_PAKS.COM
When you run @VMS_PAKS.COM, it’ll happily install all the licences. You can see what’s installed by running SHOW LICENSE:
Almost all conventional operating systems have an IP stack available by default. OpenVMS doesn’t, so it needs to be installed.
Before we install TCP/IP, we’ll need to increase the global pages value – by default, it’s 15000 and we need it to be higher.
Add the following lines:
Next, run AUTOGEN:
@SYS$UPDATE:AUTOGEN GETDATA SETPARAMS FEEDBACK
The installation is in the director TCPIP_VAX051 on the OpenVMS CD-ROM. We can mount it and install with the following commands:
SET DEFAULT DUA3:[TCPIP_VAX051.KIT]
PRODUCT INSTALL *
There isn’t much to configure with the initial installation, so we can start to configure:
Select option 1 – Core Environment, and option 2 – Interfaces. Select option 2 to use DHCP, and option 1 to enable DHCP_CLIENT.
That’s all folks
Wow, what a lot of work. Operating systems from decades past were installed and configured, then often never shut down for months or years. How times have changed.
TechRadar has a great article on securing printers, but how do you automate it? Well, with a certificate authority like Let’s Encrypt for starters, but there’s no mechanism for the printer to automatically update its certificate after it expires.
I’ve set my desktop machine to certbot and renew the certificate automatically. An evening’s hacking around the web interface showed it’s really easy to install a certificate automatically.
This is the magic command to install the certificate:
Whilst testing out some PostgreSQL replication scenarios, I needed to have multiple instances of Postgres running under my user. I could have used multiple Docker machines with ports exposed, but I opted for another solution. Here’s how I did it.
First, I created /home/user/pgdata to hold the data directories, and /tmp/postgres to hold the UNIX socket files.
Then, it’s just a case of running /usr/lib/postgresql/12/bin/initdb /home/user/pgdata/1 -E UTF-8 to create a data directory, then editing postgresql.conf and changing the TCP port from 5432 to, say, 50432, and unix_socket_directories to /tmp/postgresql.
To start the server, run /usr/lib/postgresql/12/bin/pg_ctl -D /home/user/pgdata/1 -l logfile start, and to stop it, replace "start" with "stop". The final step is to create a postgres user with /usr/lib/postgresql/12/bin/createuser -drs postgres -h localhost -p 50432 and you’re away.
Some use cases I can think of for this are:
An isolated Postgres instance for integration testing
Trying out a new version of Postgres in parallel with your existing version
I have a Philips Hue bridge which lets me control the lights in my flat in a variety of useful ways. It’s a good bit of kit, but with one major problem – it assumes you’re running PAT (Port Address Translation), and that your Hue bridge and the device you access https://account.meethue.com/bridge both access the Internet from the same source IP address. If not, even though the devices may be in the same broadcast domain and the same IPv4 subnet, you won’t be able to link your Hue to your account.
Despite tweeting for assistance, I ended up crying shibboleet and reverse-engineering the method of linking they’re using. Here’s what I found out, in the hope it’ll save somebody else a lot of time.
My Internet services are through the excellent A&A, and I can’t recommend them highly enough. I have a public IPv4 subnet, and each of my devices accesses the Internet without any address translation. Inbound connectivity is restricted – there are only a few things I need accessible from the Internet. (As an aside, I have two DSL lines, with my IPv4 subnet routed down each – load balancing and resilience)
My Hue bridge connects to https://discovery.meethue.com/, and that service makes a note of an inventory that the bridge sends to it. Here’s where the problem is – visiting discovery.meethue.com only returns the devices that registered from the IP address you’re connecting from. That’s fine if all your devices go through address translation and appear to come from a single external IP address, but useless for me – my mobile device uses an entirely different IPv4 address, as does my desktop and laptop. The Hue app reports that no devices were found.
After some frustrating interactions on Twitter, I solved the problem myself. I set up IP Masquerade – essentially port address translation behind the router’s external IPv4 address – for my Hue bridge and my mobile device, so they’d appear to be coming from the router’s external IP address. Rebooting the Hue, disabling one of the PPP connections on my router (necessary since they both have an IP address assigned, and my outbound traffic is load-balanced per TCP connection) and linking the device from my mobile phone then worked. Rolling it all back and rebooting the Hue again leaves the device linked to my account.
What a mess. Adding a “Enter the IP address of your Hue then when prompted, press the button” on the device linking page would have been a whole load easier. Not everyone’s Internet connection is the same, nor is everyone as experienced in network engineering as I am… yet still it took me three days to work out a fix.
In summary: buy Hue devices – they’re good, but beware if you’re doing anything that possibly deviates from the common case.
A project I’m working on requires a Kerberos and LDAP infrastructure. Most technology projects are easy to do badly, and more difficult to do well – and documented.
We use LetsEncrypt to issue a certificate to each server, and OpenLDAP can take the certificate and use it to encrypt and authenticate connections from other LDAP servers.
One of the problems I encountered was when setting up replication between servers. We use SaltStack to build and maintain our server estate, so deployment and configuration needs to be automated. This normally means spending a week automating a process which you’ll only do twice – once when you install it, and once again when you’re rebuilding the server and have forgotten everything you’ve done.
To secure OpenSSL, add this LDIF file to your directory:
Restarting slapd, the LDAP daemon, with full debugging showed nothing, neither did running the daemon through strace.
Several days of painful troubleshooting followed. We eventually found the issue – the LDAP daemon wasn’t able to access the TLS certificates! AppArmor was blocking access to the files under /etc/letsencrypt, and so we did two simple things.
First, we used setfacl to give the openldap user ‘rx’ permissions on /etc/letsencrypt/live and /etc/letsencrypt/archive:
How do you turn off your HP LaserJet printer’s built-in WiFi Direct SSID?
A new HP LaserJet printer arrived in the office recently. After plugging it on to the office network, the setup was easy, but I couldn’t find a way to stop the printer broadcasting a wireless network name.
Several hours of testing, failing, resetting and re-testing led me to HP’s Support Forum. I found some instructions from a user who worked out how to disable the network. It’s not logical or obvious.
First, make sure you don’t have a USB cable attached to the printer. Go to the web configuration page and select Networking > Wi-Fi Direct Setup, and turn WiFi Direct on.
Then, set the connection method to Advanced, and check “Do not broadcast the Wi-Fi Direct Name”. Click Apply, restart the printer, then go back in to the configuration page and turn Wi-Fi Direct off.
I don’t know why this isn’t an obvious option, but HP might fix it in a later release of the printer’s firmware.
When I was inexperienced in Java, I wrote a lot of tests using assert(), rather than using assertEquals().
Revisiting code today, I wanted to update many test suites to use assertEquals(), which requires I flip the expected and actual values around. Too difficult to do quickly by hand, so I used the following regular expression:
Find: assert \((.+)\)\.equals\((.+)\)\);
Replace: assertEquals($2, $1));
It worked like a treat.
Installing a TLS (or SSL) certificate on a HP LaserJet printer
The fantastic and free Let’s Encrypt service lets you issue TLS (SSL) certificates to as many devices as you want. It’s perfect for a home or small office environment.
The Let’s Encrypt service needs to validate that you are in control of the device you’ve requested a certificate for. Most of the time, it’s fine to serve up a single file from your server. What if you can’t actually serve arbitrary files from your device?
There is a way around this – you can use a TXT (text) record in DNS to authenticate the device, and that’s what I did.