As described in the earlier thread, I managed to get VAX (7.3) DECWindows sessions to run on a Linux Xserver. However, it seems pretty unreliable when compared to the Alpha (8.3) version of DECWindows running on the same server. The latter is pretty reliable and can actually be used.
What happens with the VAX version is that the mouse and/or keyboard get disconnected for some reason. Sometimes the session will run for several minutes, sometimes it locks up with no mouse or keyboard response as soon as an application is opened. This may be a timing issue, since DECWindows on SIMH VAX does seem to be quite a bit slower than it is on my Alphas.
Clearly, though, DECWindows was meant to run on DECNET rather than TCPIP (I think.) Linux has an optional protocol stack for DECNET. Is there likely any advantage to letting Linux speak OpenVMS' native language rather than asking OpenVMS to speak today's TCPIP? Has anyone tried it?
Edited by Altivo on June 20 2011 07:44
I had to first get DECnet working on my Alphas and emulated VAxen. That much is achieved, and they are talking to one another now over DECnet.
I've seen that notice that the DECnet for Linux project was "dead" too, dated late last year. But there have been updates to it in 2011, so I'm not so sure. I think what was "declared dead" was code within the Linux kernel intended to specifically support DECnet. This doesn't preclude support through loadable modules (which is the norm for most driver type stuff these days) or separate applications.
The current Ubuntu and Debian repositories do include the DECnet support packages. I'll find out soon.
Configuration was a bit touchy, and particularly so since I'm using TUN/TAP bridging for the virtual machines, but I have successfully installed the DECnet for Linux packages on two Ubuntu 10.04 machines now and they do work. The transport for X11 is not included, but is apparently available from another source, at least according to the support site for these packages.
There are three packages:
You need all three. It's DECnet Phase IV, so it will alter your MAC address. Pay attention to the (skimpy) instructions. Read them twice. Except for some difficulties getting proxy to work on the VAXes, all the expected features are functional. The syntax is altered a bit since standard VMS commands look like shell commands in Linux and the shell intercedes and messes you up. Most commands are prefixed by "dn" to keep the shell out of the way. But you get "set host" equivalence, remote directory searching, DECnet object activation, remote file copy and deletion, etc. It works and seems faster than the TCP-based equivalent operations.
I am using Virtual Box to create a dedicated Linux box to hold my SimH emulator and a dedicated W7 x64 for FreeAXP.
VirtualBox allows creating 4 virtual NICs, so I dedicated two NICs to each host and two NIC for each emulator. I have one bridged network for each VM and one internal network for local and cluster traffic. I have to run an older version of VirtualBox to allow VMS cluster traffic.
Yes, I've encountered some issues with ethernet bridging. Perhaps not surprising, but some of the code written to create and support bridges assumes that IP is the only protocol. In particular, when I installed UML (User Model Linux) utilities in order to try their bridging structure, I found that not only does their model not work with DECnet but if it is active on the system it interferes with taptap.c, the tool I've been using that does pass DECnet through. I had to disable the uml-switching process to get taptap working again.
With respect to the original question, I haven't yet found the code patch to activate DECnet transport in X11. However, I find that using the TCPIP transport becomes much more reliable if XDM is left out of the picture. When I use a DECnet task script to run DECW$STARTLOGIN on the X client machine, it avoids the bug I've mentioned before with respect to server zero vs. other server numbers. It also seems to sidestep some issues with Motif losing focus and not responding to the mouse or keyboard.
In brief, I start an X server on the Linux machine, then use
$ dntask 'REMOTENODE"SYSTEM -"::start-login'
at the Linux prompt. That's the DECnet for Linux equivalent of the VMS command
$ TYPE REMOTENODE"SYSTEM PASSWORD"::"TASK=START-LOGIN"
The REMOTENODE::START-LOGIN.COM file is run with system privileges. It creates a display, defines the DECW$DISPLAY logical at the system level, and runs DECW$STARTLOGIN.EXE
This works well to get an X session going on a Linux display from either OpenVMS VAX 7.3 or OpenVMS Alpha 8.3.
Edited by Altivo on July 03 2011 05:23
malmberg December 12 2017 HPE only makes the most current version of OpenVMS Alpha / IA64 /VAX available to hobbyists. When I had access to the Alliance 1 program it was the same. No public downloads are allowed by HPE.
nmbonao December 08 2017 Is anybody knows OpenVMS 8.2 version downloadable version? or CD copy? Thank you very much
aarommes December 02 2017 Bitcoin and Blockchain enthusiasts ( plus distributed computing ) please connec / reply: http://www.openvmshobbyis t.com/forum/viewthread.ph p?forum_id=130&thread_id= 2991
malmberg September 10 2017 https://sourceforge.net/p /vms-ports/wiki/VMSInstal lation/ For the most part just use VMS 6.1 media instead of 7.3. But why run the older release?
DoeveR August 07 2017 Where can I find the write up on running VMS 6.1 using the simh emulator?
Bart March 20 2017 Happy to have found my password again!