[HECnet] Decnet on RSX-11

[Kurt Hamm]

The last command in INS.CMD is failing:
>INS LB:[IPNET]MLTCOM/UIC=[1,54]/PRO=[RW,RW,,]
>INS LB:[IPNET]MLTNET
>INS LB:[IPNET]MLTCFG
>NCP SET PROC IP TOP
>NCP SET LIN IP-0-* ALL
>MNC SHO VER
Initialize common
MLTCFG version X0.5
>MNC SET CIR IP-0-0 TCA HOST 192.168.1.49:7705 PORT 0
>NCP SET CIR IP-0-0 STA ON
NCP -- Set failed, component in wrong state, Circuit

INS.CMD (Content)
---------------------------------------------------------------
DU0:[1,2]INS.CMD;255   08-JUN-19;9 10:55:28

.IFNINS MLTCOM INS LB:[IPNET]MLTCOM/UIC=[1,54]/PRO=[RW,RW,,]
.IFNINS MLTNET INS LB:[IPNET]MLTNET
.IFNINS ...MNC INS LB:[IPNET]MLTCFG
NCP SET PROC IP TOP
NCP SET LIN IP-0-* ALL
MNC SHO VER
.;
.; Repeat the following two lines for each connection wanted.
.; For TCP connections, use TCP or TCA, depending on passive or
.; active. Also, you should set a local port for passive TCP.
.; (PORT n)
.;
.; For the remote host, you can give host:port, if you
.; want. For active TCP connections, you can use local port 0,
.; to get a random source port.
.; For passive TCP connetions, you can set host and/or remote
.; port to 0, to accept conncetions from any.
.;
.; At Johnny's end: IP-0-10
.; Mark Matlock (area 30) - 147 ms
.; mnc set cir ip-0-10 tcp host 0.0.0.0:0 port
.; ncp set cir ip-0-10 sta
.;
.;mnc set cir ip-0-0 udp host <host> port <n>
.;ncp set cir ip-0-0 sta on
.;
MNC SET CIR IP-0-0 TCA HOST 192.168.1.49:7705 PORT 0
NCP SET CIR IP-0-0 STA ON

On Sat, Jun 8, 2019 at 8:35 PM Johnny Billquist <bqt at softjar.se> wrote:

> Well, just taking a peek at CURLY, I can also tell that you have not
> managed to get the multinet links up and running, but are communicating
> with DECnet over ethernet. :-)
>
>    Johnny
>
> On 2019-06-09 02:20, Kurt Hamm wrote:
> > That's it!  Everything works great.  I have the node list and can access
> > reachable nodes from the PDP-11.
> >
> > This is just really wonderful.
> >
> > As a thank you, I will leave the mailing list alone for a while.  ha!
> >
> > Kurt
> >
> > On Sat, Jun 8, 2019 at 6:42 PM Johnny Billquist <bqt at softjar.se
> > <mailto:bqt at softjar.se>> wrote:
> >
> >     Ok, a quick rundown on some of the more esoteric details about lines
> >     and
> >     circuits might be useful here.
> >
> >     First of all, Kurt, you might want to redo the NETGEN and not set the
> >     system up for DECnet routing. It takes a lot of resources, and if you
> >     don't really want it, you are much better off with a system not even
> >     generated for it.
> >
> >     The question about remote nodes are just for an initial population of
> >     your local nodename database. All you need there is really to define
> >     MIM
> >     (1.13). You can then copy the whole nodename database from MIM over
> to
> >     your machine that way instead. Let me know if you want some details
> on
> >     that step.
> >
> >
> >     On 2019-06-08 22:25, Mark Matlock wrote:
> >      > Kurt,
> >      >     If you want two circuits like UNA-0 and IP-0-0 I think you
> >     have to
> >      > be a routing node (but not an area router).
> >
> >     Not really, but maybe. If you define multiple circuits, and are an
> >     end-node, only one circuit will be active. But others can exist, and
> >     the
> >     system will fall back to some other circuit if the active one goes
> >     down.
> >     So it can be used as a back connection.
> >
> >     However, even as an end-node, you can have several *lines* on and
> >     active
> >     at the same time. And this is a trick to use if you want
> >     DECnet-over-IP,
> >     but want to be an endnode. You need ethernet for TCP/IP to work, but
> >     you
> >     do not want DECnet to use the ethernet circuit.
> >
> >     But you configure one IP line as well, and that is the one DECnet
> will
> >     be using. Meanwhile, TCP/IP is using the ethernet line.
> >
> >     It might sound a little confusing, but think about it for a second,
> and
> >     I hope it becomes clear. Otherwise, just ask, and I'll try to
> >     explain it
> >     better.
> >
> >      > Also, you should define all the nodes you want to talk to with
> >     NCP in
> >      > one of the startup files or do it with CFE but I prefer startup
> >     so it’s
> >      > simple to change.
> >
> >     I would actually say that you wouldn't. Of course, there is nothing
> >     wrong in doing this, but if you want the full HECnet nodename
> database,
> >     it contains 700 nodes. That will take a while to type in.
> >
> >     However, you can copy the nodename database from another DECnet node,
> >     and then use this.
> >
> >     But you need at least the nodename of someone where you will be
> copying
> >     from to start with, which is why I suggested adding MIM at the start.
> >
> >     After that, there is [5,54]NNC which is the nodename collector task.
> >     Run
> >     it, collect names from another node, and build a new database with
> >     this.
> >     Then stop and restart the node name server, and you have all the
> nodes
> >     defined.
> >
> >     (The node name server is managed by the command SCP.
> >     So, SCP STOP followed by SCP START will do it.)
> >
> >         Johnny
> >
> >     --
> >     Johnny Billquist                  || "I'm on a bus
> >                                         ||  on a psychedelic trip
> >     email: bqt at softjar.se <mailto:bqt at softjar.se>             ||
> >     Reading murder books
> >     pdp is alive!                     ||  tryin' to stay hip" - B. Idol
> >
>
> --
> Johnny Billquist                  || "I'm on a bus
>                                    ||  on a psychedelic trip
> email: bqt at softjar.se             ||  Reading murder books
> pdp is alive!                     ||  tryin' to stay hip" - B. Idol
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sonic.net/pipermail/hecnet-list/attachments/20190608/5ddeffe5/attachment.html>