I would like an account if you open up telnet. Thanks!
Sent from my iPhone
On Jan 14, 2014, at 4:52 PM, Mark Abene <phiber at phiber.com> wrote:
False alarm. Sampsa's bridge was apparently broken. I peered with
Mark Wickens and my adjacency came right up. Good to go. Would
anyone like an account on my TOPS-20 system? It's at 4.1022, node
name GLGMSH. It's also reachable via LAT. I may consider opening it
up via telnet as well.
On Tue, Jan 14, 2014 at 8:06 AM, <Paul_Koning at dell.com> wrote:
I d suggest firing up wireshark or equivalent to capture the DECnet traffic. See what it shows. Wireshark has some DECnet packet annotation (though some of it is badly broken route packets are not parsed right). Look for packets from the neighbor routers you re expecting to see. And look at the hello packets from both sides and match them against what the DECnet routing spec says should happen.
If necessary, post a few lines from the result and we can look it over.
paul
On Jan 14, 2014, at 2:38 AM, Mark Abene <phiber at phiber.com> wrote:
I'm only ever seeing LAT and MOP packets coming over the bridge; my
DECnet 0x6003 packets get silently ignored, and I never see any coming
from the far end of the bridge. Is anyone willing to bridge with me
as a test for comparison? I don't yet have an area assigned by Johnny
since I only have one machine at present (but that may change), so
I've been using the same area number as the remote bridge peer.
-Mark
On Mon, Jan 13, 2014 at 10:45 PM, Mark Abene <phiber at phiber.com> wrote:
OK, I'm at the hair-tearing-out stage... Who here has a TOPS-20 system
connected to HECnet? I'm trying to peer with Sampsa with "bridge" in
his same area as an endnode. Bridge works fine on my end, lots of
activity in debug mode, and I see lots of node number entries if I
send it a SIGUSR1. However I never "see" the adjacent node in NCP,
though my circuit and line are up and active. Adding a few nodes
manually with NCP SET NODE and then trying a SET HOST in the monitor
just times out. If I do this while watching my ethernet with tcpdump,
I see my real DECnet traffic, immediately followed by its encapsulated
udp packet going out over port 4711. However, I never get any
responses to my SET HOST from the remote bridge side. Any suggestions
on debugging this?
Thanks,
Mark
False alarm. Sampsa's bridge was apparently broken. I peered with
Mark Wickens and my adjacency came right up. Good to go. Would
anyone like an account on my TOPS-20 system? It's at 4.1022, node
name GLGMSH. It's also reachable via LAT. I may consider opening it
up via telnet as well.
On Tue, Jan 14, 2014 at 8:06 AM, <Paul_Koning at dell.com> wrote:
I d suggest firing up wireshark or equivalent to capture the DECnet traffic. See what it shows. Wireshark has some DECnet packet annotation (though some of it is badly broken route packets are not parsed right). Look for packets from the neighbor routers you re expecting to see. And look at the hello packets from both sides and match them against what the DECnet routing spec says should happen.
If necessary, post a few lines from the result and we can look it over.
paul
On Jan 14, 2014, at 2:38 AM, Mark Abene <phiber at phiber.com> wrote:
I'm only ever seeing LAT and MOP packets coming over the bridge; my
DECnet 0x6003 packets get silently ignored, and I never see any coming
from the far end of the bridge. Is anyone willing to bridge with me
as a test for comparison? I don't yet have an area assigned by Johnny
since I only have one machine at present (but that may change), so
I've been using the same area number as the remote bridge peer.
-Mark
On Mon, Jan 13, 2014 at 10:45 PM, Mark Abene <phiber at phiber.com> wrote:
OK, I'm at the hair-tearing-out stage... Who here has a TOPS-20 system
connected to HECnet? I'm trying to peer with Sampsa with "bridge" in
his same area as an endnode. Bridge works fine on my end, lots of
activity in debug mode, and I see lots of node number entries if I
send it a SIGUSR1. However I never "see" the adjacent node in NCP,
though my circuit and line are up and active. Adding a few nodes
manually with NCP SET NODE and then trying a SET HOST in the monitor
just times out. If I do this while watching my ethernet with tcpdump,
I see my real DECnet traffic, immediately followed by its encapsulated
udp packet going out over port 4711. However, I never get any
responses to my SET HOST from the remote bridge side. Any suggestions
on debugging this?
Thanks,
Mark
On Tue, 14 Jan 2014, Mark Abene wrote:
Sure. Since dynamips runs the real IOS with a real config, I can send
you a copy of my config so you'll see exactly how to config the cisco
as a packet switching router with integrated vintage X.29 PAD.
Thanks! I'd appreciate that!
-Mark
On Tue, Jan 14, 2014 at 10:38 AM, Cory Smelosky <b4 at gewt.net> wrote:
On Tue, 14 Jan 2014, Mark Abene wrote:
Yes, this is purely for nostalgic reasons. It's great having an
actual X.25 network again after so long. The routing and PAD access
is accomplished on emulated Ciscos (using dynamips/dynagen). DECnet
Phase V/Plus/OSI is unfortunately needed by VMS in order to support
the X.25 protocol itself. However "xotd" will work on any linux
system, provided you mod a network listener daemon to bind to address
family X25 (I modded telnetd as a simple test) so you have something
to connect to.
And if anyone knows where I might find an old tape installer image of
PSI for TOPS-20, let me know. :)
Add me to this. I'll use my real Cisco and some sort of terminal server
hackery.
-Mark
On Tue, Jan 14, 2014 at 4:02 AM, Sampsa Laine <sampsa at mac.com> wrote:
(Johnny - sorry for the plagiarised name, just thought it fit)
So Mark Abene and me (well 99% Mark Abene) has managed to get a working
X.25-over-IP setup going. We can currently connect Linux systems and VMS
boxes, but unfortunately they have to be running DECNET Phase V..
In any case, it's a fun toy, I don't think it's actually useful in any
sane sense but I quite like the idea of having my own X.25 network..
sampsa <sampsa at mac.com>
mobile +44 7961 149465
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
Sure. Since dynamips runs the real IOS with a real config, I can send
you a copy of my config so you'll see exactly how to config the cisco
as a packet switching router with integrated vintage X.29 PAD.
-Mark
On Tue, Jan 14, 2014 at 10:38 AM, Cory Smelosky <b4 at gewt.net> wrote:
On Tue, 14 Jan 2014, Mark Abene wrote:
Yes, this is purely for nostalgic reasons. It's great having an
actual X.25 network again after so long. The routing and PAD access
is accomplished on emulated Ciscos (using dynamips/dynagen). DECnet
Phase V/Plus/OSI is unfortunately needed by VMS in order to support
the X.25 protocol itself. However "xotd" will work on any linux
system, provided you mod a network listener daemon to bind to address
family X25 (I modded telnetd as a simple test) so you have something
to connect to.
And if anyone knows where I might find an old tape installer image of
PSI for TOPS-20, let me know. :)
Add me to this. I'll use my real Cisco and some sort of terminal server
hackery.
-Mark
On Tue, Jan 14, 2014 at 4:02 AM, Sampsa Laine <sampsa at mac.com> wrote:
(Johnny - sorry for the plagiarised name, just thought it fit)
So Mark Abene and me (well 99% Mark Abene) has managed to get a working
X.25-over-IP setup going. We can currently connect Linux systems and VMS
boxes, but unfortunately they have to be running DECNET Phase V..
In any case, it's a fun toy, I don't think it's actually useful in any
sane sense but I quite like the idea of having my own X.25 network..
sampsa <sampsa at mac.com>
mobile +44 7961 149465
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On Tue, 14 Jan 2014, Mark Abene wrote:
Yes, this is purely for nostalgic reasons. It's great having an
actual X.25 network again after so long. The routing and PAD access
is accomplished on emulated Ciscos (using dynamips/dynagen). DECnet
Phase V/Plus/OSI is unfortunately needed by VMS in order to support
the X.25 protocol itself. However "xotd" will work on any linux
system, provided you mod a network listener daemon to bind to address
family X25 (I modded telnetd as a simple test) so you have something
to connect to.
And if anyone knows where I might find an old tape installer image of
PSI for TOPS-20, let me know. :)
Add me to this. I'll use my real Cisco and some sort of terminal server hackery.
-Mark
On Tue, Jan 14, 2014 at 4:02 AM, Sampsa Laine <sampsa at mac.com> wrote:
(Johnny - sorry for the plagiarised name, just thought it fit)
So Mark Abene and me (well 99% Mark Abene) has managed to get a working X.25-over-IP setup going. We can currently connect Linux systems and VMS boxes, but unfortunately they have to be running DECNET Phase V..
In any case, it's a fun toy, I don't think it's actually useful in any sane sense but I quite like the idea of having my own X.25 network..
sampsa <sampsa at mac.com>
mobile +44 7961 149465
--
Cory Smelosky
http://gewt.net Personal stuff
http://gimme-sympathy.org Projects
On Tue, 14 Jan 2014, Sampsa Laine wrote:
I'm not _exactly_ sure of the details of the whole setup (there's a Cisco virtual router involved as well) but we'll get some instructions up soon.
OK. When you have those instructions, shoot them to me or send me a URL.
Thanks,
Fred
FYI: X.25 support exists for RSX as well. Unfortunately I do not have that, so I can't even participate in playing with it.
Johnny
On 2014-01-14 08:40, Mark Abene wrote:
Yes, this is purely for nostalgic reasons. It's great having an
actual X.25 network again after so long. The routing and PAD access
is accomplished on emulated Ciscos (using dynamips/dynagen). DECnet
Phase V/Plus/OSI is unfortunately needed by VMS in order to support
the X.25 protocol itself. However "xotd" will work on any linux
system, provided you mod a network listener daemon to bind to address
family X25 (I modded telnetd as a simple test) so you have something
to connect to.
And if anyone knows where I might find an old tape installer image of
PSI for TOPS-20, let me know. :)
-Mark
On Tue, Jan 14, 2014 at 4:02 AM, Sampsa Laine <sampsa at mac.com> wrote:
(Johnny - sorry for the plagiarised name, just thought it fit)
So Mark Abene and me (well 99% Mark Abene) has managed to get a working X.25-over-IP setup going. We can currently connect Linux systems and VMS boxes, but unfortunately they have to be running DECNET Phase V..
In any case, it's a fun toy, I don't think it's actually useful in any sane sense but I quite like the idea of having my own X.25 network..
sampsa <sampsa at mac.com>
mobile +44 7961 149465
Yes, this is purely for nostalgic reasons. It's great having an
actual X.25 network again after so long. The routing and PAD access
is accomplished on emulated Ciscos (using dynamips/dynagen). DECnet
Phase V/Plus/OSI is unfortunately needed by VMS in order to support
the X.25 protocol itself. However "xotd" will work on any linux
system, provided you mod a network listener daemon to bind to address
family X25 (I modded telnetd as a simple test) so you have something
to connect to.
And if anyone knows where I might find an old tape installer image of
PSI for TOPS-20, let me know. :)
-Mark
On Tue, Jan 14, 2014 at 4:02 AM, Sampsa Laine <sampsa at mac.com> wrote:
(Johnny - sorry for the plagiarised name, just thought it fit)
So Mark Abene and me (well 99% Mark Abene) has managed to get a working X.25-over-IP setup going. We can currently connect Linux systems and VMS boxes, but unfortunately they have to be running DECNET Phase V..
In any case, it's a fun toy, I don't think it's actually useful in any sane sense but I quite like the idea of having my own X.25 network..
sampsa <sampsa at mac.com>
mobile +44 7961 149465
I d suggest firing up wireshark or equivalent to capture the DECnet traffic. See what it shows. Wireshark has some DECnet packet annotation (though some of it is badly broken route packets are not parsed right). Look for packets from the neighbor routers you re expecting to see. And look at the hello packets from both sides and match them against what the DECnet routing spec says should happen.
If necessary, post a few lines from the result and we can look it over.
paul
On Jan 14, 2014, at 2:38 AM, Mark Abene <phiber at phiber.com> wrote:
I'm only ever seeing LAT and MOP packets coming over the bridge; my
DECnet 0x6003 packets get silently ignored, and I never see any coming
from the far end of the bridge. Is anyone willing to bridge with me
as a test for comparison? I don't yet have an area assigned by Johnny
since I only have one machine at present (but that may change), so
I've been using the same area number as the remote bridge peer.
-Mark
On Mon, Jan 13, 2014 at 10:45 PM, Mark Abene <phiber at phiber.com> wrote:
OK, I'm at the hair-tearing-out stage... Who here has a TOPS-20 system
connected to HECnet? I'm trying to peer with Sampsa with "bridge" in
his same area as an endnode. Bridge works fine on my end, lots of
activity in debug mode, and I see lots of node number entries if I
send it a SIGUSR1. However I never "see" the adjacent node in NCP,
though my circuit and line are up and active. Adding a few nodes
manually with NCP SET NODE and then trying a SET HOST in the monitor
just times out. If I do this while watching my ethernet with tcpdump,
I see my real DECnet traffic, immediately followed by its encapsulated
udp packet going out over port 4711. However, I never get any
responses to my SET HOST from the remote bridge side. Any suggestions
on debugging this?
Thanks,
Mark
On 2014-01-13 22:45, Mark Abene wrote:
OK, I'm at the hair-tearing-out stage... Who here has a TOPS-20 system
connected to HECnet? I'm trying to peer with Sampsa with "bridge" in
his same area as an endnode. Bridge works fine on my end, lots of
activity in debug mode, and I see lots of node number entries if I
send it a SIGUSR1. However I never "see" the adjacent node in NCP,
though my circuit and line are up and active. Adding a few nodes
manually with NCP SET NODE and then trying a SET HOST in the monitor
just times out. If I do this while watching my ethernet with tcpdump,
I see my real DECnet traffic, immediately followed by its encapsulated
udp packet going out over port 4711. However, I never get any
responses to my SET HOST from the remote bridge side. Any suggestions
on debugging this?
If you don't have the adjacency status up when looking in NCP, there is no point in trying to do a SET HOST, or anything else, because that depends on the circuit status.
If you see both incoming and outgoing DECnet traffic, and yet don't get the adjacency status up, then there is only one thing I'd want to check.
When looking at incoming packets from a remote system, the bridge will send them out on your local ethernet. Check what the source MAC address is of those packets. If the source MAC is the machine where you run the bridge, then it will not work. I know that some machines and ethernet controllers will not allow you to set the source MAC address yourself, and in that situation, the bridge can never work.
Johnny
