On 22 Aug 2008, at 13:06, gerry77 at mail.com wrote:
BTW, our network is a lot smaller than HECnet and we ran without any router
for almost two years. In other words, one size does not fit all and YMMV. :-)
G.
Out of curiosity I looked at your node list on your webpage and you seem to have quite a few active hosts. It's a pity you used the same area number as us, it might be fun to link the two networks together at some point.
Sampsa
On Fri, 22 Aug 2008 08:11:37 +0200, you wrote:
Unfortunately you can't specify the timeouts that the DNS lookups should use. :-(
This the main reason why I choose not to have that functionality in my code.
This reminds me that we had some circuit stability problems in the past, due
to network transients and probably also to abnormal DNS lookup delays which
stalled the bridge. For this, and to reduce overall idle network traffic, long
ago we decided to set DECnet hello timer and LAT keepalive delays up to 90
seconds replacing the 10 (15?) seconds default value. Since then our network
is a lot more stable, i.e. no more up'n'down in operator.log and drawbacks
caused by that change (like slower adjacency change detection and so on) are
far more acceptable than the original problem. :-)
BTW, our network is a lot smaller than HECnet and we ran without any router
for almost two years. In other words, one size does not fit all and YMMV. :-)
G.
John Wilson wrote:
From: gerry77 at mail.com
Also, name lookups can take time, and meanwhile the bridge is stopped. If someone have a seriously bad setup dns, this could means hanging the bridge for half a minute per such host (or more). It also increases the
Honestly I think we have overlooked that issue and never really considered it,
maybe because we never stumbled upon it. I wonder if it would be possible to
specify a very short timeout for DNS lookups while performed by the bridge...
... or spin off a thread and do the lookup there (the Right Way :-).
That would be the right way, yes. :-)
However, I don't really feel like rewriting it as a threaded application. If I would, then I think I'd get rid of the select code as well, and have one thread per link, or possibly two...
But I'm just not enough motivated. Right now, I'm more interested in doing some more hacking in RSX...
Speaking of which. Is there anyone who have the complete distribution of Modula-2 for RSX? I have a somewhat damaged distribution, but would like to get it complete.
Johnny
Zane H. Healy wrote:
By they way, Angela, you don't need to be running dnroute if your node
is an end-node (which it seems to be). In a situation where the rest of
it doesn't seem to be working properly it's only confusing the issue I
suspect :)
Chrissie,
Am I reading this to mean that it is possible to use a Linux box as a DECnet
Area Router? If so how hard is it to setup on Ubuntu? I haven't played
with Linux DECnet in close to a decade.
It should be fairly easy, though I haven't done it for ages!
If Ubuntu has the same startup scripts as Debian then it should be just
a case of enabling routing in /etc/defaults/decnet and starting the
dnroute daemon (which you can also specify in that file). dnroute will
listen for (and send) routing messages and set routes to the nodes it
finds. To use Linux as an area router you need to do a few more things
such as start dnroute with the '-2' switch and tell the kernel the
router level eg:
echo "2" > /proc/sys/net/decnet/conf/eth0/forwarding
echo "2" > /proc/sys/net/decnet/conf/eth1/forwarding
echo "10" > /proc/sys/net/decnet/conf/eth0/priority
echo "10" > /proc/sys/net/decnet/conf/eth1/priority
the dnetinfo command will query the status of the routing daemon and
give you a display similar to the VMS SHOW NETWORK[/OLD] command.
/me makes a note to update the FAQ
Chrissie
gerry77 at mail.com skrev:
On Thu, 21 Aug 2008 11:58:01 +0200, you wrote:
What about http://decnet.ipv7.net/files/decnet-bridge_0.7-4.tar.gz ? :-P
A few things should perhaps be pointed out. The config file is not compatible with my bridge program, so people who change need to be aware that they have to modify that.
... Or you (or someone else) could integrate into your original bridge some of
our changes when- where- and however you like. You granted us permission to
use your code as a starting point for our modifications, now you are obviously
entitled to do whatever you want with our code. Don't even think to ask. :-)
Well, I would keep LAT and Infoserver traffic separate. :-)
Also, name lookups can take time, and meanwhile the bridge is stopped. If someone have a seriously bad setup dns, this could means hanging the bridge for half a minute per such host (or more). It also increases the
Honestly I think we have overlooked that issue and never really considered it,
maybe because we never stumbled upon it. I wonder if it would be possible to
specify a very short timeout for DNS lookups while performed by the bridge...
Unfortunately you can't specify the timeouts that the DNS lookups should use. :-(
This the main reason why I choose not to have that functionality in my code.
Johnny
--
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
With all this activity going on I just couldn't stay offline. :^) I swapped out my VLC for a VAXstation 4000/60 tonight (much quieter), so PDXVAX and MONK are back online. This also means that http://www.avanthar.com:8080/nodes/ works again.
Zane
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com (primary) | OpenVMS Enthusiast |
| MONK::HEALYZH (DECnet) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| PDP-10 Emulation and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
By they way, Angela, you don't need to be running dnroute if your node
is an end-node (which it seems to be). In a situation where the rest of
it doesn't seem to be working properly it's only confusing the issue I
suspect :)
Chrissie,
Am I reading this to mean that it is possible to use a Linux box as a DECnet
Area Router? If so how hard is it to setup on Ubuntu? I haven't played
with Linux DECnet in close to a decade.
Thanks,
Zane
I wouldn't trust the output from "SHOW NETWORK". I've got a similar issue on my system where "SHOW NETWORK" gives me out of date information.
What'd I'd suggest you look at is:
NCP> show exec status
or
NCP> show exec char
This will show you DECnet's currently configured address.
The information from "SHOW NETWORK" is merely poked in at system-startup time and may not reflect subsequent configuration changes.
--Marc
From: gerry77 at mail.com
Also, name lookups can take time, and meanwhile the bridge is stopped.
If someone have a seriously bad setup dns, this could means hanging the
bridge for half a minute per such host (or more). It also increases the
Honestly I think we have overlooked that issue and never really considered it,
maybe because we never stumbled upon it. I wonder if it would be possible to
specify a very short timeout for DNS lookups while performed by the bridge...
... or spin off a thread and do the lookup there (the Right Way :-).
John Wilson
D Bit
On Thu, 21 Aug 2008 11:58:01 +0200, you wrote:
What about http://decnet.ipv7.net/files/decnet-bridge_0.7-4.tar.gz ? :-P
A few things should perhaps be pointed out. The config file is not
compatible with my bridge program, so people who change need to be aware
that they have to modify that.
... Or you (or someone else) could integrate into your original bridge some of
our changes when- where- and however you like. You granted us permission to
use your code as a starting point for our modifications, now you are obviously
entitled to do whatever you want with our code. Don't even think to ask. :-)
Also, name lookups can take time, and meanwhile the bridge is stopped.
If someone have a seriously bad setup dns, this could means hanging the
bridge for half a minute per such host (or more). It also increases the
Honestly I think we have overlooked that issue and never really considered it,
maybe because we never stumbled upon it. I wonder if it would be possible to
specify a very short timeout for DNS lookups while performed by the bridge...
G.
Hi all, trying to set up my Alpha (running VMS 8.3) on HECnet.
I installed DECnet Phase IV and configured it using NETCONFIG.COM,
told it to use my new node ID 2.150
I have set up a Multinet DECnet-over-IP link
However, SHOW NET gives me "Product: DECNET Node: LOKI
Address(es): 1.1"
Do I need to change the node address somewhere else? Have I missed a step?
It has been a long time since I've installed and configured DECnet, but do
you simply need to stop and restart DECnet?
Zane
Hi all, trying to set up my Alpha (running VMS 8.3) on HECnet.
I installed DECnet Phase IV and configured it using NETCONFIG.COM,
told it to use my new node ID 2.150
I have set up a Multinet DECnet-over-IP link
However, SHOW NET gives me "Product: DECNET Node: LOKI
Address(es): 1.1"
Do I need to change the node address somewhere else? Have I missed a step?
Thanks
John
--
Ph'nglui mglw'nafh Cthulhu R'lyeh wgah'nagl fhtagn
Yup, Fidonet is still around, still some traffic on the echos even. I get my feed from http://www.easternstar.info/ (I LOVE that website, so 1994).
Until then you can of course get echo/netmail through my BBS :)
Sampsa
On 21 Aug 2008, at 16:21, Bob Armstrong wrote:
I've added a second guest account, B4BBS, that connects to my bulletin
board.
Some games (main menu option X) and Fidonet echos are available :)
I'm kinda out of touch here - is Fidonet still alive? I thought that
disappeared ages ago.
My wife used to administer a Fidonet node; if Fidonet is still around
she'll really want to try it out.
Bob
I've added a second guest account, B4BBS, that connects to my bulletin
board.
Some games (main menu option X) and Fidonet echos are available :)
I'm kinda out of touch here - is Fidonet still alive? I thought that
disappeared ages ago.
My wife used to administer a Fidonet node; if Fidonet is still around
she'll really want to try it out.
Bob
Well the ANSI backgammon is pretty kick ass..
Sampsa
On 21 Aug 2008, at 16:13, Zane H. Healy wrote:
At 1:01 PM +0100 8/21/08, Sampsa Laine wrote:
I've added a second guest account, B4BBS, that connects to my bulletin board.
Some games (main menu option X) and Fidonet echos are available :)
You're trying to convince me to get my DECnet router back up and running aren't you...
Zane
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com (primary) | OpenVMS Enthusiast |
| MONK::HEALYZH (DECnet) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| PDP-10 Emulation and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
At 1:01 PM +0100 8/21/08, Sampsa Laine wrote:
I've added a second guest account, B4BBS, that connects to my bulletin board.
Some games (main menu option X) and Fidonet echos are available :)
You're trying to convince me to get my DECnet router back up and running aren't you...
Zane
--
| Zane H. Healy | UNIX Systems Administrator |
| healyzh at aracnet.com (primary) | OpenVMS Enthusiast |
| MONK::HEALYZH (DECnet) | Classic Computer Collector |
+----------------------------------+----------------------------+
| Empire of the Petal Throne and Traveller Role Playing, |
| PDP-10 Emulation and Zane's Computer Museum. |
| http://www.aracnet.com/~healyzh/ |
Yeah, did that earlier, Chrissie suggested it over the IRC, it works now.
Sampsa
On 21 Aug 2008, at 15:22, Marc Chametzky wrote:
I'm also trying to add a gateway (i.e. TELNET out) to my BBS via a captive account, but keep getting the following error:
%SYSTEM-W-NODEVAVL, no device available
%SMG-F-INVPAS_ID, invalid pasteboard-id
I'm not positive, but I bet that one of your logicals isn't quite right inside a command procedure.
I *think* what you want is, before you run the command that's giving that error:
$ DEFINE/USER SYS$INPUT SYS$COMMAND
By default, SYS$INPUT refers to the command procedure itself, and that's not a valid source for SMG$ operations. This would redirect the input for the next run program to SYS$COMMAND which should be the controlling terminal.
--Marc
I'm also trying to add a gateway (i.e. TELNET out) to my BBS via a captive account, but keep getting the following error:
%SYSTEM-W-NODEVAVL, no device available
%SMG-F-INVPAS_ID, invalid pasteboard-id
I'm not positive, but I bet that one of your logicals isn't quite right inside a command procedure.
I *think* what you want is, before you run the command that's giving that error:
$ DEFINE/USER SYS$INPUT SYS$COMMAND
By default, SYS$INPUT refers to the command procedure itself, and that's not a valid source for SMG$ operations. This would redirect the input for the next run program to SYS$COMMAND which should be the controlling terminal.
--Marc
Hi again,
I've added a second guest account, B4BBS, that connects to my bulletin board.
Some games (main menu option X) and Fidonet echos are available :)
Sampsa
I'm also trying to add a gateway (i.e. TELNET out) to my BBS via a captive account, but keep getting the following error:
%SYSTEM-W-NODEVAVL, no device available
%SMG-F-INVPAS_ID, invalid pasteboard-id
Any ideas?
Sampsa
On 21 Aug 2008, at 12:25, Johnny Billquist wrote:
Sampsa Laine wrote:
Guys,
I've added a captive account to run a Tetris clone on CHIMPY, login TETRIS, no password needed.
Well, I don't think I have any TETRIS around, but MIM::SYS$GAMES: contains several fun programs, such as PACMAN (PAC.TSK), EMPIRE, ADVENTURE and DUNGEON.
There are some other fun programs there as well...
Johnny
Sampsa Laine wrote:
Guys,
I've added a captive account to run a Tetris clone on CHIMPY, login TETRIS, no password needed.
Well, I don't think I have any TETRIS around, but MIM::SYS$GAMES: contains several fun programs, such as PACMAN (PAC.TSK), EMPIRE, ADVENTURE and DUNGEON.
There are some other fun programs there as well...
Johnny
Johnny Billquist wrote:
Angela Kahealani wrote:
On Wed, 2008-08-20 08:16:57 Johnny Billquist wrote:
Hmm. Is this the same machine as is running the bridge?
Johnny
Yes, is that a problem?
Could very well be. The problem is that the bridge program cannot
exactly feed the ethernet input on the same machine as it is running,
nor hear data sent out on the ethernet port of the same machine. That
is, unless the hardware will do this for you. Some hardware do, but
most don't. So you don't see your own data, which makes it impossible
for my bridge program to handle the machine it itself is running on.
Johnny
I think I separated the bridge and linuxDECnet dnetd onto separate
NICs so they can see each other...
dnetinfo shows:
The next hop to the nearest area router is node 0.0
Node Cost Hops Next hop to node
1.10 4 1 if0 -> 1.266
1.500 kahea 0 0 lo -> 1.500 kahea
but I don't seem to see any neighbors, so maybe still not right.
That table looks so wrong that it's horrible. :-)
I don't even dare guessing exactly what is wrong at your end...
It's not quite as bad as it looks. 'lo' is a loopback link to the local
node (which is correct). I'm not sure where the other two nodes come
from though!
By they way, Angela, you don't need to be running dnroute if your node
is an end-node (which it seems to be). In a situation where the rest of
it doesn't seem to be working properly it's only confusing the issue I
suspect :)
Chrissie
Bob Armstrong wrote:
http://www.avanthar.com:8080/nodes/ seems to not be working to list currently active HECnet nodes.
Assuming you have a VMS, RSX or RSTS system AND your node is a routing
node (and your local node database is up to date), you can do NCP SHOW ACTIVE NODES
to see the nodes that are adjacent to you. You can also do
Technically speaking, that will list active nodes, which isn't exactly the same as adjacent nodes.
In RSX (atleast), you can also do
NCP SHOW ADJACENT NODES
to show which nodes are adjacent.
And don't forget TOPS-20. :-)
NCP SHOW KNOWN NODES
to see the up/down (reachable or unreachable) status of every node.
Well, not exactly. For nodes not in your own area, it will show which the routing node is, but if the final destination is up or down is not discernible. You'll have to ask the routine node next, to get more information on the actual destination.
Unfortunately if your node is NOT a routing node then this won't tell you
anything useful. BUT, don't despair - in that case you can ask your
favorite routing node to tell you. On your machine you can say -
NCP TELL LEGATO SHOW KNOWN NODES
and this will actually send the command from your node to NML ("Network
Management Listener") on LEGATO to be executed. You'll see the nodes that
LEGATO thinks are currently up. This will work on any target node where the
owner has been kind enough to enable the default DECnet account, e.g. MIM,
and probably most others too.
And to nit pick a bit more. Atleast in RSX, this is not related to the default DECnet account. In fact, in RSX there is no such thing as a default DECnet account.
In RSX atleast, this is an attribute called "verification" that you control on the object. #19 (NIC$$$) in this case.
Johnny