Bit of a puzzler here, but it'll probably turn out to be my configuration.
When the SIMH VAX/VMS 7.3 instance SIMVAX is serving hecnet.eu I get the
following area/node stats reported:
Known Area Volatile Summary as of 20-SEP-2012 11:53:22
Area State Circuit Next node to area
4 reachable QNA-0 0 (SIMVAX)
Perhaps SIMVAX cannot see any DECnet routers and is only declaring area 4
to be reachable because it is in area 4 itself?
Known Node Volatile Summary as of 20-SEP-2012 11:53:37
Executor node = 4.248 (SIMVAX)
State = on
Identification = DECnet for OpenVMS VAX V7.3
Node State Active Delay Circuit Next node
Links
1.1 (MAGICA) QNA-0 0
1.2 (ERNIE) QNA-0 0
1.3 (FNATTE) QNA-0 0
1.4 (GOBLIN) QNA-0 0
1.5 (ZEKE) QNA-0 0
1.6 (GNOME) QNA-0 0
1.7 (BJARNE) QNA-0 0
1.8 (KRILLE) QNA-0 0
...
So only my area known, and weird stats for nodes. Can anyone tell me
what is up with my configuration?
I'm running Johnny's bridge and if I bring up SLAVE, the AlphaServer
1000A, on the same area it reports all stats correctly. SIMVAX is
currently unable to see MIM. I set it up as a non-routing node. When I
tried setting it up via NETCONFIG as a routing
node I get everything unreachable.
Are you running SIMH and the bridge on the same machine?
I only ask because I find when I run SIMH on VMS, the simulated machine cannot
communicate directly with the host machine but can communicate with other
machines on the network. This might not apply in the case of whatever host OS
you are using to run SIMH on but I thought I'd mention it, just in case.
If this is the problem, try bringing up SIMVAX and SLAVE at the same time. They
should be able to see each other.
Regards,
Peter Coghlan.
You'll have to forgive me, I'm not entirely familiar with how networking on
VMS works at this level.
I'm not hugely familiar with it myself either. I only know enough to be
dangerous :-)
Can you setup something that uses pseudo devices but doesn't require
promiscuous mode and see if that works? If so then I think we can
definitely narrow it down to your promiscuous theory. If that's the
case it just then becomes a question of is the issue with the pseudo
devices themselves or the way PCAP is setting them up?
As I understand it, when VMS initially detects an ethernet adapter, it creates
a template device, such as EWA0 for example. As soon as something want's to use
the adapter for something, a pseudodevice is cloned from the template device.
For example, when DECnet is started, it creates a pseudodevice called EWA2.
When the Multinet IP stack is started, it creates another two pseudodevices,
EWA4 and EWA6, one to handle IP and one to handle ARP. Another pseudodevice
handles cluster traffic. Each pseudodevice handles a particular ethernet
protocol type. There is no problem with DECnet, IP and cluster traffic, for
example sharing the same ethernet adapter.
I don't know how PCAP-VMS interacts with this model. What it wants to do does
not seem a good fit for the way the model was designed.
I have no idea what I'm talking about, by the way. Just thinking out loud. :)
I don't know much about this either. As I said, just enough to be dangerous :-)
One day I may get around to researching it in more detail. I'd also like to
find a way to use networking on SIMH running without requiring privileges on
the host machine. Until I can do that, I am not likely to do very much with
it anyway other than some limited testing on a sacrifical system.
Regards,
Peter Coghlan.
On Thu, Sep 20, 2012 at 8:09 AM, Brian Hechinger <wonko at 4amlunch.net> wrote:
You'll have to forgive me, I'm not entirely familiar with how networking on VMS works at this level. Can you setup something that uses pseudo devices but doesn't require promiscuous mode and see if that works? If so then I think we can definitely narrow it down to your promiscuous theory. If that's the case it just then becomes a question of is the issue with the pseudo devices themselves or the way PCAP is setting them up?
I have no idea what I'm talking about, by the way. Just thinking out loud. :)
-brian
On Sep 19, 2012, at 19:25, Peter Coghlan <HECNET at beyondthepale.ie> wrote:
I would also like to run multible instances of simh. This works fine for me as
long as only one of them tries to do networking. Once I attempt to start
networking on a second instance, the networking on the first instance stops
working. This makes it difficult to run a vaxcluster on my alpha :-(
How are the simh instances sharing the Ethernet interface?
I don't really know. It appears a new EWA pseudodevice is created for each
instance. However, PCAP-VMS seems to require the pseudodevice to be in
promiscuous mode and I suspect that this is only allowed for one of them.
It's not clear to me whether networking is even supposed to work for multiple
instances. I should try installing an extra ethernet adapter and see
if that allows it to work.
Regards,
Peter Coghlan.
Hello!
I agree. As it happens despite having a good interest in what is going
on here, I happen to have some lacking of understanding regarding how
VMS does networking. However I do almost completely understand how the
modern UNIX descendants do networking. With the exception of Ultrix
however.
-----
Gregg C Levine gregg.drwho8 at gmail.com
"This signature fought the Time Wars, time and again."
Dan,
Thanks for the help.
Unfortunately it didn't make any difference.
License Management Facility V1.2
License Database File: SYS$COMMON:[SYSEXE]LMF$LICENSE.LDB;1
Created on: 16-SEP-2012
Created by user: SYSTEM
Created by LMF Version: V1.2
-----------------------------------
DVNETRTG DEC
$ show network
Product: DECNET Node: SIMVAX Address(es): 4.248
Product: TCP/IP Node: simvax.hecnet.eu Address(es): 192.168.1.127
NCP>show known areas
Known Area Volatile Summary as of 20-SEP-2012 14:46:17
Area State Circuit Next node to area
4 unreachable
NCP>show exec characteristics
Node Volatile Characteristics as of 20-SEP-2012 14:45:58
Executor node = 4.248 (SIMVAX)
Identification = DECnet for OpenVMS VAX V7.3
Management version = V4.0.0
Incoming timer = 45
Outgoing timer = 60
Incoming Proxy = Enabled
Outgoing Proxy = Enabled
NSP version = V4.1.0
Maximum links = 32
Delay factor = 80
Delay weight = 5
Inactivity timer = 60
Retransmit factor = 10
Routing version = V2.0.0
Type = routing IV
Routing timer = 600
Broadcast routing timer = 180
Maximum address = 1023
Maximum circuits = 16
Maximum cost = 1022
Maximum hops = 30
Maximum visits = 63
Maximum area = 63
Max broadcast nonrouters = 64
Max broadcast routers = 32
Maximum path splits = 1
Area maximum cost = 1022
Area maximum hops = 30
Maximum buffers = 100
Buffer size = 576
Nonprivileged user id = DECNET
Nonprivileged password = AUMIAWOSAJ
Default access = incoming and outgoing
Pipeline quota = 4032
Alias maximum links = 32
Path split policy = Normal
Maximum Declared Objects = 31
On Thu, 20 Sep 2012, Dan Williams wrote:
Hi,
You need :
ncp>def exec type routing iv
ncp>set exec state off
$@startnet
Also have you got dvnetrtg license loaded ?
That's what I normally forget.
Thanks
Dan
On 20 Sep 2012, at 12:05, Mark Wickens <mark at wickensonline.co.uk> wrote:
Bit of a puzzler here, but it'll probably turn out to be my configuration.
When the SIMH VAX/VMS 7.3 instance SIMVAX is serving hecnet.eu I get the following area/node stats reported:
Known Area Volatile Summary as of 20-SEP-2012 11:53:22
Area State Circuit Next node to area
4 reachable QNA-0 0 (SIMVAX)
Known Node Volatile Summary as of 20-SEP-2012 11:53:37
Executor node = 4.248 (SIMVAX)
State = on
Identification = DECnet for OpenVMS VAX V7.3
Node State Active Delay Circuit Next node
Links
1.1 (MAGICA) QNA-0 0
1.2 (ERNIE) QNA-0 0
1.3 (FNATTE) QNA-0 0
1.4 (GOBLIN) QNA-0 0
1.5 (ZEKE) QNA-0 0
1.6 (GNOME) QNA-0 0
1.7 (BJARNE) QNA-0 0
1.8 (KRILLE) QNA-0 0
...
So only my area known, and weird stats for nodes. Can anyone tell me what is up with my configuration?
I'm running Johnny's bridge and if I bring up SLAVE, the AlphaServer 1000A, on the same area it reports all stats correctly. SIMVAX is currently unable to see MIM. I set it up as a non-routing node. When I tried setting it up via NETCONFIG as a routing
node I get everything unreachable.
Regards, Mark.
--
http://www.wickensonline.co.ukhttp://declegacy.org.uk
You'll have to forgive me, I'm not entirely familiar with how networking on VMS works at this level. Can you setup something that uses pseudo devices but doesn't require promiscuous mode and see if that works? If so then I think we can definitely narrow it down to your promiscuous theory. If that's the case it just then becomes a question of is the issue with the pseudo devices themselves or the way PCAP is setting them up?
I have no idea what I'm talking about, by the way. Just thinking out loud. :)
-brian
On Sep 19, 2012, at 19:25, Peter Coghlan <HECNET at beyondthepale.ie> wrote:
I would also like to run multible instances of simh. This works fine for me as
long as only one of them tries to do networking. Once I attempt to start
networking on a second instance, the networking on the first instance stops
working. This makes it difficult to run a vaxcluster on my alpha :-(
How are the simh instances sharing the Ethernet interface?
I don't really know. It appears a new EWA pseudodevice is created for each
instance. However, PCAP-VMS seems to require the pseudodevice to be in
promiscuous mode and I suspect that this is only allowed for one of them.
It's not clear to me whether networking is even supposed to work for multiple
instances. I should try installing an extra ethernet adapter and see
if that allows it to work.
Regards,
Peter Coghlan.
Hi,
You need :
ncp>def exec type routing iv
ncp>set exec state off
$@startnet
Also have you got dvnetrtg license loaded ?
That's what I normally forget.
Thanks
Dan
On 20 Sep 2012, at 12:05, Mark Wickens <mark at wickensonline.co.uk> wrote:
Bit of a puzzler here, but it'll probably turn out to be my configuration.
When the SIMH VAX/VMS 7.3 instance SIMVAX is serving hecnet.eu I get the following area/node stats reported:
Known Area Volatile Summary as of 20-SEP-2012 11:53:22
Area State Circuit Next node to area
4 reachable QNA-0 0 (SIMVAX)
Known Node Volatile Summary as of 20-SEP-2012 11:53:37
Executor node = 4.248 (SIMVAX)
State = on
Identification = DECnet for OpenVMS VAX V7.3
Node State Active Delay Circuit Next node
Links
1.1 (MAGICA) QNA-0 0
1.2 (ERNIE) QNA-0 0
1.3 (FNATTE) QNA-0 0
1.4 (GOBLIN) QNA-0 0
1.5 (ZEKE) QNA-0 0
1.6 (GNOME) QNA-0 0
1.7 (BJARNE) QNA-0 0
1.8 (KRILLE) QNA-0 0
...
So only my area known, and weird stats for nodes. Can anyone tell me what is up with my configuration?
I'm running Johnny's bridge and if I bring up SLAVE, the AlphaServer 1000A, on the same area it reports all stats correctly. SIMVAX is currently unable to see MIM. I set it up as a non-routing node. When I tried setting it up via NETCONFIG as a routing
node I get everything unreachable.
Regards, Mark.
--
http://www.wickensonline.co.ukhttp://declegacy.org.uk
Bit of a puzzler here, but it'll probably turn out to be my configuration.
When the SIMH VAX/VMS 7.3 instance SIMVAX is serving hecnet.eu I get the following area/node stats reported:
Known Area Volatile Summary as of 20-SEP-2012 11:53:22
Area State Circuit Next node to area
4 reachable QNA-0 0 (SIMVAX)
Known Node Volatile Summary as of 20-SEP-2012 11:53:37
Executor node = 4.248 (SIMVAX)
State = on
Identification = DECnet for OpenVMS VAX V7.3
Node State Active Delay Circuit Next node
Links
1.1 (MAGICA) QNA-0 0
1.2 (ERNIE) QNA-0 0
1.3 (FNATTE) QNA-0 0
1.4 (GOBLIN) QNA-0 0
1.5 (ZEKE) QNA-0 0
1.6 (GNOME) QNA-0 0
1.7 (BJARNE) QNA-0 0
1.8 (KRILLE) QNA-0 0
...
So only my area known, and weird stats for nodes. Can anyone tell me what is up with my configuration?
I'm running Johnny's bridge and if I bring up SLAVE, the AlphaServer 1000A, on the same area it reports all stats correctly. SIMVAX is currently unable to see MIM. I set it up as a non-routing node. When I tried setting it up via NETCONFIG as a routing
node I get everything unreachable.
Regards, Mark.
--
http://www.wickensonline.co.ukhttp://declegacy.org.uk
I would also like to run multible instances of simh. This works fine for me as
long as only one of them tries to do networking. Once I attempt to start
networking on a second instance, the networking on the first instance stops
working. This makes it difficult to run a vaxcluster on my alpha :-(
How are the simh instances sharing the Ethernet interface?
I don't really know. It appears a new EWA pseudodevice is created for each
instance. However, PCAP-VMS seems to require the pseudodevice to be in
promiscuous mode and I suspect that this is only allowed for one of them.
It's not clear to me whether networking is even supposed to work for multiple
instances. I should try installing an extra ethernet adapter and see
if that allows it to work.
Regards,
Peter Coghlan.
Right, however that has nothing to do with my problem. :)
It's the host side support for IDLE that's not working for some reason that I really wish I could figure out. :(
-brian
On 9/19/2012 11:04 AM, Paul_Koning at Dell.com wrote:
Idle handling depends on the guest OS doing what SIMH expects. On some emulated machine types, like the PDP11, the answer is usually easy because there is a WAIT instruction that is widely used. The VAX doesn't have any such thing, and as a result SIMH tries to recognize the idle loop. That tends to be a messy heuristic, and it is vulnerable to breakage. For example, SIMH VAX knows about NetBSD, but that heuristic only works in old NetBSD releases, not in the current one. (I've been trying to figure out a good answer for the current one; there may not be other than getting some help from NetBSD by inserting a magic instruction that SIMH can look for.)
paul
On Sep 19, 2012, at 9:07 AM, Brian Hechinger wrote:
On Sep 16, 2012, at 14:23, Peter Coghlan <HECNET at beyondthepale.ie> wrote:
Perhaps there is a host platform specific problem with set cpu idle?
That's the conclusion I've come to as it works fine most places. Something about Solaris trips it up but I haven't been able to figure out what yet.
I would also like to run multible instances of simh. This works fine for me as
long as only one of them tries to do networking. Once I attempt to start
networking on a second instance, the networking on the first instance stops
working. This makes it difficult to run a vaxcluster on my alpha :-(
How are the simh instances sharing the Ethernet interface?
-brian
Idle handling depends on the guest OS doing what SIMH expects. On some emulated machine types, like the PDP11, the answer is usually easy because there is a WAIT instruction that is widely used. The VAX doesn't have any such thing, and as a result SIMH tries to recognize the idle loop. That tends to be a messy heuristic, and it is vulnerable to breakage. For example, SIMH VAX knows about NetBSD, but that heuristic only works in old NetBSD releases, not in the current one. (I've been trying to figure out a good answer for the current one; there may not be other than getting some help from NetBSD by inserting a magic instruction that SIMH can look for.)
paul
On Sep 19, 2012, at 9:07 AM, Brian Hechinger wrote:
On Sep 16, 2012, at 14:23, Peter Coghlan <HECNET at beyondthepale.ie> wrote:
Perhaps there is a host platform specific problem with set cpu idle?
That's the conclusion I've come to as it works fine most places. Something about Solaris trips it up but I haven't been able to figure out what yet.
I would also like to run multible instances of simh. This works fine for me as
long as only one of them tries to do networking. Once I attempt to start
networking on a second instance, the networking on the first instance stops
working. This makes it difficult to run a vaxcluster on my alpha :-(
How are the simh instances sharing the Ethernet interface?
-brian
