Idiots quide to networking QNX & Windows?

Hi there,

We have a QNX 4.25 development system here at work,
and a TCPIP ethernet network for all our Win98/NT
PCs.

I’ve been trying to get the QNX machine onto the network,
but with no success. Our connection to the 'Net is done
using an ISDN hub which hands out IP addresses as needed, it
also does address translation. As a result, all
the Windows machines are set to obtain IP addresses
automatically - so I don’t have any entries to put into
my /etc/hosts file

I suspect I’ll need to use dhcp.client to set up the
QNX machine, but when I try this, I still can’t ping
anything.

If I try entering a default route to the hub’s IP address,
the only way I can get “route” to accept anything is
to enter

route add default aaa.bbb.ccc.ddd

where aaa.bbb.ccc.ddd is our internal IP address for the hub
( our gateway onto the 'Net )

But this seems to remove a lot of stuff from the routing
tables, and stops ping from working.

Is there an easy-to-understand FAQ or guide to how to
get a QNX machine onto the 'Net through such a network?
I have tried looking through the knowledge base on QDN,
but there don’t seem to be any examples dealing with
a DHCP system like ours…

Thanks

Richard Bebbington
Grid Xitek Ltd.

“Richard Bebbington” <richard@rbebb.demon.co.uk> wrote in message
news:L741TDAX24A8EwLF@demon.co.uk

Hi there,

We have a QNX 4.25 development system here at work,
and a TCPIP ethernet network for all our Win98/NT
PCs.


I’ve been trying to get the QNX machine onto the network,
but with no success. Our connection to the 'Net is done
using an ISDN hub which hands out IP addresses as needed, it
also does address translation. As a result, all
the Windows machines are set to obtain IP addresses
automatically - so I don’t have any entries to put into
my /etc/hosts file

I suspect I’ll need to use dhcp.client to set up the
QNX machine, but when I try this, I still can’t ping
anything.

That should work. Make sure you have the latest
version of TCP/IP 5.0

If it doesn’t make sure syslog is running and restart
dhcp.client with -d and then examine the content of the
syslog file for error messages returned by dhcp.client.

If I try entering a default route to the hub’s IP address,
the only way I can get “route” to accept anything is
to enter

route add default aaa.bbb.ccc.ddd

where aaa.bbb.ccc.ddd is our internal IP address for the hub
( our gateway onto the 'Net )

That is the proper syntax but if you use dhcp.client you shouldn’t
need this.


But this seems to remove a lot of stuff from the routing
tables, and stops ping from working.

Is there an easy-to-understand FAQ or guide to how to
get a QNX machine onto the 'Net through such a network?

QNX TCP/IP pretty much behaves like other UNIX TCP/IP
stack. Any documenation about TCP/IP on the net should
be applicable.

I have tried looking through the knowledge base on QDN,
but there don’t seem to be any examples dealing with
a DHCP system like ours…

Your system doesn’t seems to have anything special.

Can you give more info as to how your ip mapping is
organised .

Thanks

Richard Bebbington
Grid Xitek Ltd.

In article <9u06ov$8se$1@inn.qnx.com>, Mario Charest
<mcharest@clipzinformatic.com> writes

That should work. Make sure you have the latest
version of TCP/IP 5.0

If it doesn’t make sure syslog is running and restart
dhcp.client with -d and then examine the content of the
syslog file for error messages returned by dhcp.client.

I’ve just downloaded and applied the crypt patch,
and installed TCPIP 5.

It works! I can now get the machine to browse the 'Net
through our network. Thanks for prodding me to
get the TCPIP 5 update! :wink:

But I still can’t get the machine to talk to our
Windows PCs, probably because we don’t have a
dedicated server here. Does SMBfsys require a
SMB server, or will it work with any old
WinNT4 or Win98 machine that’s on a network?

One other question:

Our DHCP/ISDN box is set up to drop the ISDN
connection after 1 minute of inactivity, this
saves us lots of money on telephone charges
( we have to pay for local calls here in the UK )

I noticed the QNX machine was keeping the connection
going even when I wasn’t doing anything
( Voyager was idle, it had finished loading a page)

I could see the network activity LED on the network
card flashing every half a second
( none of our other machines were using the Net )
and this was keeping our ISDN connection alive.

This means I have to yank the network cable out,
or the QNX PC will keep the ISDN phone line
up forever… = huge phone bill !!

Is there some way to stop this - perhaps an
option to one of the utilities?

Thanks

Richard Bebbington
Grid Xitek Ltd.

But I still can’t get the machine to talk to our
Windows PCs, probably because we don’t have a
dedicated server here. Does SMBfsys require a
SMB server, or will it work with any old
WinNT4 or Win98 machine that’s on a network?

A Win98 machine is an SMB server :wink:
What exaclty are doing?

One other question:

Our DHCP/ISDN box is set up to drop the ISDN
connection after 1 minute of inactivity, this
saves us lots of money on telephone charges
( we have to pay for local calls here in the UK )

I noticed the QNX machine was keeping the connection
going even when I wasn’t doing anything
( Voyager was idle, it had finished loading a page)

I could see the network activity LED on the network
card flashing every half a second
( none of our other machines were using the Net )
and this was keeping our ISDN connection alive.

Use netstat -a to detect what program is using the net.
To my knowledge unless a program is trying to
connect to internet your gateway shouldn’t receive
any data to route.

It’s possible that Voyager makes DNS request,
even when idle.

The blinking leds on the netword card could be cause
by such think as nameloc, ARP request etc.

This means I have to yank the network cable out,
or the QNX PC will keep the ISDN phone line
up forever… = huge phone bill !!

Is there some way to stop this - perhaps an
option to one of the utilities?

Thanks

Richard Bebbington
Grid Xitek Ltd.

Mario Charest wrote:


A Win98 machine is an SMB server > :wink:
What exaclty are doing?

Ok. That’s a relief!

I have been running the following:

SMBfsys &

then

mount_smb

(sorry I can’t remember the exact command line, I’m at home now)

After trying mount_smb, with the username “Richard”
(using the -u option)
and password ***** that I normally use to get into another
PC on the network,
( let’s use “Jeremy” as an example here)
Then the command prompt in the shell
changes from a hash ( logged in as root )
to > , and does nothing.
I’ve specified to mount_smb that I want to connect to the
PC with a NetBIOS name “Jeremy”, and mount it as “/jeremy’s_PC”,
but this never appears in the filesystem.

Do I need to specify the IP address of that PC as well as it’s
NetBIOS name?


I noticed the QNX machine was keeping the connection
going even when I wasn’t doing anything
( Voyager was idle, it had finished loading a page)

I could see the network activity LED on the network
card flashing every half a second

Use netstat -a to detect what program is using the net.
To my knowledge unless a program is trying to
connect to internet your gateway shouldn’t receive
any data to route.

That’s what I thought…


It’s possible that Voyager makes DNS request,
even when idle.

The blinking leds on the netword card could be cause
by such think as nameloc, ARP request etc.

Ok, I’ll try it again tomorrow, but without Voyager.
Voyager does tend to do wierd things sometimes :wink:



Or maybe I need to tell nameloc to only register

locally ( we only have 1 QNX PC here ) ?

Thanks,

Richard

“Richard Bebbington” <richard.bebbington@ntlworld.com> wrote in message
news:3C06AD93.7020606@ntlworld.com

Mario Charest wrote:


A Win98 machine is an SMB server > :wink:
What exaclty are doing?


Ok. That’s a relief!

I have been running the following:

SMBfsys &

then

mount_smb

(sorry I can’t remember the exact command line, I’m at home now)

After trying mount_smb, with the username “Richard”
(using the -u option)
and password ***** that I normally use to get into another
PC on the network,
( let’s use “Jeremy” as an example here)
Then the command prompt in the shell
changes from a hash ( logged in as root )
to > , and does nothing.

Usually that means the program is waiting for more input.
It happends when non matching delimiters (" ’ `)are used.
I’m guessing you are using one of these characters somewhere.

Here is the sequence I use:

SMBfsys &
user_smb user_name user_password
mount_smb //name:name/sharedrive /mount_point

The first instance of “name” must be in ascii.
The second must be resolvable. If it can’t
be resolved (no DNS) you can use the IP address.


Or maybe I need to tell nameloc to only register
locally ( we only have 1 QNX PC here ) ?

If you don’t have a netmap file, nameloc shouldn’t
send anything out.

Thanks,

Richard

In article <9ug0u2$5e3$1@inn.qnx.com>, Mario Charest
<mcharest@clipzinformatic.com> writes

Here is the sequence I use:

SMBfsys &
user_smb user_name user_password
mount_smb //name:name/sharedrive /mount_point

The first instance of “name” must be in ascii.
The second must be resolvable. If it can’t
be resolved (no DNS) you can use the IP address.

Ok, I got better results with this, if I enter

mount_smb //Nick:Nick/C /nick_c

to try to access C drive on Nick’s machine,
eventually it comes back with

Can’t find IP for ‘Nick’

How can I get it to find out the IP address of
Nick’s machine, bearing in mind that ALL our
PCs are set in Windows to “obtain an IP address automatically” ?
The only fixed IP address is the ISDN hub.


Re: the constant network activity that keeps our
ISDN line active:

Or maybe I need to tell nameloc to only register
locally ( we only have 1 QNX PC here ) ?


If you don’t have a netmap file, nameloc shouldn’t
send anything out.

After doing a netstat -a , I get the following output:

Active Internet connections ( including servers )
Proto Recv-Q Send-Q Local Address Foreign Address (state)
udp 0 0 *.bootpc .
udp 0 0 . .

So something’s using UDP and keeping the connection up.
Is *.bootpc something to do with remote booting of nodes
over the network ( we don’t have any nodes like that ).
If so, how can I disable it?

Thanks,

Richard Bebbington

“Richard Bebbington” <richard@rbebb.demon.co.uk> wrote in message
news:zj$7lCAEOQD8EwV0@demon.co.uk

In article <9ug0u2$5e3$> 1@inn.qnx.com> >, Mario Charest
mcharest@clipzinformatic.com> > writes

Here is the sequence I use:

SMBfsys &
user_smb user_name user_password
mount_smb //name:name/sharedrive /mount_point

The first instance of “name” must be in ascii.
The second must be resolvable. If it can’t
be resolved (no DNS) you can use the IP address.


Ok, I got better results with this, if I enter

mount_smb //Nick:Nick/C /nick_c

to try to access C drive on Nick’s machine,
eventually it comes back with

Can’t find IP for ‘Nick’

How can I get it to find out the IP address of
Nick’s machine, bearing in mind that ALL our
PCs are set in Windows to “obtain an IP address automatically” ?

You ISDN hub should have passed along the
IP address of the DNS. And the DNS must
have knowledge of Nick. Not all DNS have
hook to DHCP to learn about machine name.
Windows usually rely on it’s on mecanism (WINS)
to learn about machine name/IP.

If the DNS isn’t dynamic you may be force to
specify the IP address (which you would obtain
from doing a ping to Nick from a Windows machines).


After doing a netstat -a , I get the following output:

Active Internet connections ( including servers )
Proto Recv-Q Send-Q Local Address Foreign Address (state)
udp 0 0 *.bootpc .
udp 0 0 . .

So something’s using UDP and keeping the connection up.

I doubt it. It would only be the case if the Foriegn Address is set
to some machine outside your segment and the state be set
to ESTABLISHED.

Is *.bootpc something to do with remote booting of nodes
over the network ( we don’t have any nodes like that ).
If so, how can I disable it?

I think if you remove the entries in /etc/protocol that should do it.
I’m not sure.

Thanks,

Richard Bebbington