starting and configuring qnet

QNX Newsgroups (Archives) qdn.public.qnxrtp.os
1

I’m trying to use the qnet protocol between two machines running the RTP. I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–

Jason A. Farqué
CDI at http://www.pigging.com/

2

Hi Jason,

Search the QNX Knowledge base for “qnet”.

http://qdn.qnx.com/support/bok/index.html

Erick

I’m trying to use the qnet protocol between two machines running the RTP. I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–

3

I’ve done that.

There seems to be a lot of things that I’m “just supposed to know” that I
don’t. Following is the entire contents of the “How to get qnet up and
running” KB article 9583:

  1. The network i/o support should already be running. To check, from a
    terminal type:

pidin ar | grep io-net

and you should see
io-net -pttcpip -ppppmgr

For more information on io-net see the online docs at
http://qdn.qnx.com/support/docs/neutrino_qrp/utilities/i/io-net.html

  1. Mount QNET by typing:

mount -Tio-net -o npm-qnet.so

QNET issues:

The QNET nodes reside under /net/ and look like directories. You can rmdir
/net/othernode to remove them from your namespace or you can rmdir
/net/yournode to rebroadcast your presence.

As of now, QNET cannot span multiple domains. This support will be available
soon.
ie. ls /net/node.foo.com/net/joe

I can do this (and have), but this doesn’t explain how to share and export
directories between machines.

Jason A. Farqué
CDI at http://www.pigging.com/



“Hardware Support Account” <hw@qnx.com> wrote in message
news:9g5t5u$eca$1@nntp.qnx.com…

Hi Jason,

Search the QNX Knowledge base for “qnet”.

http://qdn.qnx.com/support/bok/index.html

Erick

I’m trying to use the qnet protocol between two machines running the
RTP. I
can install the drivers and qnet appears to start correctly but I only
get
‘localhost’ in the /net directory. The documentation for io-net and
qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO
on
this?

Thanks in advance.

–

\

4

Ok, some clarification.

I have RTP running on a standard pc booting a network with tcp/ip and qnet
both. I have implemented adding in the qnet protocal using the
/etc/rc.d/rc.local file with the command “mount -Tio-net
/lib/dll/npm-qnet.so”. The “full blown” pc comes up and has a /dev/io-net
directory and a /net directory that has ‘localhost’ in it. So in essence
the PC running RTP seems to be running fine.

My second machine is an embedded PC platform also running RTP, however, it’s
a command line only embedded system with only necessary files on it.
Currently I am manually starting its networking with
“io-net -dcrys8900 -pqnet &”. This does in fact create a /dev/io-net
directory with en0, qnet0 and qnet_ip as its only files, and a /net
directory. However, on the /net directory is empty - does not contain even
a “localhost” entry. And attempts for these two machines to communicate
with one another are met with “No route to host” messages.

I’m pretty confident that the piece I’m missing is on my embedded PC side -
some file I’m not copying over or some some command line switch that I’m not
using. However, what I have been unable to find in the KB or elsewhere is a
dependency list, and an order in which these components must be initialized.
Also, I understand that directories must be exported between the machines
but I can’t figure out how that is accomplished. I can’t find more than a
few words on the ‘export=’ command line option.

Can anyone straighten me out here?

–

Jason A. Farqué
CDI at http://www.pigging.com/



“Jason A. Farque’” <jasonf@pigging.com> wrote in message
news:9g5qlc$gri$1@inn.qnx.com…

I’m trying to use the qnet protocol between two machines running the RTP.
I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–

Jason A. Farqué
CDI at > http://www.pigging.com/

\

5

To be totally honest I am not certain, perhaps someone in the OS
group would be able to shed some light on the issue… Barry?.

Have you setup your hostname and domain?

Erick.




I have RTP running on a standard pc booting a network with tcp/ip and qnet
both. I have implemented adding in the qnet protocal using the
/etc/rc.d/rc.local file with the command “mount -Tio-net
/lib/dll/npm-qnet.so”. The “full blown” pc comes up and has a /dev/io-net
directory and a /net directory that has ‘localhost’ in it. So in essence
the PC running RTP seems to be running fine.

My second machine is an embedded PC platform also running RTP, however, it’s
a command line only embedded system with only necessary files on it.
Currently I am manually starting its networking with
“io-net -dcrys8900 -pqnet &”. This does in fact create a /dev/io-net
directory with en0, qnet0 and qnet_ip as its only files, and a /net
directory. However, on the /net directory is empty - does not contain even
a “localhost” entry. And attempts for these two machines to communicate
with one another are met with “No route to host” messages.

I’m pretty confident that the piece I’m missing is on my embedded PC side -
some file I’m not copying over or some some command line switch that I’m not
using. However, what I have been unable to find in the KB or elsewhere is a
dependency list, and an order in which these components must be initialized.
Also, I understand that directories must be exported between the machines
but I can’t figure out how that is accomplished. I can’t find more than a
few words on the ‘export=’ command line option.

Can anyone straighten me out here?

–



“Jason A. Farque’” <> jasonf@pigging.com> > wrote in message
news:9g5qlc$gri$> 1@inn.qnx.com> …
I’m trying to use the qnet protocol between two machines running the RTP.
I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–


\

6

First of all, you better determain the hostname of each node,
QNET will not allow 2 nodes have same name.

On your standard pc, in the rc.local file, try do:

mount -Tio-net -o bind=ether,host=nodea /lib/dll/npm-qnet.so

On your embedded PC, you start it with:

io-net -dcrys8900 -p qnet bind=ether,host=nodeb

Then you should see “nodea” and “nodeb” in both machine’s /net,
and to acces nodea’s file from nodeb, just:

ls /net/nodea/any/directory

to use the tcpip stack on nodea (from nodeB), do:

SOCK=/net/nodea telnet some_node_nodea_can_reach

-xtang

Jason A. Farque’ <jasonf@pigging.com> wrote:

Ok, some clarification.

I have RTP running on a standard pc booting a network with tcp/ip and qnet
both. I have implemented adding in the qnet protocal using the
/etc/rc.d/rc.local file with the command “mount -Tio-net
/lib/dll/npm-qnet.so”. The “full blown” pc comes up and has a /dev/io-net
directory and a /net directory that has ‘localhost’ in it. So in essence
the PC running RTP seems to be running fine.

My second machine is an embedded PC platform also running RTP, however, it’s
a command line only embedded system with only necessary files on it.
Currently I am manually starting its networking with
“io-net -dcrys8900 -pqnet &”. This does in fact create a /dev/io-net
directory with en0, qnet0 and qnet_ip as its only files, and a /net
directory. However, on the /net directory is empty - does not contain even
a “localhost” entry. And attempts for these two machines to communicate
with one another are met with “No route to host” messages.

I’m pretty confident that the piece I’m missing is on my embedded PC side -
some file I’m not copying over or some some command line switch that I’m not
using. However, what I have been unable to find in the KB or elsewhere is a
dependency list, and an order in which these components must be initialized.
Also, I understand that directories must be exported between the machines
but I can’t figure out how that is accomplished. I can’t find more than a
few words on the ‘export=’ command line option.

Can anyone straighten me out here?

–

Jason A. Farqué
CDI at > http://www.pigging.com/



“Jason A. Farque’” <> jasonf@pigging.com> > wrote in message
news:9g5qlc$gri$> 1@inn.qnx.com> …
I’m trying to use the qnet protocol between two machines running the RTP.
I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–

Jason A. Farqué
CDI at > http://www.pigging.com/

\

7

“Jason A. Farque’” a écrit :

Ok, some clarification.

I have RTP running on a standard pc booting a network with tcp/ip and qnet
both. I have implemented adding in the qnet protocal using the
/etc/rc.d/rc.local file with the command “mount -Tio-net
/lib/dll/npm-qnet.so”. The “full blown” pc comes up and has a /dev/io-net
directory and a /net directory that has ‘localhost’ in it. So in essence
the PC running RTP seems to be running fine.

My second machine is an embedded PC platform also running RTP, however, it’s
a command line only embedded system with only necessary files on it.
Currently I am manually starting its networking with
“io-net -dcrys8900 -pqnet &”. This does in fact create a /dev/io-net
directory with en0, qnet0 and qnet_ip as its only files, and a /net
directory. However, on the /net directory is empty - does not contain even
a “localhost” entry. And attempts for these two machines to communicate
with one another are met with “No route to host” messages.

I’m pretty confident that the piece I’m missing is on my embedded PC side -
some file I’m not copying over or some some command line switch that I’m not
using. However, what I have been unable to find in the KB or elsewhere is a
dependency list, and an order in which these components must be initialized.
Also, I understand that directories must be exported between the machines
but I can’t figure out how that is accomplished. I can’t find more than a
few words on the ‘export=’ command line option.

Can anyone straighten me out here?

–

Jason A. Farqué
CDI at > http://www.pigging.com/

“Jason A. Farque’” <> jasonf@pigging.com> > wrote in message
news:9g5qlc$gri$> 1@inn.qnx.com> …
I’m trying to use the qnet protocol between two machines running the RTP.
I
can install the drivers and qnet appears to start correctly but I only get
‘localhost’ in the /net directory. The documentation for io-net and qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO on
this?

Thanks in advance.

–

Jason A. Farqué
CDI at > http://www.pigging.com/


\

I’m afraid that you are going to waste your time; since the beginning I try to
play with qnet and it’s allways so incomprehensible. Sometimes some machines
goes out of /net, without any clear reason, sometimes while copying files you
get a 'network is down! message, sometimes when machines boot up, you only have
the local machine in /net, etc…, etc…, etc…

Don’t know if some machines can run tiny stack while some others run full stack,
anyway I’m using Samba or nfs to share with NT, nfs to share between qrtp
machines. Of course, not so easy to use, not as easy as qnet would be able to
do!!!

Some noises say that qnet is much stable in 6.1, … wait and see!
…


regards,
Alain.

8

xtang is my friend.

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on
something such as this? One interesting thing is that on my embedded PC
there’s clearly a config file missing because it is listing the other
machine as “nodea.” (note the period) while the full PC is showing the
embedded PC as “nodeb.net.intra”. So now I have to root around and figure
out how to fix that.

Is this stuff documented anywhere? Yes I’ve searched the KB and gone
through the online docs - the information is not sufficient for a neophyte
such as myself to know for sure that I’ve got everything set up correctly.
I may get everything working and looking proper, but how can I be sure that
there’s not some serious performance flaw or misconfiguration of my system?

–
Jason A. Farqué
CDI at http://www.pigging.com/



“Xiaodan Tang” <xtang@qnx.com> wrote in message
news:9g6ddo$mj7$1@nntp.qnx.com…

First of all, you better determain the hostname of each node,
QNET will not allow 2 nodes have same name.

On your standard pc, in the rc.local file, try do:

mount -Tio-net -o bind=ether,host=nodea /lib/dll/npm-qnet.so

On your embedded PC, you start it with:

io-net -dcrys8900 -p qnet bind=ether,host=nodeb

Then you should see “nodea” and “nodeb” in both machine’s /net,
and to acces nodea’s file from nodeb, just:

ls /net/nodea/any/directory

to use the tcpip stack on nodea (from nodeB), do:

SOCK=/net/nodea telnet some_node_nodea_can_reach

-xtang

Jason A. Farque’ <> jasonf@pigging.com> > wrote:
Ok, some clarification.

I have RTP running on a standard pc booting a network with tcp/ip and
qnet
both. I have implemented adding in the qnet protocal using the
/etc/rc.d/rc.local file with the command “mount -Tio-net
/lib/dll/npm-qnet.so”. The “full blown” pc comes up and has a
/dev/io-net
directory and a /net directory that has ‘localhost’ in it. So in
essence
the PC running RTP seems to be running fine.

My second machine is an embedded PC platform also running RTP, however,
it’s
a command line only embedded system with only necessary files on it.
Currently I am manually starting its networking with
“io-net -dcrys8900 -pqnet &”. This does in fact create a /dev/io-net
directory with en0, qnet0 and qnet_ip as its only files, and a /net
directory. However, on the /net directory is empty - does not contain
even
a “localhost” entry. And attempts for these two machines to communicate
with one another are met with “No route to host” messages.

I’m pretty confident that the piece I’m missing is on my embedded PC
side -
some file I’m not copying over or some some command line switch that I’m
not
using. However, what I have been unable to find in the KB or elsewhere
is a
dependency list, and an order in which these components must be
initialized.
Also, I understand that directories must be exported between the
machines
but I can’t figure out how that is accomplished. I can’t find more than
a
few words on the ‘export=’ command line option.

Can anyone straighten me out here?

–

Jason A. Farqué
CDI at > http://www.pigging.com/



“Jason A. Farque’” <> jasonf@pigging.com> > wrote in message
news:9g5qlc$gri$> 1@inn.qnx.com> …
I’m trying to use the qnet protocol between two machines running the
RTP.
I
can install the drivers and qnet appears to start correctly but I only
get
‘localhost’ in the /net directory. The documentation for io-net and
qnet
aren’t really clear how to perform the proper exports to make machine A
appear to machine B and vice versa. Could someone point me to a HOW-TO
on
this?

Thanks in advance.

–

Jason A. Farqué
CDI at > http://www.pigging.com/



\

9

Jason A. Farque’ <jasonf@pigging.com> wrote:

xtang is my friend.

Heh :slight_smile:

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on
something such as this? One interesting thing is that on my embedded PC

There is warning print out on stderr, “Name Reject !” or sth like that.
The usual problem is because io-net’s stderr didn’t assoicate to any
console. If you do a “reopen” before starting io-net in startup script,
you might be see it.

there’s clearly a config file missing because it is listing the other
machine as “nodea.” (note the period) while the full PC is showing the
embedded PC as “nodeb.net.intra”. So now I have to root around and figure
out how to fix that.

The “node name” works like this:

A node name is a “hostname.domain” string. If your machine’s CS_DOMAIN
is NULL, qnet will add a default domain “net.intra” to it.

Under /net, if the nodes’ domain match local domain, only the node’s
name get displayed.

To your problem, since the your embedded PC don’t have CS_DOMAIN set,
it call himself as “nodeb.net.intra” (as seen on nodea).
The reason nodea seen as “nodea.” is because your full pc does have
a CS_DOMAIN as " ". You may want to check your /etc/net.cfg file, or
run phlip on FULL PC (under photon), and fill out the domain part.

Is this stuff documented anywhere? Yes I’ve searched the KB and gone
through the online docs - the information is not sufficient for a neophyte
such as myself to know for sure that I’ve got everything set up correctly.
I may get everything working and looking proper, but how can I be sure that
there’s not some serious performance flaw or misconfiguration of my system?

There used to be a “network administrition document” covering QNET,
but I am not sure if it is still online (and up to date :slight_smile:

-xtang

10

Previously, Xiaodan Tang wrote in qdn.public.qnxrtp.os:

Jason A. Farque’ <> jasonf@pigging.com> > wrote:
xtang is my friend.

Heh > :slight_smile:

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on
something such as this? One interesting thing is that on my embedded PC

There is warning print out on stderr, “Name Reject !” or sth like that.
The usual problem is because io-net’s stderr didn’t assoicate to any
console. If you do a “reopen” before starting io-net in startup script,
you might be see it.

there’s clearly a config file missing because it is listing the other
machine as “nodea.” (note the period) while the full PC is showing the
embedded PC as “nodeb.net.intra”. So now I have to root around and figure
out how to fix that.

The “node name” works like this:

A node name is a “hostname.domain” string. If your machine’s CS_DOMAIN
is NULL, qnet will add a default domain “net.intra” to it.

Under /net, if the nodes’ domain match local domain, only the node’s
name get displayed.

To your problem, since the your embedded PC don’t have CS_DOMAIN set,
it call himself as “nodeb.net.intra” (as seen on nodea).
The reason nodea seen as “nodea.” is because your full pc does have
a CS_DOMAIN as " ". You may want to check your /etc/net.cfg file, or
run phlip on FULL PC (under photon), and fill out the domain part.

Is this stuff documented anywhere? Yes I’ve searched the KB and gone
through the online docs - the information is not sufficient for a neophyte
such as myself to know for sure that I’ve got everything set up correctly.
I may get everything working and looking proper, but how can I be sure that
there’s not some serious performance flaw or misconfiguration of my system?

There used to be a “network administrition document” covering QNET,
but I am not sure if it is still online (and up to date > :slight_smile:

I’ve been avidly following this thread since I haven’t been able to set
a domain name on my system. When I run phlip - it refuses to allow
setting a domain. If I edit /etc/net.cfg and edit the domain line - phlip
deletes the line.

The file looks like this:

nto network config file v1.2

version v1.2

[global]
hostname shadow
domain
nameserver 24.0.48.33
nameserver 24.0.48.34
route 192.168.1.1 0.0.0.0 0.0.0.0
lookup file bind

[en0]
type ethernet
mode dhcp
dhcp_id shadow
manual_ip 192.168.1.100
manual_netmask 255.255.255.0

Is it some problem with dhcp?

\

David L. Hawley D.L. Hawley and Associates

11

David Hawley <David.L.Hawley@computer.org> wrote:

Previously, Xiaodan Tang wrote in qdn.public.qnxrtp.os:
Jason A. Farque’ <> jasonf@pigging.com> > wrote:
xtang is my friend.

Heh > :slight_smile:

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on
something such as this? One interesting thing is that on my embedded PC

There is warning print out on stderr, “Name Reject !” or sth like that.
The usual problem is because io-net’s stderr didn’t assoicate to any
console. If you do a “reopen” before starting io-net in startup script,
you might be see it.

there’s clearly a config file missing because it is listing the other
machine as “nodea.” (note the period) while the full PC is showing the
embedded PC as “nodeb.net.intra”. So now I have to root around and figure
out how to fix that.

The “node name” works like this:

A node name is a “hostname.domain” string. If your machine’s CS_DOMAIN
is NULL, qnet will add a default domain “net.intra” to it.

Under /net, if the nodes’ domain match local domain, only the node’s
name get displayed.

To your problem, since the your embedded PC don’t have CS_DOMAIN set,
it call himself as “nodeb.net.intra” (as seen on nodea).
The reason nodea seen as “nodea.” is because your full pc does have
a CS_DOMAIN as " ". You may want to check your /etc/net.cfg file, or
run phlip on FULL PC (under photon), and fill out the domain part.

Is this stuff documented anywhere? Yes I’ve searched the KB and gone
through the online docs - the information is not sufficient for a neophyte
such as myself to know for sure that I’ve got everything set up correctly.
I may get everything working and looking proper, but how can I be sure that
there’s not some serious performance flaw or misconfiguration of my system?

There used to be a “network administrition document” covering QNET,
but I am not sure if it is still online (and up to date > :slight_smile:

I’ve been avidly following this thread since I haven’t been able to set
a domain name on my system. When I run phlip - it refuses to allow
setting a domain. If I edit /etc/net.cfg and edit the domain line - phlip
deletes the line.

The file looks like this:

nto network config file v1.2

version v1.2

[global]
hostname shadow
domain
nameserver 24.0.48.33
nameserver 24.0.48.34
route 192.168.1.1 0.0.0.0 0.0.0.0
lookup file bind

[en0]
type ethernet
mode dhcp
dhcp_id shadow
manual_ip 192.168.1.100
manual_netmask 255.255.255.0

Is it some problem with dhcp?

I believe it because your dhcp server either do not send you a
domain, or send you a blank domain. Since you specificly say
you want dhcp, phlip will just trust dhcp, and erase whatever
you type in. (The information phlip displayed, could be read
as “currect setting”, since you are dhcp, you are not allowed
to change anything :slight_smile:

Of cause, this is not very nice. The behaver is changed in
later version of RTP.

For now, you can move your en0 to “mode manual”, and set everything
by yourself.

-xtang

12

David Hawley a écrit :

Previously, Xiaodan Tang wrote in qdn.public.qnxrtp.os:
Jason A. Farque’ <> jasonf@pigging.com> > wrote:
xtang is my friend.

Heh > :slight_smile:

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on
something such as this? One interesting thing is that on my embedded PC

There is warning print out on stderr, “Name Reject !” or sth like that.
The usual problem is because io-net’s stderr didn’t assoicate to any
console. If you do a “reopen” before starting io-net in startup script,
you might be see it.

there’s clearly a config file missing because it is listing the other
machine as “nodea.” (note the period) while the full PC is showing the
embedded PC as “nodeb.net.intra”. So now I have to root around and figure
out how to fix that.

The “node name” works like this:

A node name is a “hostname.domain” string. If your machine’s CS_DOMAIN
is NULL, qnet will add a default domain “net.intra” to it.

Under /net, if the nodes’ domain match local domain, only the node’s
name get displayed.

To your problem, since the your embedded PC don’t have CS_DOMAIN set,
it call himself as “nodeb.net.intra” (as seen on nodea).
The reason nodea seen as “nodea.” is because your full pc does have
a CS_DOMAIN as " ". You may want to check your /etc/net.cfg file, or
run phlip on FULL PC (under photon), and fill out the domain part.

Is this stuff documented anywhere? Yes I’ve searched the KB and gone
through the online docs - the information is not sufficient for a neophyte
such as myself to know for sure that I’ve got everything set up correctly.
I may get everything working and looking proper, but how can I be sure that
there’s not some serious performance flaw or misconfiguration of my system?

There used to be a “network administrition document” covering QNET,
but I am not sure if it is still online (and up to date > :slight_smile:

I’ve been avidly following this thread since I haven’t been able to set
a domain name on my system. When I run phlip - it refuses to allow
setting a domain. If I edit /etc/net.cfg and edit the domain line - phlip
deletes the line.

Yes, because of phlip, I think, don’t be under photon to modify this file.

About your problem with the domain, I think - not really sure - that it’s normal if
you configure your tcp/ip stack to use a dhcp server. As says Xiaodan, if you want
to set it, look at Devices tab and choose ‘manual’.

…

Regards,
Alain.

13

Ok, my domain name is working, but I’m doing it this way:

io-net is starting on my primary PC for tcp/ip as it defaulted when it
installed. Everything works including my dhcp. Then in the rc.d directory
I created a rc.local file (as was suggested in another thread in this
newsgroup - it ain’t documented anywhere that I can find) and included this
line:

mount -Tio-net -o bind=ether,host=master,mount=/net:pig.net
/lib/dll/npm-qnet.so

That line mounts the qnet protocol to io-net with the machine name of
“master” and the domaing of “pig.net” (don’t ask). Works fine for me, and
in fact my net.cfg file has the domain as “pig.net” and the line has not
been deleted even though that is not the name of the domain that my dhcp
server uses.

Couple of links that you’ve probably already seen but I’ll post just in
case:

http://support.qnx.com/support/docs/neutrino_qrp/technotes/native-net.html

http://staff.qnx.com/~docs/network.html#998599

–

Jason A. Farqué
CDI at http://www.pigging.com/


I’ve been avidly following this thread since I haven’t been able to set
a domain name on my system. When I run phlip - it refuses to allow
setting a domain. If I edit /etc/net.cfg and edit the domain line - phlip
deletes the line.

The file looks like this:

nto network config file v1.2

version v1.2

[global]
hostname shadow
domain
nameserver 24.0.48.33
nameserver 24.0.48.34
route 192.168.1.1 0.0.0.0 0.0.0.0
lookup file bind

[en0]
type ethernet
mode dhcp
dhcp_id shadow
manual_ip 192.168.1.100
manual_netmask 255.255.255.0

Is it some problem with dhcp?

David L. Hawley D.L. Hawley and Associates

14

Jason A. Farque’ <jasonf@pigging.com> wrote in message
news:9g7vlu$9ko$1@inn.qnx.com…

xtang is my friend.

My qnet network fired up and is now showing other nodes in the /net
directory, thanks for the help. Apparently it was the conflicting node
names that was causing problems - why is there no warning or error on

slogger - sloginfo ???

something such as this? One interesting thing is that on my embedded PC

15

Previously, Joerg Kampmann wrote in qdn.public.qnxrtp.os:

hi,

under QNX 4 when entering “su otherUser” this user’s shell sourced the
file which was in the environment variable ENV (e.g. export
ENV=~/.kshrc) - It seems that this is no more the case with 6.1.0 -

It is very helpful because then within the .kshrc I changed the LOGNAME
variable and could use it e.g. within a export
PS1=’$LOGNAME@$(hostname):$PWD->’ - and I immediately could see what
user (normal or superuser) was working …

How about something like:

PS1=’id -nu@$(hostname):$PWD->’


It should be easy to fix - or ?
Anybody aware of this stuff ?

cheers Jörg

====================================================================> Dr. Jörg Kampmann - IBK-Consult for Real-Time and Embedded Systems
D-31228 Peine - Tel.:+49-177-276-3140 - Fax: +49-5171-13385
http://www.ibk-consult.de
===== QNX is the better Choice for Real-Time: > http://www.qnx.com > ===

–
David L. Hawley D.L. Hawley and Associates

16

this is it!!
thanks
Jörg

David Hawley schrieb:

Previously, Joerg Kampmann wrote in qdn.public.qnxrtp.os:
hi,

under QNX 4 when entering “su otherUser” this user’s shell sourced the
file which was in the environment variable ENV (e.g. export
ENV=~/.kshrc) - It seems that this is no more the case with 6.1.0 -

It is very helpful because then within the .kshrc I changed the LOGNAME
variable and could use it e.g. within a export
PS1=’$LOGNAME@$(hostname):$PWD->’ - and I immediately could see what
user (normal or superuser) was working …

How about something like:

PS1=’id -nu@$(hostname):$PWD->’


It should be easy to fix - or ?
Anybody aware of this stuff ?

cheers Jörg

=====================================================================
Dr. Jörg Kampmann - IBK-Consult for Real-Time and Embedded Systems
D-31228 Peine - Tel.:+49-177-276-3140 - Fax: +49-5171-13385
http://www.ibk-consult.de
===== QNX is the better Choice for Real-Time: > http://www.qnx.com > ====

\

David L. Hawley D.L. Hawley and Associates

–

Dr. Jörg Kampmann - IBK-Consult for Real-Time and Embedded Systems
D-31228 Peine - Tel.:+49-177-276-3140 - Fax: +49-5171-13385
http://www.ibk-consult.de
===== QNX is the better Choice for Real-Time: http://www.qnx.com ====