Hello,
I’m using the utility “on” to start processes on an embedded system with no
keyboard and monitor. This works great as long as the process which I start
doesn’t access any files or devices (serial ports in this case). It appears
as if the remote machine is using the file system structure of my dev. PC,
as it complains about certain files not existing when I run programs or
about devices not existing, as my dev. PC doesn’t have as many serial ports
as the embedded system. An example: I start a test program located on
machine B from my dev. PC, machine A. Machine B has a /dev/ser3 device, but
Machine A only has /dev/ser1 and /dev/ser2. When I start the test program
from A, the program complains that the device doesn’t exist. If, however, I
add the same program to a start script called during the boot process, the
program runs fine.
Is this actually what I’m observing or is there another problem? If it is a
path resolution problem, how do I force the remote system to only use its
own, local file system?
Thanks in advance.
Jason
That is correct behaviour for “on -n”. In the next QNX release you will be
able to do “on -f” which will set the root of the process to be on the remote
node. You can also use explicit paths like /net/machine/dev/ser1 so you
can be protected from the node you are running on verses the node on which
you want to access resources.
chris
Jason Wycoff <jason@socratec.de> wrote:
Hello,
I’m using the utility “on” to start processes on an embedded system with no
keyboard and monitor. This works great as long as the process which I start
doesn’t access any files or devices (serial ports in this case). It appears
as if the remote machine is using the file system structure of my dev. PC,
as it complains about certain files not existing when I run programs or
about devices not existing, as my dev. PC doesn’t have as many serial ports
as the embedded system. An example: I start a test program located on
machine B from my dev. PC, machine A. Machine B has a /dev/ser3 device, but
Machine A only has /dev/ser1 and /dev/ser2. When I start the test program
from A, the program complains that the device doesn’t exist. If, however, I
add the same program to a start script called during the boot process, the
program runs fine.
Is this actually what I’m observing or is there another problem? If it is a
path resolution problem, how do I force the remote system to only use its
own, local file system?
Thanks in advance.
Jason
\
–
Chris McKillop <cdm@qnx.com> “The faster I go, the behinder I get.”
Software Engineer, QSSL – Lewis Carroll –
http://qnx.wox.org/
How are you calling on?m Make sure you are using on -f to use the remote
file system as well.
-Peter
Jason Wycoff <jason@socratec.de> wrote:
Hello,
I’m using the utility “on” to start processes on an embedded system with no
keyboard and monitor. This works great as long as the process which I start
doesn’t access any files or devices (serial ports in this case). It appears
as if the remote machine is using the file system structure of my dev. PC,
as it complains about certain files not existing when I run programs or
about devices not existing, as my dev. PC doesn’t have as many serial ports
as the embedded system. An example: I start a test program located on
machine B from my dev. PC, machine A. Machine B has a /dev/ser3 device, but
Machine A only has /dev/ser1 and /dev/ser2. When I start the test program
from A, the program complains that the device doesn’t exist. If, however, I
add the same program to a start script called during the boot process, the
program runs fine.
Is this actually what I’m observing or is there another problem? If it is a
path resolution problem, how do I force the remote system to only use its
own, local file system?
Thanks in advance.
Jason
Thanks for the tips, guys. I’m using the online documentation at QSSL and
didn’t see the -f option for “on” described at:
http://qdn.qnx.com/support/docs/neutrino_2.11_en/utilities/o/on.html
Is there a newer version of the online documentation which includes the most
current options?
Thanks,
Jason
“Peter Graves” <pgraves@qnx.com> wrote in message
news:aam2hp$7em$3@nntp.qnx.com…
How are you calling on?m Make sure you are using on -f to use the remote
file system as well.
-Peter
Jason Wycoff <> jason@socratec.de> > wrote:
Hello,
I’m using the utility “on” to start processes on an embedded system with
no
keyboard and monitor. This works great as long as the process which I
start
doesn’t access any files or devices (serial ports in this case). It
appears
as if the remote machine is using the file system structure of my dev.
PC,
as it complains about certain files not existing when I run programs or
about devices not existing, as my dev. PC doesn’t have as many serial
ports
as the embedded system. An example: I start a test program located on
machine B from my dev. PC, machine A. Machine B has a /dev/ser3 device,
but
Machine A only has /dev/ser1 and /dev/ser2. When I start the test
program
from A, the program complains that the device doesn’t exist. If,
however, I
add the same program to a start script called during the boot process,
the
program runs fine.
Is this actually what I’m observing or is there another problem? If it
is a
path resolution problem, how do I force the remote system to only use
its
own, local file system?
Thanks in advance.
Jason
Yes, there is a newer version. However, it isn’t generally available yet.
The -f option for ‘on’ is described in the docs that will accompany the
next release. When the new versions completes its beta test cycle and
we make it generally available, we’ll update the docs on the web site.
The option is
-f nodename
and it will cause the process to run on the remote node you name, using the
pathname space of the remote node.
“Jason Wycoff” <jason@socratec.de> wrote in message
news:aam8ha$3u5$1@inn.qnx.com…
Thanks for the tips, guys. I’m using the online documentation at QSSL and
didn’t see the -f option for “on” described at:
http://qdn.qnx.com/support/docs/neutrino_2.11_en/utilities/o/on.html
Is there a newer version of the online documentation which includes the
most
current options?
Thanks,
Jason
“Peter Graves” <> pgraves@qnx.com> > wrote in message
news:aam2hp$7em$> 3@nntp.qnx.com> …
How are you calling on?m Make sure you are using on -f to use the remote
file system as well.
-Peter
Jason Wycoff <> jason@socratec.de> > wrote:
Hello,
I’m using the utility “on” to start processes on an embedded system
with
no
keyboard and monitor. This works great as long as the process which I
start
doesn’t access any files or devices (serial ports in this case). It
appears
as if the remote machine is using the file system structure of my dev.
PC,
as it complains about certain files not existing when I run programs
or
about devices not existing, as my dev. PC doesn’t have as many serial
ports
as the embedded system. An example: I start a test program located on
machine B from my dev. PC, machine A. Machine B has a /dev/ser3
device,
but
Machine A only has /dev/ser1 and /dev/ser2. When I start the test
program
from A, the program complains that the device doesn’t exist. If,
however, I
add the same program to a start script called during the boot process,
the
program runs fine.
Is this actually what I’m observing or is there another problem? If
it
is a
path resolution problem, how do I force the remote system to only use
its
own, local file system?
Thanks in advance.
Jason
\