Hi,
I have created a boot image that provides IDE, FTP, WWW and Telnet services.
The target is a DSP Design TB400 PC104 CPU.
The image does not include any 3rd party applications and no code of my own.
The problem is that after a day or so the entire system locks up, there are
4 shells running and all are completely inaccessible, the keyboard dies,
everything gone, the only option is to reboot.
The next step is to take bits out of the image to ilsolate the problem, this
will take some time, any ideas anyone ?
Also I’ve noticed that just typing pidin results in the last process ID
(pidin) being shown as ever increasing by large amounts…in a short period
the PID will jump by 50000. Is this normal ?
My feeling is that the system is running out of resources and ultimately has
nothing left, any ideas on trapping this?
Thanks,
–
Simon Platten, Senior Software Engineer
VT Controls
Bessemer Way, Harfreys Industrial Estate, Great Yarmouth, Norfolk, NR31 0LX.
Tel: (0)1493 668811 Fax: (0)1493 651137
Email: simon.platten@vtcontrols.co.uk
Any views or opinions expressed are solely those of the author and do not
necessarily
represent those of VT Group, its holding company or any of its subsidiaries
or associates.
I have created a boot image that provides IDE, FTP, WWW and Telnet
services.
The target is a DSP Design TB400 PC104 CPU.
The image does not include any 3rd party applications and no code of my
own.
The problem is that after a day or so the entire system locks up, there
are
4 shells running and all are completely inaccessible, the keyboard dies,
everything gone, the only option is to reboot.
The next step is to take bits out of the image to ilsolate the problem,
this
will take some time, any ideas anyone ?
Run a shell on the console at high priority, or have a high priority process
blink an LED and see if that stops. If not, you probably have a run away
process starving others out. You mentioned IDE, what version of 6.2 are you
using? Does the board stop responding to a network ping?
Also I’ve noticed that just typing pidin results in the last process ID
(pidin) being shown as ever increasing by large amounts…in a short
period
the PID will jump by 50000. Is this normal ?
That’s fine, and expected.
My feeling is that the system is running out of resources and ultimately
has
nothing left, any ideas on trapping this?
The system shouldn’t become completely unresponsive - certain services may
not operate gracefully under low memory conditions but it should still be
recoverable. You can start logging pidin and hopefully it will give you
some idea as to what was happening at the time. If you’ve got 6.2PE you
could use the system monitoring features on the host, to keep an eye on the
target.
-Adam
Hi, thanks for the response.
When it does lock, I can’t ping it or indeed do anything, its locked up in
every sense. I am using Momentics 6.2.0 with all the patches applied and
QNX to create the boot image. I do have the PE version. I’ll look into
the system monitor features, still new to Momentics.
Thank you,
Simon
“Adam Mallory” <amallory@qnx.com> wrote in message
news:atab99$qpg$1@nntp.qnx.com…
I have created a boot image that provides IDE, FTP, WWW and Telnet
services.
The target is a DSP Design TB400 PC104 CPU.
The image does not include any 3rd party applications and no code of my
own.
The problem is that after a day or so the entire system locks up, there
are
4 shells running and all are completely inaccessible, the keyboard dies,
everything gone, the only option is to reboot.
The next step is to take bits out of the image to ilsolate the problem,
this
will take some time, any ideas anyone ?
Run a shell on the console at high priority, or have a high priority
process
blink an LED and see if that stops. If not, you probably have a run away
process starving others out. You mentioned IDE, what version of 6.2 are
you
using? Does the board stop responding to a network ping?
Also I’ve noticed that just typing pidin results in the last process ID
(pidin) being shown as ever increasing by large amounts…in a short
period
the PID will jump by 50000. Is this normal ?
That’s fine, and expected.
My feeling is that the system is running out of resources and ultimately
has
nothing left, any ideas on trapping this?
The system shouldn’t become completely unresponsive - certain services may
not operate gracefully under low memory conditions but it should still be
recoverable. You can start logging pidin and hopefully it will give you
some idea as to what was happening at the time. If you’ve got 6.2PE you
could use the system monitoring features on the host, to keep an eye on
the
target.
-Adam