Hi
I have QNX 6.2.0 running on a fast x86 platform. I have an application that
uses TCP/IP that runs on it and communicates with a server. After starting
the run at night, the system crashes after approximately 12 hours.
This is what happens:
After about 12 hours of normal run, the application log (my program’s log)
says that it timed out on the communication and so shuts down normally. No
problem there. But then the machine goes into a crashed state. Trying to
connect to it (like ftp or telnet) remotely fails. But pinging it works.
Also the machine stops responding locally too…
I tried the run with and without photon (my app doesn’t need photon). It
crashes both times. So I guess photon is ruled out.
Also since my app shuts down normally, I guess the app isn’t to blame either
Without photon running, typing anything at the shell prompt after the crash
shows a simple “Memory Fault” error.
Then on only way to recover is to cycle power.
Any help would be appreciated…
Thanks,
Santosh
Sounds like you may be running out of some resource. Try
monitoring with pidin, netstat etc as it’s running but before
you get in the state below to see if anything stands out.
-seanb
Santosh Patil <Santosh.Patil@gehh.ge.com> wrote:
Hi
I have QNX 6.2.0 running on a fast x86 platform. I have an application that
uses TCP/IP that runs on it and communicates with a server. After starting
the run at night, the system crashes after approximately 12 hours.
This is what happens:
After about 12 hours of normal run, the application log (my program’s log)
says that it timed out on the communication and so shuts down normally. No
problem there. But then the machine goes into a crashed state. Trying to
connect to it (like ftp or telnet) remotely fails. But pinging it works.
Also the machine stops responding locally too…
I tried the run with and without photon (my app doesn’t need photon). It
crashes both times. So I guess photon is ruled out.
Also since my app shuts down normally, I guess the app isn’t to blame either
Without photon running, typing anything at the shell prompt after the crash
shows a simple “Memory Fault” error.
Then on only way to recover is to cycle power.
Any help would be appreciated…
Thanks,
Santosh
Hello, Sean!
SB> Sounds like you may be running out of some resource. Try
SB> monitoring with pidin, netstat etc as it’s running but before
SB> you get in the state below to see if anything stands out.
I have the same problem. When I trying to compile the simply application,
like that with huge static array:
static unsigned char a[10000000];
int main()
{
return 0;
}
gcc torturing my HDD approx. 20 minutes then it “Memory faulted”, after that
I can’t even run shutdown, I’m always get “Memory fault”.
I have 384Mb RAM, so I think problem not in memory count.
With best regards, Mike Gorchak.
I just tried your static array case on a 6.2 system (I even went up to 1000
million on the array) and it only took a second or two to compile. In fact,
the compile time didn’t seem to have any relation to the size of the array.
There’s definitely something funny going on here. Give us some more time to
ponder some other possible tests/questions for you. If you can think of any
other info that might be relevant…
cheers,
Kris
“Mike Gorchak” <mike@malva.ua.remove.this.for.no.spam> wrote in message
news:alp8ra$gph$1@inn.qnx.com…
Hello, Sean!
SB> Sounds like you may be running out of some resource. Try
SB> monitoring with pidin, netstat etc as it’s running but before
SB> you get in the state below to see if anything stands out.
I have the same problem. When I trying to compile the simply application,
like that with huge static array:
static unsigned char a[10000000];
int main()
{
return 0;
}
gcc torturing my HDD approx. 20 minutes then it “Memory faulted”, after
that
I can’t even run shutdown, I’m always get “Memory fault”.
I have 384Mb RAM, so I think problem not in memory count.
With best regards, Mike Gorchak.
Hello, Kris!
You wrote on Thu, 12 Sep 2002 09:12:01 -0400:
KW> I just tried your static array case on a 6.2 system (I even went up to
KW> 1000 million on the array) and it only took a second or two to compile.
KW> In fact, the compile time didn’t seem to have any relation to the size
KW> of the array. There’s definitely something funny going on here. Give
KW> us some more time to ponder some other possible tests/questions for
KW> you. If you can think of any other info that might be relevant…
Sorry, I’m slipped some important information, I’m using constraint to
section, like that:
char overlay[10000000] attribute((section(".overlay")));
and in “main” function do some silly access to array like a=overlay[100];
for prevent compiler optimization.
I will try one more my test program at home … at evening today.
With best regards, Mike Gorchak.