once more i have to report my problem here, because now it’s fatal for our
system!
Our System now runs QNX 6.2. It is a small DIN-Rail mounted PC with compact
flash card as harddrive.
For test only photon is started and postgrsql gets evey second a
testlog-entry from a script.
A second script runs in a loop: logs datetime, does a pidin on screen and
waits two seconds.
Thats my test environment.
After an unspecified time (a couple of hours) the second script hangs up,
because pidin hangs.
Manually testings shows this:
ps shows a couple of processes (may be all? i don’t know) but never
terminates any more
mount shows the mount-points but: does not terminate.
ls /proc finishes, but ls -l /proc dies without printing a single line …
I am hopeful that anyone can ask a good question or tells me what is going
on ?!?
FATAL is:
12 hours later postgresql stop its function because there are no free client
connections left…
But if the postmaster dies because of this: My system dies also…
I found an entry in the newsgroups where it should have something to do with
a programming failure. I can’t remember exactly. But there is nothing
running that is programmed by myself… (except the testscripts
AND:
Mysterios is that this bug was never seen on my standard desktop system with
normal harddrive (i don’t know it the compact flah/harddrive thing has
something to do with it…?)
It seems for me, that it is a problem of QNX not of postgreSQL because of
the last point and because of the VERY strange behavior of QNX while the bug
is switched on.
once more i have to report my problem here, because now it’s fatal for
our
system!
Our System now runs QNX 6.2. It is a small DIN-Rail mounted PC with
compact
flash card as harddrive.
For test only photon is started and postgrsql gets evey second a
testlog-entry from a script.
A second script runs in a loop: logs datetime, does a pidin on screen and
waits two seconds.
Thats my test environment.
After an unspecified time (a couple of hours) the second script hangs up,
because pidin hangs.
Manually testings shows this:
ps shows a couple of processes (may be all? i don’t know) but never
terminates any more
mount shows the mount-points but: does not terminate.
ls /proc finishes, but ls -l /proc dies without printing a single line
…
I am hopeful that anyone can ask a good question or tells me what is going
on ?!?
FATAL is:
12 hours later postgresql stop its function because there are no free
client
connections left…
But if the postmaster dies because of this: My system dies also…
I found an entry in the newsgroups where it should have something to do
with
a programming failure. I can’t remember exactly. But there is nothing
running that is programmed by myself… (except the testscripts >
AND:
Mysterios is that this bug was never seen on my standard desktop system
with
normal harddrive (i don’t know it the compact flah/harddrive thing has
something to do with it…?)
It seems for me, that it is a problem of QNX not of postgreSQL because of
the last point and because of the VERY strange behavior of QNX while the
bug
is switched on.
i am very grateful for your answer, but what makes you think in this
direction?
The second problem (the “leaking connections”) should have a cause. But i
have no idea where to look.
Why does QNX behave so strange? If i am no longer able to kill processes, is
there a possibility that postgres also cannot end all processes (that ones
which it has to kill)?
once more i have to report my problem here, because now it’s fatal for
our
system!
Our System now runs QNX 6.2. It is a small DIN-Rail mounted PC with
compact
flash card as harddrive.
For test only photon is started and postgrsql gets evey second a
testlog-entry from a script.
A second script runs in a loop: logs datetime, does a pidin on screen
and
waits two seconds.
Thats my test environment.
After an unspecified time (a couple of hours) the second script hangs
up,
because pidin hangs.
Manually testings shows this:
ps shows a couple of processes (may be all? i don’t know) but never
terminates any more
mount shows the mount-points but: does not terminate.
ls /proc finishes, but ls -l /proc dies without printing a single line
…
I am hopeful that anyone can ask a good question or tells me what is
going
on ?!?
FATAL is:
12 hours later postgresql stop its function because there are no free
client
connections left…
But if the postmaster dies because of this: My system dies also…
I found an entry in the newsgroups where it should have something to do
with
a programming failure. I can’t remember exactly. But there is nothing
running that is programmed by myself… (except the testscripts >
AND:
Mysterios is that this bug was never seen on my standard desktop system
with
normal harddrive (i don’t know it the compact flah/harddrive thing has
something to do with it…?)
It seems for me, that it is a problem of QNX not of postgreSQL because
of
the last point and because of the VERY strange behavior of QNX while the
bug
is switched on.
}