Cumulus 1.8.8 hangs

[hendrikr]

Hi Steve,

I'm using your fabulous Cumulus 1.8.8 together with PCE-FWS 20 (FineOffset) equipment. Everything is working absolutely smooth and fine, excepting that today the second time the akquiring of data seems to hang. A couple of days ago this occurred the first time and I had to restart the PC to get it working again. Sending data to the internet and copying within the network is working, but data transmitted are those gathered hours ago. Please find attached the Feb10log.txt, latest diag file(cumulus.009) and some screenshots shoing the data akquisition has stopped around 10:16 this morning. Is there any other information I can provide for analysing this behaviour?

Kind regards
Hendrik

[steve]

I can't tell whether the thread which reads the data has stopped, or whether it's just getting the same data every time. Could you please turn on the debug log, wait 10 minutes or so, and then attach the debug.log file.

When you say you had to restart the PC to get it going again - I assume you tried just restarting Cumulus, so what happened when you did that?

[hendrikr]

thanks for the speedy answer. I'm not at home at the moment (got the files via VPN, unfortunately I haven't turned on virtual desktop), therefore have to post the debug log tonight. Last time I restarted Cumulus w/o restarting the PC the result was the same, i.e. data were repeatedly sent instead of the latest ones. Data on the local display of the FWs20 are current.

[hendrikr]

Steve,
attached the debug log. As expected, not much to read from USB ...

Kind regards
Hendrik

[steve]

Yes, it's trying to read the data but nothing's coming back. This explains why you have to restart your PC for it to start working again. It looks like a problem with your station, or some sort of USB driver problem...

[steve]

It had been running for about 6 days when the problem occurred - do you know how long it had been running when it happened before?

[hendrikr]

as a rough guess I could say also about six days. Looks like some buffer overflow within the USB driver or something similar. I will restart that PC thing now and see how long it lasts this time. I can also install some of M$'s powerful PowerToys to see if there is some DLL stopping activity, maybe that helps ...
If everything else fails, I can reboot that machine every other day

[steve]

If it gets near to six days, please turn on the debug log again (best not to leave it on, it may cause problems, and the file will be huge), and then we might see something at the time the problem occurs.

[hendrikr]

ok, I'll do that and come back to you then.

Kind regards and many thanks for your efforts
Hendrik

[hendrikr]

Unfortunately I was not quick enough. Cumulus stopped akquiring data this morning, after a little more than five days. I restarted and will turn on debugging when four days are over, i.e. next friday.
What is puzzling me a little, that I have to start Cumulus twice to get rid of the "Downloading Data" message box, although even on the first time data are gathered and shown correctly, only the message box refuses to disappear and the progress bar shows nothing.
After start the "New Record" LED will stay off for some hours, but "Last Data read at ..." status line shows that records are read and also everything else works fine. After some hours, the LED starts working. When Cumulus stops (after five days) the status line constantly shows the time the last record was handled correctly, but the "New Record" LED continues flashing.

[steve]

Unfortunately I was not quick enough. Cumulus stopped akquiring data this morning, after a little more than five days.

So it's not after a specific time that this happens. I have a feeling that when you finally do get the debug log information that it won't help anyway, as your station simply stops responding... but we'll see.

What is puzzling me a little, that I have to start Cumulus twice to get rid of the "Downloading Data" message box, although even on the first time data are gathered and shown correctly, only the message box refuses to disappear and the progress bar shows nothing.

It must be having problems reading the data from the logger. Or are you saying that it continues to display 'live' data behind the 'downloading data' box?

After start the "New Record" LED will stay off for some hours, but "Last Data read at ..." status line shows that records are read and also everything else works fine. After some hours, the LED starts working. When Cumulus stops (after five days) the status line constantly shows the time the last record was handled correctly, but the "New Record" LED continues flashing.

I think you are misunderstanding what the 'new record' LED means: http://wiki.sandaysoft.com/a/FAQ#My_.E2 ... d_flashing

[beteljuice]

"New Record" light has nothing to do with data load. http://wiki.sandaysoft.com/a/FAQ#My_.E2 ... d_flashing

Click on it at any time - short cut to "All Time Records"

[hendrikr]

Thank you for the pointers, I really misinterpreted "New Record" as "New Record received". My apologies.

So it's not after a specific time that this happens. I have a feeling that when you finally do get the debug log information that it won't help anyway, as your station simply stops responding... but we'll see.

Unfortunately you are right, Things stopped working yesterday after three hours runtime. I did a restart this morning and now it's running since then.

It must be having problems reading the data from the logger. Or are you saying that it continues to display 'live' data behind the 'downloading data' box?

That's exactly what it does, actual data are being shown behind the message box.

[steve]

It must be having problems reading the data from the logger. Or are you saying that it continues to display 'live' data behind the 'downloading data' box?

That's exactly what it does, actual data are being shown behind the message box.

That's strange. Basically, it opens the window, downloads and processes the archive data, then closes the window and goes into 'live' mode. I guess it's possible there's some error path where the window doesn't get closed. The diags file for that run might give some clues.

[hendrikr]

I moved the message box into some corner where it did not bother
As far as I remember it later disappeared, I will have a look tonight, when I'm home again. Maybe it just takes longer time for WM_CLOSE, whyever. I'm running a Java engine on the same machine, maybe this could also be some cause for strange behaviours (Sun and Microsoft not always cooperate smoothly, right?)
The diag file attached shows "PleaseWaitForm.Close executed" shortly after application start, so I am sure the problem for delayed closing is somewhere inside Windows. It's not really a problem, I'll take a look at the time it takes for closing next time.

I will install another logger the next time data akquisition stops - without and with rebooting the PC. I have two of those logger devices running. Maybe it's really a problem of that specific hardware.