Problems with Cumulus closing and high rainfall totals

[Steve Willington]

Steve,

Thanks for your help - I think the rainfall stats are sorted now and have started to keep the website up to date.

I have had Cumulus stop working again this evening though (was hoping that the WU update removal had stopped that). Diags are attached as requested.

[steve]

There's nothing in the diags files to tell me the cause of the crash. And the fact that no bugreport.txt file was produced suggests that it's some kind of Windows system error.

Perhaps if you could tell me exactly what happens when it crashes? You were rather vague about the error message that Windows was showing you. Windows will normally give error codes when it forces an application to close.

[steve]

A thought occurred to me - what value are you using in cumulus.ini for VP2SleepInterval?

[Steve Willington]

Steve,

Not ignoring this request - been at work again. However the crashes/freezes are much less frequent now (only the one time so far as mentioned above).

I'll check the VP2sleepinterval value tomorrow and post it up.

[Steve Willington]

Well it has taken some time for Cumulus to stop working and be able to capture that message. It is attached.

As far as I can tell it definitely WU update related. I was getting this on average 4 or 5 times a day, but since switching off WU it is now just once a week.

[steve]

Look in the Windows Event Viewer. There should be an event in there for the crash (look for an Application Error at the time of the crash), with some error codes.

[Steve Willington]

Steve,

Is this what you are looking for?

[steve]

Googling AppHangB1 shows a number of people having this issue with all kinds of software (including commercial software); so far theories seem to be either a resource issue on the PC or a bug in Windows. It seems to happen when Windows decides rightly or wrongly that the application has hung.

What have you set for VP2SleepInterval in cumulus.ini?

[Steve Willington]

VP2SleepInterval=1100

[steve]

In that case I've run out of ideas (not that I had really had any to begin with). Windows on your PC is deciding that it needs to force Cumulus to close for some reason. I'm not aware of anyone else having this problem.

[Steve Willington]

Hi Steve,

The problem is not a big one now that WU update is switched off. It occurs about once every 24 hours of running so will only be a concern when I am away for any length of time and wanted to leave the AWS unattended and updating the web (obviously not on 1 min intervals...).

Thanks for looking into it.

[Steve Willington]

Steve, I have another one of those rainfall issues. On starting Cumulus today I got the message "Cumulus did not initialise correctly" (or very similar). Apart from today and yesterday values all other values were zero. I closed Cumulus and copied the backup files into data. I started Cumulus again, there appears to be no data gap, but the last data line had a value of 270mm rain in it.

I tried correcting using toolbox but the 270mm re-appears as soon as I restart Cumulus - from the last close down point only 5 or 10 mins earlier.

BTW dayfile.txt does not have a 'today' line yet - so that maybe the cause? Otherwise any other ideas?

[steve]

Please zip up the diags folder and attach it.

BTW dayfile.txt does not have a 'today' line yet

That's right, the entries in dayfile.txt are summaries of each day, and 'today' hasn't finished yet, so there can be no summary.

[Steve Willington]

Diags folder attached. Don't forget I have tried to 'fix' this using 'toolbox' - I'll have another go in a minute now that dayfile.txt has yesterdays data in it.

[steve]

On your first run yesterday, the Davis DLL couldn't contact the station (or Virtual VP in your case):

31/07/2011 19:01:03.483 : VP2: InitStation_V, res = -32701

On your second run, the connection worked, but Cumulus somehow thought that the last time it ran (from today.ini) was 29/07/2011 22:30:00, but it was actually LastUpdateTime = 30/07/2011 21:18:00. Perhaps you had been editing today.ini to change the timestamp - I can't see how the timestamp could have changed otherwise. This change of date meant that the data in the logger was not enough to fill the gap - it only went back to 30/07/2011 00:25:00

31/07/2011 19:04:54.603 : VP2: Downloading history from 29/07/2011 22:30:00
31/07/2011 19:04:54.603 : y=2011 m=7 d=29 h=22 m=30
31/07/2011 19:05:02.684 : VP2: DownloadData_V = 2559
31/07/2011 19:05:02.684 : VP2: GetNumberOfArchiveRecords_V = 2559
31/07/2011 19:05:02.684 : VP2: Number of history records = 2559
31/07/2011 19:05:02.684 : Logger entry zero: Y = 2011, M = 7, D = 30, H = 0, M = 25
31/07/2011 19:05:02.684 : VP2: Earliest timestamp 30/07/2011 00:25:00

In 1.9.1. Cumulus resets the rain counter because it can't cope with the data not being sufficient to fill the gap.

31/07/2011 19:05:02.996 : Rain counter = 0

I changed this behaviour in build 1004 of 1.9.2, so it at least doesn't reset the rain counter. The data still won't be very good, though, because of the missing data caused by rewinding too far.