Cumulus Support

I've just recently started using Cumulus and firstly I must say that it is great! Thanks Steve.

I'm using a WH1081 weather station but I am having a few issues with the Cumulus software which I would like to query.

1. The Cumulus software initially behaves very well with the new record light flashing every few seconds. If I leave the system alone then it will run for around 10 hours approximately before the error light will flash shortly after each new record light flash. When this happens one of two things happen :
a) Everything seems to be logged correctly with the system seemingly logging correct data.
b) The system begins logging identical data - it seems to be logging a set of data results from approximately the time the red error light started flashing.
(The only way I can recover this is to backup the folder, then copy one of the previous backup directories data over the data folder and restart Cumulus to reload the missing data from the WH1081 device - it might be nice to have a force reload option ...)

2. If I interact with the software then occasionally (and I haven't been able to pinpoint a pattern to this yet) it will crash. I first get what looks like a Borland crash screen which is displayed for a second or two and then I get the Windows 'submit a error' dialog box. I can supply the windows crash data Steve next time it happens if you like with a description of what I have been doing.

3. I have also noticed that if I am using the menus then the application appears to freeze when the system is uploading to the website - this is the 'Uploading' message is displayed in the status bar.

I'm not sure if these problems are related to the fact that I'm using a WH1081 or not.

For information, I currently have the system saving data every minute, uploading the website pages each minute, and the realtime enabled updating the data every 5 seconds.

It sounds like it is losing the USB connection - this would account for logging the same data repeatedly, and also probably the crash. What you're describing for the crash sounds very odd - you should get the 'madExcept' error screen - looks a bit like the one on this page - http://www.madshi.net/madExceptDescription.htm and you should be able to save the bugreport.txt file and upload it here.

Regarding the flashing error light - I'm not sure from your description whether you realise that it does actually 'flash' (about once a second, I think) once activated; it doesn't do a single flash to indicate a single error. It will continue to flash until you click on it, when it should show what the error was - it's normally an ftp error. It's the same with the 'new record' light, it just starts flashing once a new record has been broken, and doesn't stop until you display the all-time records screen (e.g. by clicking on the light).

You could try increasing the upload interval to say 5 minutes and see if that has any effect on the problem.

Steve, many thanks for the quick response.

steve wrote:It sounds like it is losing the USB connection - this would account for logging the same data repeatedly, and also probably the crash. What you're describing for the crash sounds very odd - you should get the 'madExcept' error screen - looks a bit like the one on this page - http://www.madshi.net/madExceptDescription.htm and you should be able to save the bugreport.txt file and upload it here.

Yes, that is the dialog I get for about 1 to 2 seconds (I said the Borland crash screen as when I was coding in Borland sometime ago it had the same/very similar crash screen). Perhaps the reason it goes quickly to the Windows dialog is due to the fact that I am running under WinXP in a user restricted account.

steve wrote:Regarding the flashing error light - I'm not sure from your description whether you realise that it does actually 'flash' (about once a second, I think) once activated; it doesn't do a single flash to indicate a single error. It will continue to flash until you click on it, when it should show what the error was - it's normally an ftp error. It's the same with the 'new record' light, it just starts flashing once a new record has been broken, and doesn't stop until you display the all-time records screen (e.g. by clicking on the light).

Ah I see - ok. During my trials I was starting the system fresh with no data and seeing the 'New Record' light flashing one a second. I assumed this was as it retrieve the new data record, but I now understand this was not a 'new data record' but a new minimum or maximum record since it was new data. Ah - thanks for that - suppose I should have read the help more closely.

Right - I now understand the working of the 'new record' and 'error' light. I can click on these to see the all time records and an error report.

steve wrote:You could try increasing the upload interval to say 5 minutes and see if that has any effect on the problem.

I will do and will see let you know my findings.


One other thing I have found - I have the 'data log interval' set to 1 minute which works fine however if I shut Cumulus down for a period of time and then restart it then the historical data it catches up is only logged using the default 5 minute interval and once it gets back to logging in realtime it obeys the setting in the 'data log interval'.

maumoon wrote:One other thing I have found - I have the 'data log interval' set to 1 minute which works fine however if I shut Cumulus down for a period of time and then restart it then the historical data it catches up is only logged using the default 5 minute interval and once it gets back to logging in realtime it obeys the setting in the 'data log interval'.

This is because you have your station logger set to 5 minute intervals. Cumulus just logs all of the entries from the logger; there wouldn't be any point in Cumulus logging at 1-minute intervals using data created at 5-minute intervals.

steve wrote:This is because you have your station logger set to 5 minute intervals. Cumulus just logs all of the entries from the logger; there wouldn't be any point in Cumulus logging at 1-minute intervals using data created at 5-minute intervals.

Now that makes perfect sense Steve.