Cumulus Support

Hello,

I hope somebody can help me here. I've been running Cumulus for four months without any problems but noticed today that "This month's highs and lows" page is still showing (predominantly) data from September.

http://www.spireweather.co.uk/records/r ... his-month/

How do I resolve this please? Thanks in advance for any help.

Simon

You didn't get a midnight rollover on the 30th Sept/1st October, so the monthly highs and lows didn't get reset. If you zip up the diags folder and attach it, I may be able to see why that happened. You can use Edit | This month's records to fetch the correct values from dayfile.txt and/or the monthly log file.

Thank you Steve. I have attached a zip of my Diags folder.

Also I checked and there is an entry for 23:55 on the 30th September and 00:00 on 1st October in the logfiles.

Regards,

Unfortunately you have restarted Cumulus more than ten times since the start of the month, so I can't say what happened then (the diags file no longer exists).

I auto restart once a day at 1:30am which helps keep my Yawcam software functioning properly. Perhaps I'll reduce that to every other day.

Did you notice my modified post above re the logfile entries?

SpireWeather wrote:Did you notice my modified post above re the logfile entries?

No, I didn't. It's not a good idea to edit posts because I have no way of knowing that you've done that once I've read it (which I had). The information doesn't really help to say what happened, it just rules out a few possible explanations, and without the diags file I can only speculate - perhaps your PC clock changed at around that time, for example?

I suppose that's possible but I do try and ensure it doesn't update around midnight.

Do you see any problem with my PC being rebooted at 1.30am. I do that for two reasons: -

One to keep Yawcam running smoothly and two to ensure my Twitter updates go out at the times I want.

SpireWeather wrote:Do you see any problem with my PC being rebooted at 1.30am.

No, that should be fine. The 'danger' time is within a few minutes of midnight, and even then Cumulus copes in most situations.

Do those Diags files show anything untoward happening on the 6th by any chance, as Cumulus registered 120.6mn erroneous rainfall at the 1.30am restart which took ages to sort out.

Thanks Steve,

Yes; after the restart at 01:26, your station (or the Davis DLL) started saying that the annual total so far was about 567:

06/10/2013 01:29:28.182 : VP2: Setting raintotal from current station value: 567.435974121094

Prior to the restart it had said that it was about 446:

06/10/2013 01:20:01.654 : Writing today.ini, LastUpdateTime = 06/10/2013 01:20:00 raindaystart = 446.799987792969 rain counter = 446.799987792969

It continued to say that it was about 446 until the next restart:

07/10/2013 01:20:01.590 : Writing today.ini, LastUpdateTime = 07/10/2013 01:20:00 raindaystart = 567.400024414063 rain counter = 567.400024414063

after which it started saying that it was about 567 again:

07/10/2013 01:29:30.450 : 01:29:30 ****Rain reset? First chance: raindaystart = 567.400024414063, raintotal = 446.799987792969
07/10/2013 01:29:30.450 : Leaving counter at 446.799987792969
07/10/2013 01:29:30.611 : 01:29:30 ****Rain counter reset confirmed: raindaystart = 567.400024414063, raintotal = 446.799987792969
07/10/2013 01:29:30.611 : Setting raindaystart to 446.799987792969

It was still saying 446 this morning:

08/10/2013 10:05:12.865 : Writing today.ini, LastUpdateTime = 08/10/2013 10:05:01 raindaystart = 446.799987792969 rain counter = 446.799987792969

Restarting Cumulus shouldn't change the annual rain counter in the station, so I don't know why it should change between restarts like that. Cumulus just reads the value from the Davis DLL; perhaps the problem was actually there rather than in the console itself. The discrepancy is approximately the same as the difference between a metric gauge and an imperial gauge, so I would guess there was a miscommunication between the console and DLL about which type of gauge was fitted.

OK thanks for looking. I'll have to keep an eye on that. Have you come across this before?

I don't remember suspecting that the rain gauge type had been read incorrectly, but it's possible that someone has had strange rain figures before and it didn't occur to me that this might be the cause. I'll have a look at the DLL documentation to see if there's any way I can check for this happening.

Great, will follow this thread with interest...