Cumulus Support

I was adding a new file to the internet file upload page. I wanted it to be sent at realtime frequency and I didn't know if it took effect immediately so I ended and restarted Cumulus.

When Cumulus was restarted it went through all the motions normally and I carried on with what I was doing (adding some new stuff to my web site).

After a while I realised that the wind reading was not changing (it is WINDY at the moment). I checked that cumulus was saying it was downloading from the Davis okay (bottom right hand corner of the screen). All seemed okay. So I ended and restarted it again. Same. So I rebooted the PC (I know, stupid, since it was highly unlikely that this would fix it). Still, I rebooted.

After that things went wrong bigtime. Upon restarting Cumulus now, all seemed okay with Cumulus except the Davis consolr lost contact with the ISS (and my Anemometer which has its own transmitter). I got an "R" in the bottome corner of the console.

I decided it was a coincidence and went into setup and out again on the Davis. It found the two transmitters (ISS and Anemometer) but as soon as I put it into normal mode it lost contact again and I got the "R" back.

I rebooted the console (power off and batteries out). It found the transmitters and again, as soon as I exited setup to normal mode it lost them again.

All the time, Cumulus had been sitting there apparently doing nothing.

I ended cumulus and went ina nd out of setup on the Davis. It started working normally. Data being retrieved from ISS and anemom. I left it a couple of minutes and then started Cumulus.

This time Cumulus stopped at the downloading from logger splash screen, the Davis froze, lost contact with the transmitters and displayed "R".

I ended Cumulus, went in and out of setup. Waited for data to start flowing from the sensors again and started Cumulus again. As soon as Cumulus started to download data the Davis lost contact and dirplayed "R".

I repeated the process once more with the same results.

I decided to restore the critical files from the last good backup before all this started (I have cumulus toolbox backing up every hour).

As soon as I did this, Cumulus started normally, downloaded the missing data and has run fine since.

Has anyone seen anything like this before?

Steve, if you are around and are interested, I have a series of log files etc from a backup of cumulus folder that I took just before restoring that I could send if you wish.

I know that the console display can freeze during logger download; I have also observed this with Weatherlink. I think that it is even documented by Davis that this happens. I'm not aware of the console losing contact with the ISS when this happens, but I restart Cumulus infrequently and I can't see the console when I do it.

If you zip up the diags folder and attach it, I'll take a look, but note that Cumulus takes no part at all in the download of the logger data, as with every other operation it's all handled by the Davis DLL.

Note that restarting Cumulus quickly (particularly if the PC clock and station clock are out of sync) can bring on the same symptoms in the console as described in this FAQ about logger download taking a long time at start up, caused by not having the Cumulus logger interval the same as the station logger interval: http://wiki.sandaysoft.com/a/FAQ#Cumulu ... t_start_up. I suspect that the effort of supplying the entire contents of the logger would be more likely to cause the console to freeze and perhaps temporarily lose contact with the ISS.

John Dann has a note about this on his web site - http://www.weatherstations.co.uk/vp_wireless_tech.htm

"Also, the console CPU does not seem to multitask. If the Weatherlink data logger is fitted, there may be a brief and transient effect on reception on occasions when the CPU is handling the requested download of archive data from a linked PC."

Hello Steve,

Thank you for your reply. I forgot, you said that comms is all handled by the Davis DLL. Since its all okay now, I don't really want to waste any of your time.

Perhaps we should put it down to a cosmic ray hitting a bit or something and if it happens again I will take lots of notes and lots of copies of data/logs etc etc.

Adrian

Adrian Hudson wrote:Perhaps we should put it down to a cosmic ray hitting a bit or something

Rather than the likely explanation that I've given? You're not a believer in Occam's Razor, then?

Ahh, good old Mr Occam...

I do believe you are probably correct. I misread your initial response where you say

"Note that restarting Cumulus quickly [...] can bring on the same symptoms in the console as described in this FAQ about logger download taking a long time at start up"

.

The bit I skipped was "restarting Cumulus quickly". Without taking that part in I read it that you thought my logging interval settings had become changed and that this was the problem and that restoring the cumulus.ini file had probably fixed the problem.

I must read things more carefully.

In fact, I probably did end and restart Cumulus pretty quickly most (but not all) of the times that I did restart it so this could indeed have been the issue. If restarting cumulus quickly can make it end up downloading everything, then the Davis would probably (because of the inability to multitask) lose contact with the sensors.

So, yes, that could have been the issue and you are probably correct - although I did reboot the PC at least twice (hence causing a long delay between Cumulus ending and restarting) and yet still got the problem immediately Cumulus started its download.

Its interesting that restoring the ini files seemed to put an end to the problem immediately - but this could have been a coincidence - and the act of restoring the files also introduced a delay between ending and restarting the program of course.

So... cosmic rays or Davis hardware/DLL bugs? On reflection, I'll go for the latter.