Proper way of recovering from station data transfer hangup

[SamiS]

Hi,

What is the proper way of recovering things when Fineoffset stops giving data to cumulus, but the station itself seems to be still working and logging to its internal memory correctly? This is the first time my station has hung up, and it has been running since September 2010...

Cumulus stopped getting data yesterday at 10:13, and I have 10min logging interval.

I tried to modify today.ini date and time to the point when data stopped. This gave me quite good data, but I had to manually correct timestamps on monthly logfile, and remove duplicate entry from dayfile.txt. For some reason the rows generated after restart were not at 10.20, 10.30 etc, but showed at 10.15, 10.25 etc... Strange?

Are there other files that I should check for inconsistencies after this?

PS. I'm running 1.9.3 beta 1047... and just getting ready to update to 1051.

[steve]

Normally, when the station stops communicating, it will require a full reset (remove batteries and USB cable) and this will usually cause all the logger data to be effectively lost, so there is no recovery possible in this situation.

If the station doesn't need to be reset, and in any other situation requiring a 'rewind', the easiest way to do the rewind is to find the latest set of backup files created by Cumulus (in the "backup" folder hierarchy) - one set is created each time Cumulus starts, and one set every day after midnight - and copy the files from the backup folder into the data folder, overwriting the ones that are there. This ensures that all of the data files are reset to the same point in time, requiring no editing of any files.

Timestamps on data from the logger are approximate because of the way the logger works, and entries created from logger data will typically not match the timestamps of entries created from live data (the station does not necessarily create logs at the same time as Cumulus would).

[SamiS]

I see the point, thanks.

I ended up resetting the station anyway, and unfortunately I did it before reading your post, so I can't revert to the backup data anymore without losing even more data. The station started to respond after shutting down cumulus and unplugging usb connection, but the data was a bit strange (for example pressure was way above normal etc). But now I know how to proceed if and when this would ever occur again. In my opinion this could be something to write to the FAQ?

PS. Is there a easy way to trigger a email alert of station hangup? I could have corrected this situation yesterday if I just would have noticed it...

Anyway, Thank You Very Very Much for Your Great Job Steve!

Best Regards

Sami

[steve]

In my opinion this could be something to write to the FAQ?

It's in the readme.txt file; the one that appears when you install Cumulus which you are asked to read in full.

PS. Is there a easy way to trigger a email alert of station hangup? I could have corrected this situation yesterday if I just would have noticed it...

The Toolbox (see 3rd party apps in the wiki) may offer this, I'm not sure. If not, there may be other ways. Cumulus currently has no email facility.

[SamiS]

Sorry, I have forgotten all about readme, since I have read it only during the first install. I have only read fixed bug lists from the forum.

But there is still something fishy going on... It seems that even while cumulus is up and running, it plots some graphs "like being read from logged data on 10min interval", but some graphs update correctly. Example pictures attached... I have noticed this occasionally now for a year maybe, but I can't really tell when it happens or how it is corrected (restart of cumulus, next rollover or what).

[steve]

I can't see anything wrong with any of those graphs; where exactly do you think the problem is?

[SamiS]

As I said, the pressure graph seems "realtime" as it has a lot of small changes very rapidly and constantly.

But the temp and wind graph look like their values change only rarely as if it would be based only on 10min interval logged data. Sometimes the temp and wind graphs look more "alive" like the pressure graph. Maybe it is depending on the weather also, can't say for sure.

But no worries if it looks okay to you. It just got my attention again because of the station hangup incident. I'll try to send a sample in next few days if the graph would change for some reason.

[Spider-Vice]

You need to look at the scale of each graph. Temperature doesn't usually spike as much as wind or pressure. Pressure is always varying and so is wind. The gust plot is like that because you don't have the "Plot latest gust" option on, on the Display menu.

So, everything looks fine