Sensor error count confusion (???)

[elagache]

Dear Stu and WeatherCat automated data "accountants,"

Okay, I found the log entries associated with the sensor error event in the early morning hours of November 28th.  They are attached to this posting for the curious.  However, I've got something else that seems odd, and I don't understand what is going on here.  Blick had some confusion about the number of sensor errors on this thread:

http://athena.trixology.com/index.php?topic=1477.0

In that thread he puzzles over the Live Data not looking the same as the WeatherCat Status window.  However, Stu concludes that the number indeed are consistent.  Okay, I have two episodes.  One on the afternoon of the 27th  Here is the Status window from that episode:



Here is the Live View window at about the same time:



The WeatherCat status window is reporting 338 sensor errors, but the Live Data only has a total of 110.  The overnight incident has a similar discrepancy.  Here is the WeatherCat status window:



Here is the Live View window:



I'm guessing that the Live View is somehow "live" while the Status window is the cumulative report.  So exactly what is the Live View displaying?  Over what period of time does it keep the counts?

One other weird thing about my problem is that the email sent by WeatherCat only reports problems with the Thermometer/Humidity unit:

Code: [Select]

Subject: WeatherCat Admin Alert (Sensor Failure)
To: lwcadmin@canebas.org
From: lwcadmin@canebas.org


Sensor failure at sample time.
Failed sensors are:
External temperature
External humidity


WeatherCat TimeStamp: 00:18:01 28-Nov-14
That suggests the errors start with the Temperature/Humidity probe and only later on does the power situation start to cause errors with the anemometer.  However, the Live View lists the number of errors for each sensor as about the same.  It would appear at least during the sampling period of the Live View, both sensors were failing at an approximately equal rate.

Anyway, a curious mind would like to know how all this is supposed to add up.

Cheers, Edouard 

[Blicj11]

Blick had some confusion about the number of sensor errors on this thread:

http://athena.trixology.com/index.php?topic=1477.0

In that thread he puzzles over the Live Data not looking the same as the WeatherCat Status window.  However, Stu concludes that the number indeed are consistent.

I am still confused. On that thread, although it pains me to say it, I believe that Stu incorrectly concluded the counts were consistent. It you look at the screenshots I posted there, you see that Weather Station Gauges and WeatherCat Status both showed 16 sensor errors whilst the Live Data showed only 8 sensor errors.

Ironically, I received this admin email alert yesterday although I have no sensor or comms errors showing anywhere:

Code: [Select]

Sensor failure at sample time.
Failed sensors are:
External temperature
Pressure
Precipitation
Wind
External humidity
Internal temperature
Internal humidity

WeatherCat TimeStamp: 15:56:04 30-Nov-14
I received an identical email alert at approximately the same time of day on 16 November. What does this email alert mean and what should we do when we receive one of these?

[WCDev]

I received an identical email alert at approximately the same time of day on 16 November. What does this email alert mean and what should we do when we receive one of these?

Check the log in WeatherCat - for some reason none of your channels were valid at sample time.

Edouard, the additional errors are coming from channels that are dependent on the channel that has failed - for example wind chill or dewpoint (these aren't visible in the interface but will be adding to the total). If you scroll down to the bottom of the Live Data View, the total shown there should match the total shown in the Status window and gauges window.

[Edit: Blick, apologies, I missed your point - yes, in your case with the windspeed dead, windchill would also have been affected, hence the 8 errors indicated against windspeed, but the total was 16 because the wind chill channel would also have been affected - the errors on the wind chill are not displayed in the ui]

 

[Blicj11]

Check the log in WeatherCat - for some reason none of your channels were valid at sample time.

Right you are Stu. Apparently, I had problems all day long yesterday.

It probably all started with this:

Code: [Select]

11/30/14 10:01:38.835 AM WeatherCat[567]: ***WARNING*** This Mac is going to sleep, WeatherCat will not issue any email alerts.

I don't know how this is possible, but I am going to call Apple to find out.

I have several of these:

Code: [Select]

11/30/14 11:53:54.274 AM WeatherCat[567]: ***Hardware comms fail (stale data)***
11/30/14 11:53:55.543 AM WeatherCat[567]: ***Hardware comms restored***
11/30/14 11:53:55.560 AM WeatherCat[567]: WeatherCat weather related processes have been started.

I also have this:

Code: [Select]

11/30/14 2:10:30.441 PM WeatherCat[567]: FTP: Uploading realtimegaugesWC.txt. (4 files left to upload.)
11/30/14 2:10:30.460 PM WeatherCat[567]: FTP: curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
curl: (6) Could not resolve host: ftp.timberlakesutah.com
11/30/14 2:10:30.476 PM WeatherCat[567]: FTP: FTP ERROR (6) - Couldn't resolve host. Check Server, Internet is working and FTP URL.
11/30/14 2:10:30.491 PM WeatherCat[567]: FTP: FTP ERROR 910: Failed to complete upload (null) - retrying in 1 second.
11/30/14 2:10:30.692 PM WeatherCat[567]: Historical record upload FAILED - retrying in 1 minute
11/30/14 2:10:30.709 PM WeatherCat[567]: Wunderground: Historical record upload FAILED - retrying in 1 minute

Ttwice, I see messages like this one in the log, followed by downloads from the data logger:

Code: [Select]

11/30/14 5:46:50.230 PM WeatherCat[567]: DM: Want 465 historical minutes.
11/30/14 5:46:52.167 PM WeatherCat[567]: Starting archive download
11/30/14 5:46:58.105 PM WeatherCat[567]: ***Hardware comms fail (stale data)***
11/30/14 5:47:13.686 PM WeatherCat[567]: Starting archive download

During the day, the log contained several of these messages:

Code: [Select]

11/30/14 6:03:02.364 PM WeatherCat[567]: ***WARNING*** WeatherCat was not able to fetch all the weather data from the hardware after repeated tries.
11/30/14 6:03:02.612 PM WeatherCat[567]: Not entering data - station comms failure, please check station and reboot WeatherCat.

Since I don't sit around and look at the log all day I did not realize this was happening.

Eventually, I got this email (although I did not see it until this morning):

Code: [Select]

WeatherCat watchdog: WeatherCat appears to have hung; rebooting as a precautionary measure.

WeatherCat TimeStamp: 18:29:31 30-Nov-14
Following that reboot, it once again downloaded data from the logger and all appears well now.

Quite the day!

[elagache]

Howdy Stu, Blick, and WeatherCat station caregivers,

Edouard, the additional errors are coming from channels that are dependent on the channel that has failed - for example wind chill or dewpoint (these aren't visible in the interface but will be adding to the total). If you scroll down to the bottom of the Live Data View, the total shown there should match the total shown in the Status window and gauges window.

Okay, that's the missing ingredient I was looking for.  I was worried to present this info on the WXForum because of the apparent inconsistency.  They are one tough group of customers over there - so I need to be at the top of my game!

Thanks Stu!

Cheers, Edouard