Cumulus Support

Right this is a bit complicated to explain so here goes (hope you all understand it)

I have a Raspberry pi at my parents house due to live in a flat its private land and the flat will cause wind directions problems because its so high so can't have it here.

The weather station is at my parents connected via USB to my raspberry pi. I updated the kernel on the raspberry pi to 4.19.118 from 4.19.97 and now this is showing on the gauges page (Weather Station is currently offline.

but the ftp is showing no errors and the gauges are updating with the temperature every minute and so is the gauges.htm.

I don't understand why i am getting the error?

http://desford-weather.co.uk/gauges.htm

Any ideas be much appreciated.

Thanks.

What Richard means is that on his copy of the standard Cumulus external webpage, at the top of the 'Gauges' page, the canvas_led is flashing red and the scrolling canvas_status message is reading that his station has been offline for 147 days, though it is clearly working and updating every minute.

I don't use this page and so don't know the source of this data - but it's probably a simple little something that someone can educate us about ....

A look at the Wiki found https://cumuluswiki.org/a/SteelSeries_Versions. In section 0.8 near the bottom, an entry says So as a guess, does it mean Richard has turned off realtime updates, or is not uploading the 'realtime.txt' file?

http://desford-weather.co.uk/realtimegauges.txt

http://desford-weather.co.uk/realtime.txt

Its updating.

Oops.. ignore this post.
Missed the bit about the data being uptodate

Perhaps adding the date/time component has screwed something up.

Anyone got any ideas surely updating the raspberrypi with a new kernel shouldn't make it say that?

Data is uploading every minute.

Richard
In another topic viewtopic.php?p=143359#p143359, you say you have edited the gauges script to try to copy Mapantz code in viewtopic.php?p=142371#p142371 So it could be related to that or it could as you say be related to the Kernel or anything else that you have changed.

Why not investigate whether that edit has caused your offline message. Keep a copy of your edited script somewhere, but then swap back to script as supplied in distribution. Do you still get the offline message. That is generally the best way to solve a problem, reverse all changes since it last worked to try to isolate what change makes it happen. Then remake changes one by one until you get error.

If I am going to use a ladder, I spend far more time making sure it is safe to use that I typically spend standing on that ladder, because I don't want to fall and need to be sure it is safe. It is the same with editing code written by someone else, spend longer understanding that code, than you will spend making your edit, to be sure making the edit is safe.

Before I added the code modifications I mention in viewtopic.php?p=143355#p143355 I actually spent a lot of time learning bootstrap and researching the AltEditor Mark selected to use in version 3.4.4 when he implemented the code I wanted to modify, so I knew what each part of the script did. As a result, I was not surprised to find that all the existing functionality still worked perfectly when I successfully coded in my extra actions. As noted in that topic, I did miss that when I converted a function written in SQL to its equivalent in JavaScript, the latter language could read "+" to mean something other than addition.

According to the code, there are two conditions which turn the LED RED:

• The date/time in the message is longer than three minutes before NOW (default value)
• The sensor contact is lost (is a field in the message)

If you understand programming, you could get the relevant code by looking in the gauges.js (look for ledIndicator).
Anyway, it is one of these two, so check the contents of the realtimegauges,txt, check the sensors, if the problem persists, no doubt the dignostict may say something. Mark may jump in then. Interesting though, because indeed it may be the OS which causes some sensor disconnection. I am still on 4.19.97 so i'll wait

I have copied a new gauges.js and GaugesT.htm and same problem so i changed it back.

I can't be due to a kernel upgrade it's still on 4.19 its just gone from .98 to 118. probably a security fix.

I've reinstalled mono and no different.

only thing in MXdiags is this

2020-06-02 01:09:00.991 Minimum data set of pressure, temperature, and wind is not available and NoSensorCheck is not enabled. Skip processing

Follow that instruction from MXDiags to see if it helps?

Where do you set that?

In Cumulus.ini. I think you will have to edit it by hand (with Cumulus switched off). I think it goes in the [Station] section - but I would search both the forum and the wiki to check first.

Ok thanks ill look at it later.

I wish the gauges.js was done on the client side then server side.

Just to be aware that setting "No Sensor Check" is meant for use on a temporary basis only, it is not a permanent solution.

richard_newberry wrote: ↑Tue 02 Jun 2020 9:46 am Minimum data set of pressure, temperature, and wind is not available

So which one of those 3 is not working for you, Richard, and why not?


Steve Loft designed Cumulus on basis that for a given weather station all of these 3 sensors are working, if a sensor does not respond for less than 10 minutes, the previous readings are reused again and again. If 10 minutes pass without one of those sensors returning a value, then Cumulus 1 stopped working. Of course users found that weather station components do not last for ever, and so the "No Sensor Check" option was added to allow Cumulus 1 to continue. I'm not familiar enough with MX coding yet to say authoritatively how it works, but it is likely to be similar as a lot of Cumulus 1 code was ported across.

Why does Cumulus need sensors to be working? Two reasons:
1) Cumulus cannot handle "null values". The way that Cumulus interrogates each type of weather station varies, but it is fixed for all time, so for example part of that processing is that what is read as wind speed (or pressure) will always be converted to whatever units you have selected regardless of whether a value is available or not.
2) Cumulus (1 and MX) calculate lots of derived values, and in some cases that derived value can only be calculated if (a sub-set of) pressure, temperature, humidity, and wind speed, are all available.

How can i fix this problem.

If the data wasnt being received then the data wouldnt be transfering from client to server.

richard_newberry wrote: ↑Tue 02 Jun 2020 3:59 pm If the data wasnt being received then the data wouldnt be transfering from client to server.

Not strictly true. As was pointed out above, if Cumulus does not receive data for a short while (under 10 mins, such as a short break in communications), it uses the most recent data it had and ploughs on. It will then pass that slightly-out-of-date data on to your webpage.

I didn't check your webpage thoroughly. Are all of your sensors updating, every single one of them? Are they all updating on the console?