Cumulus Support

I just upgraded to 3.14.2 (build 3162)

After I started CMX again, all readings dropped to 0.0 accept for the wind.

All seems OK here.

It was the realtime data that zeroed out, I saw it all flash up on my website briefly, and get logged in to SQL

https://i.postimg.cc/sV5sqxHx/Screensho ... 155003.png

Checked my monthly SQL table and I did not get a zero filled line.

Edit just saw you said your realtime table. Checked mine all OK.

I just had to close CMX to do a few things, reopened it, and all the readings went to zero again..

This has been a problem with the last few releases and I'm having a hard time tracking it down, I think it may be a timing issue, but I've been unable to reproduce it on my debug instance so far. This one gets the focus now...

Which entry went to zero...

The first/last catch-up entry?
The first live reading?
Did it affect all your records?
Was it logged in the interval data?

From your screen-shot not all the variables are zero. Can you list which ones are zero, and which had real values?

It will all help tracking down the problem.
Thanks

mcrossley wrote: ↑Fri 21 Jan 2022 11:21 pm This has been a problem with the last few releases and I'm having a hard time tracking it down, I think it may be a timing issue, but I've been unable to reproduce it on my debug instance so far. This one gets the focus now...

Which entry went to zero...

The first/last catch-up entry?
The first live reading?
Did it affect all your records?
Was it logged in the interval data?

From your screen-shot not all the variables are zero. Can you list which ones are zero, and which had real values?

It will all help tracking down the problem.
Thanks

I removed them from my database, so I'll test it tomorrow to get the data which zeroed and which did not.

It only affected the realtime.txt briefly, and logged a single entry in to the realtime table.

It looks like it was the last catch-up entry and the first live reading after that. It hasn't affected any records for the day/month/year etc.

Do you think it might be related to the "Limit real time table inserts to once a minute" ?

As it stands, it isn't a calamitous bug. I thought it was going to alter all of the records, but thankfully it was just a single line in to the realtime database. That's why I wondered if it might be related to limit table inserts..

OK, if it is just the real time, then I have found the cause and it is an easy fix.

Another user - with an Ecowitt - reported interval data logged with zero readings on start-up which is proving more difficult to track down.