Why my PH and Temp graphs has out of order entries?

[Paulo Hanashiro]

Hi.

I was wondering if anyone had experienced this issue before as this is bugging me for quite some time, The graph suddenly starts to log the most recent entries close to the middle of the graph as you can see below.

If I reload the reef-pi, it fixes the problem, but after a few minutes it starts again to log the most recent entries in a similar position in the middle of the graph, instead of the far right side.


Just after the restart:

After a few minutes, issue is back:

Any clue on how to get this fixed?

Thanks in advance!!!
Paulo

 

[elysics]

Yes, frequently happens to me everytime reefpi restarts. The problem goes away if it runs long enough, around 4-6 hours for me I think.

I think I remember @Ranjib saying it had something to do with the ring structure for the data points or something similar.

 

[Ranjib]

Does this issue persist if you reload reef-pi ?

 

[elysics]

Does this issue persist if you reload reef-pi ?

For me at least, this issue is caused by reloading reefpi, although the previous fault in the graph is fixed, it starts the same problem at a more recent point in the graph.


Also happens if there is no problem and I reload reefpi, then the problem occurs

 

[Ranjib]

For me at least, this issue is caused by reloading reefpi, although the previous fault in the graph is fixed, it starts the same problem at a more recent point in the graph.


Also happens if there is no problem and I reload reefpi, then the problem occurs

got it,. what version of reef-pi you are using? I know we had this bug and in older versions where we didn't store the year (usage data) it would cause this issue. Could be we still have some that type of issue..

 

[elysics]

got it,. what version of reef-pi you are using? I know we had this bug and in older versions where we didn't store the year (usage data) it would cause this issue. Could be we still have some that type of issue..

I am on 4.1, also have a DS1307 RTC, though i got that in hopes of fixing precisely this problem, so the problem was definitely there before adding the DS1307. I assume @Paulo Hanashiro is also on 4.1 due to the minimum and maximum chart bounds

 

[Paulo Hanashiro]

Does this issue persist if you reload reef-pi ?

Hi Ranjib.

Yes, soon after a reef-pi reload (just a few minutes) issue starts again.

I've tried so far:
1-Delete the reading entries in the DB
2-Delete the entire DB
3-Reinstall from scratch (Even Rasbian fresh install)

Had no success with any of the above.

@elysics, last time I reloaded the reef-pi was around 8 hours ago , I just connected to check and issue is there (It is early morning in Sydney/AU).

I've also also tried a few different browsers, I can confirm the issue happens on both Safari and Firefox.
Also across platforms, It happens on Mac, iPhone and iPad.

I'm also running version 4.1

Cheers.
Paulo

 

[Paulo Hanashiro]

Hello to everyone.

I just noticed the ph-current label on the Top left corner is not showing the actual ph of the Tank.

I also noticed the same behaviour starting to show on the ph-historical (noticeable in the spikes shown in the right corner).

@elysics, I think you're right, I also noticed the misplaced entries are catching up and will eventually meet the right corner of the graph where it was supposed to be. I hope when it gets there it does not restart the drift pattern again.

Below another screenshot for reference:

 

[Ranjib]

This is a bug. Just to elaborate on the exact nature of this bug, it’s due to a memory trick (to keep it within limits of pi), that we use a ring buffer in the usage data. Which is causing this. This does not implicate the behavior of the controller, as it is solely used for dashboard and not any control logic.

 

[Paulo Hanashiro]

This is a bug. Just to elaborate on the exact nature of this bug, it’s due to a memory trick (to keep it within limits of pi), that we use a ring buffer in the usage data. Which is causing this. This does not implicate the behavior of the controller, as it is solely used for dashboard and not any control logic.

Thanks Ranjib.

Do we have any feasible workaround that could work on reef-pi 4.1?

 

[Ranjib]

The only thing i can think of is reducing or increasing the number of past usage values (specified in configuration -> telemetry section). But I have not tested this.

 

[Paulo Hanashiro]

is this what you're talking about?