Page 1 of 1

EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Fri Oct 23, 2020 10:17 pm
by pvanmastrigt
Dear Forum/Jameson,

I have upgraded my iMac to OS 10.15.7 - Catalina. The EKM Dash software has been re-installed and I have version 2.2.4 (the latest at the time of writing this post).

Performance of the Dash software has really deteriorated after this upgrade.

I note that whilst the software starts and seems to get on with reading remote meters, the performance drops significantly after a while (hours/days) resulting in skipping 200 or so reads and interchange that with longer times of no reads (day(s)). The software often (I think now: always) becomes unresponsive and the only way out is force quit. The error dumps are forwarded to Apple for analysis, but of cause I don't hear back.

I did not find any such report on this forum. Does anyone else experience this? Any solutions? I'll wait for the next stall and post the error dump.

With regards
Piet van Mastrigt

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Fri Oct 30, 2020 2:59 am
by Jameson
Hello Piet,

Great to hear from you. Sorry that you are having trouble with the stability of Dash 2.2.4 with Catalina (OSX Catalina has certainly been its own ball of wax!) Yes, please post the error reports here or email them to support@ekmmetering.com

I have also experienced similar things and have not yet narrowed down as to why it happens (so any insights any of you have out there, please let us know). But if I run the EKM Dash for something like a week at a time on a Mac, it sometimes will go unresponsive. I have to "Force Quit" Dash in order to get it back. I have not noticed the "skipping 200 or so reads" that you mentioned. Is this when your mac is asleep, or does it also happen when you are using your mac and it is not asleep?

The more evidence we can gather as to when this happens, or why this happens, the easier it will be for us to fix it. Thank you for your help, we really do appreciate it :ugeek:

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Sat Nov 07, 2020 5:11 am
by pvanmastrigt
Hello Jameson,

Recently everything seems to be stable - I have not had any hick-ups anymore for a few days now. The version of MacOS is 10.15.7 Frankly I have not really paid sufficient attention and did not realise the Dash software was not running for a while. Meanwhile there may have been some Catalina updates. I restarted Dash about 3 days ago and there seem to be no issues. I like problems that solve themselves.... with regards,
Piet

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Fri Nov 13, 2020 11:36 pm
by pvanmastrigt
Hello Jameson, unfortunately I was too early stating that there seem no issues: the software went unresponsive just a day after the post above. A simple restart resolved the matter. There was no error log generated somehow. I'll wait for the next time.

Regards
Piet

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Sat Jan 30, 2021 7:56 am
by pvanmastrigt
Hello Jameson,

I have gone back to MacOS High Sierra on a separate (old) dedicated machine and in general that works fine. One annoying feature is the daily "zero-reads". This occurs actually no more than once/twice a day and sets off all triggers (see attached screenshot). I have never seen this before and the data looks fine - even for the zero points (most, but not all values are zero's). Has this been reported? Is there a solution recommended?

With thanks for advice,

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Thu Feb 04, 2021 2:44 pm
by Jameson
Interesting. I have not seen this before. Are you saying that you get this drop out consistently once per day? Is it always at the same time of day?

How are you reading this meter? Is it via a USB converter or the EKM Push system? Does both Total Watts and Voltage drop to zero? Is it just for 1 read? In your screenshot it looks like the green line and the orange line drop at different times. That is very odd.

You can avoid the triggers going off by making your triggers fire only if they see > 1 reads in a row with the trigger value.

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Sat Feb 27, 2021 9:25 pm
by pvanmastrigt
Hello Jameson,

I was away for quite some time, but now manage to check your reply, thanks again for reaction.

So here are some answers as I know them now:

-The dropout is not consistent and roughly about 1 x per day to as low as 1x per two days. Sometimes there are 2 dropouts on the same day for different measurements (such as in the example photo).
- Occurs at seemingly totally random time.
- I am reading from the Dash push system via www.
- The title of this blog "EKM Dash 2.2.4 performing poorly on MacOS Catalina" is actually no longer applicable here: to get away from Catalina I have switched to an old machine which runs macOS High Sierra.
- Indeed it is just one read that has some zeros where on the next read things are back to normal."Some zero's" mean that more than one measurement can be 0 at the same time. I cannot see any pattern.

On the suggested solution: I am not sure how this is meant to be implemented. For example I have this trigger for the garden watering system (which is supposed to be off for like 10 months):

Rule: pulse count 2 increases by more than 1. Period = last 24 hours (trigger activate again after one hour, restart trigger after rule failure followed by successes for a given period)

So after one spurious "0" read the increase is a large number, way larger than 1 (back to the original count i.e. thousands). I don't see a way around this. Otherwise this method works well (and used to work well in the past).

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Tue Mar 02, 2021 2:10 pm
by Jameson
Hello Piet,

If you send us your EKM Dash database via email (support@ekmmetering.com) we can also gather some data and see if we see the same dropouts. If you look at the Push data in the EKM Widget or in encompass.io do you see the same dropouts?

I had another user recently that was seeing random negative watts, I am wondering if the issue you are seeing is related. But when I loaded his database into my Dash software and gathered data from the Push system I did not experience the same anomalies. Is this guys case the Push data behind the scenes did not have anomalies in it. So it seemed in that case like a Dash issue, it is just a tricky one to track down because the anomalies are infrequent and random.

Will you send me a screenshot of your trigger setup in your Dash software? Are you triggering based on pulse counts? Or on watts? Your screenshot that you sent with the dropout was a graph of watts.

It sounds like pulse counts, so maybe you can add something like volts, because when you have a dropout, I assume all data (such as volts goes to zero as well, but Im not sure without looking at your data). So I would try something like this:
Screen Shot 2021-03-02 at 9.09.07 PM.png
Screen Shot 2021-03-02 at 9.09.07 PM.png (68.62 KiB) Viewed 44940 times
So in this case the voltage would have to be above 80 volts AND the pulse count would have to increase by more than 40 within a span of 10 meter reads for the trigger to fire. How does this idea sound?

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Fri Mar 05, 2021 8:44 pm
by pvanmastrigt
Hello Jameson,

Again apologise for delayed replies - I did forward the requested data to the support e-mail just now.
Hopefully this gets some insights.

Also note that further to your questions:

- I bravely state that in my e-mail that the EKM Widget shows the zero read problem also. Note that today this is from memory as no zero read has happened in the last 1000 reads so I cannot be 100% sure - take with a pinch of salt...I am simply not sure on this, but next time I'll double check.
- Encompass.io does not have the resolution to show a single read?

- my triggers are indeed on pulse counts (which show a whopping increase after a zero read).

I did not get my brain around setting 2 triggers so as to avoid the issue. Once you receive my data, perhaps something specific can be recommended?
Also I forwarded a sample trigger e-mail under separate cover so you can correlate the time and find the zero read in the database that was dumped.

Thanks for continued support,
Piet

Re: EKM Dash 2.2.4 performing poorly on MacOS Catalina

Posted: Fri Apr 09, 2021 9:59 pm
by pvanmastrigt
Hello Jameson,

It would seem that all is solved: I do get triggers that are valid now and no zero reads anymore. The solution that you suggested works fine as follows:

1.) The zero reads disappear when the wifi connection is improved e.g. to moving my laptop closer to the router. Somewhat surprising as the list of parameters that gets read seems to contain otherwise valid data and the zero reads occur largely on the same measurement "voltage 1" or "pulse 1" mostly. With poorer internet connection one would expect random/garbage, but this is not the case. Yet the improvement by moving closer is clear as all reads were stable and valid for days.

2.) Unfortunately this is not quite practical for me. So another solution was suggested which is super simple. I have attached an example, which exclude zero reads by inserting a rule that the count needs to be greater than 10, :shock: :lol:



My set-up now works fine as follows:

Local:
- MacBook Air (Late 2010)
- MacOS High Sierra 10.13.6
- EKM Dash 2.2.4 (64 bit) software

Remote site:
- Push: EKM-Omnimeter Pulse v.4

Thanks for the support,
Piet