Platform information: Release = Raspbian GNU/Linux 10 (buster) Kernel = Linux 5.10.11-v7+ Platform = Raspberry Pi 3 Model B Rev 1.2 CPU Usage = 25.06% avg over 4 cpu(s) (4 core(s) x 1 socket(s)) CPU Load = 1m: 1.00, 5m: 1.00, 15m: 0.89 Memory = Free: 0.04GB (5%), Used: 0.90GB (95%), Total: 0.…

[image] ifank: the whole system became unresponsive A more general loss of resource, then. Likely related - [image] openHAB 3 runs out of memory / java heap space errors, CPU 100%+ after a few hours Installation For anyone looking to test this, 3.1.0-S2…

[image] ifank: both rules are constantly in running state and are triggered every 20s by an update of one of the channels Can you show us what makes you think this? Example, your events.log would show timestamped trigger events and results updates.

Example (I had to restart openhab, because the whole system became unresponsive, see below events.log after restart): Line 7548: 2021-02-22 12:34:07.283 [INFO ] [openhab.event.ItemStateChangedEvent ] - Item 'Norm_CurrentDCPowerofString1' changed from 0.09377778 to 0.09244444 Line 7554: 2021-02-22…

Thanks a lot, it looks exactly like this - the second thing I did besides implementing the rules was an update of openhabian … The 3-4s time between trigger and rule result is ok from your point of view?

The OH2 DSL rules environment always had a one off “first time” execution lag of a few seconds, we imagined this was some kind of compilation effect. Further, if a rule was unused for long periods, it appeared it could get “rolled out” or otherwise reacquire this “first time” effect. The OH3 engin…

Would you then recommend file based definition of rules in OH3 for faster execution? Does it make also a difference if DSL rules are used or ECMAScript in the UI (I have implemented UI-based DSL Rules so far)?

As said, I do not know. You can experiment. I wouldn’t bother until/unless you incorporate the current fixes for performance issues. I do think it’s worth looking into. That kind of delay is a right pain for, say, motion triggered lighting. You’ve got a nice example, easily translated into diff…

You’re right, it doesn’t make sense to investigate further as long as 1 cycle is calculated correctly or if you can extend the cycle time. Another question, as I don’t want to mess up my system more than necessary and manually installing snapshot-versions: Wouldn’t it be easier to change the branch…

Simple rules running for 20+ seconds and slowing down system

Setup, Configuration and Use Scripts & Rules

rossko57 (Rossko57) February 22, 2021, 1:03pm 4

A more general loss of resource, then.
Likely related -