@mstormi @rossko57
After moving through the Java settings, I noticed some difference I think with option 2. Just to reiterate the issue, its slow response to rules after the rules havnt been processed for a period of hours, though direct on/off mqtt commands that arent rule driven, run perfectly fine.
So my initial experiences pre any changes was about a 22 second from button press to activation. After moving to setting 2 (add option 1 and option 2 to the EXTRA_JAVA_OPTS)… My response time seems to have dropped to 13 seconds (when the issue occurs) and has remained that way over testing, during the last 2 days. Ive not experienced a longer delay. (please note, I am using -XX:+UseParNewGC and not -XX:+UseParallelGC)
Options for EXTRA_JAVA_OPTS
-XX:-UseAdaptiveSizePolicy -XX:+UseParallelGC -XX:InitiatingHeapOccupancyPercent=95-XX:MinHeapFreeRatio=15 -XX:MaxHeapFreeRatio=30-XX:+UseCodeCacheFlushing
Example from this morning with a 13 second delay
2020-05-02 10:14:21.201 [vent.ItemStateChangedEvent] - RFDataLights changed from NULLFORTESTING to 811301
2020-05-02 10:14:33.634 [ome.event.ItemCommandEvent] - Item 'gUpstairs' received command ON
2020-05-02 10:14:33.677 [ome.event.ItemCommandEvent] - Item 'lsDesk' received command ON
2020-05-02 10:14:33.716 [ome.event.ItemCommandEvent] - Item 'lsWillow' received command ON
2020-05-02 10:14:33.718 [vent.ItemStateChangedEvent] - RFDataLights changed from 811301 to NULLFORTESTING
2020-05-02 10:14:33.720 [nt.ItemStatePredictedEvent] - lsDesk predicted to become ON
2020-05-02 10:14:33.757 [nt.ItemStatePredictedEvent] - lsWillow predicted to become ON
2020-05-02 10:14:33.774 [vent.ItemStateChangedEvent] - lsDesk changed from OFF to ON
2020-05-02 10:14:33.787 [vent.ItemStateChangedEvent] - lsWillow changed from OFF to ON
My system has 2462mb ram free, or 72% of memory free, so it shouldnt be a memory issue.
As mentioned, once the rule has been used, things remain responsive for the next X hours, dropping down to a 1-2 second response time.
Also, on one of my tests where I experienced a delay, I had literally been on a remote desktop session on the server 1 minute before, though not interacting with Openhab… I only mention this, so as to point out its not like the system was idling in a low power state or something (which I dont think it drops into anyway).
So I guess its something else.
As an aside, I have a Raspberry Pi 3 A+ turning up for another project im working on. I can for a couple of days, setup Openhabian on it, take all my rules etc over and see if I experience the same issues… so we will have a comparative to see if the problem follows with the hardware/OS.
Happy to look at any other suggestions you guys have. Again, thanks for looking into this with me.
