Is there any limit in the amount of items to be used within one openHAB instance?
A couple of month ago I suspected to see misbehavior in OH when using too many items. I reduced the amount of items and everything was fine again.
Now, some months later, the amount of items raised up again, and now it seems that I am facing again some problems. I do not have counted the exact amount of items in my configuration, but I guess, that it will be around 100 - not only sensors/actors, but also pure text-items and virtual switches for configurations.
several different symptoms, mostly in rules.
e.g. in rules the debug log sometimes shows, that items didn’t exist, but in the defaults.items file they do.
I do also have issues in displaying graphs. I do have around 10 and 2 of them do no longer show up (in the classic web interface like a broken img-tag), although the items still persist in rrd4j folder (with updated timestamps)
I have far more than 100 items. A quick (sloppy) count is 45 groups and 180 items in default.items and 5 groups and 150 items in a different .items file. So 50 groups and 330 items in two different files.
Many of the items are pretty stagnant as they are mostly attributes of my dsc alarm sensors or zwave devices.
I have no ill effects from this number of items. I am running on a quad core 2+ghz celeron with 4gb of memory and sata drive though. Perhaps a pi or similar might struggle?
OK, I do have a RPi 2, RAM often over 80%, maybe low memory might cause issues? Should I change jvm settings? Any suggestions?
I also realise that sometimes e.g. a single FS20 switch with a timer rule also does not work with the timer. Anyone else with RPi experience?
I never ran my instance on a pi but do note that OH’s java process can take a fair amount of memory. Mine is using 233mb for java and 44.3mb for mysql. The virtual memory for the java process is 3.4gb and 1.5gb for mysql. I mention that because if your disk is slow (sd cards are) then time taken to retrieve parts of files associated with that process could take longer than normal.
Also depends on what else the pi is doing. If it has other processes, they potentially could push OH out of working memory more often.