Autor Thema: THZ Tecalor (LWZ Stiebel Eltron) module support and code improvement.  (Gelesen 174922 mal)

Offline immi

  • Developer
  • Hero Member
  • ****
  • Beiträge: 1027
Yesterday, when I did not get any reading out of the THZ it seemed that the RAM-consumption did not increase. It was only a few hours from afternoon to today morning, so not the best sample, but I just noticed it.
Do you have any doubt that there might be some memory-issue also with the "blocking"-implementation?
well I cannot tell, if your perl leaks memory; you can monitor it via fhem for some time or you can restart your system each time it grows too much.

with "more memory with non-blocking" I meant something else.
non-blocking in thz:
  • each time you have an automatic refresh (e.g. interval_sDHW 400) another fhem process is created in parallel with the same size of the main process
  • after the comunication with tecalor is done, the parallel process is immediately deleted


The benefit of non-blocking for 5-20ms per reading, is not worth the 2x82956 in memory (even for short time) and  the additional cpu load to create each time the second process.

I wrote it because it could make sense for very old tecalors with 25-50ms reading time

   immi

 

decade-submarginal