Date

05 Oct 2020

Attendees

Who	Notes
Andrei Lebedev	Started writing test for making the same actions as was done on the Longevity stand. The test is ready. Both Postgres and memory problems can be checked by that test. Next step: find a reason for the problem and solve it. Postgres issue does not have an obvious reason for now. There are no consistent results. Mikhail proposed to save all logs and timings Memory consumption: we have a code which counts objects and pointers. Михаил Болдырев merged that code. Protobuf-generated classes make a problem with the pointer counting, as they cannot be easily wrapped. I need to also update the test for optimization of the memory in the test In Postgres, we currently have all data saved, so we can proceed with the analysis. I can share the test on Github, so Stephan can start making experiments, gather data and make an analysis.
Михаил Болдырев	Now working on the code with pointers and objects counting for catching the problem with the synchronization memory over-consumption. We need a task for making the pointer counting for reuse For making the code ready to merge in "master"
Stepan Lavrentev	I will get the test and parameters from Andrey and check the data. Also, working on the preparation of Locust image for longevity stand with Sofiane
Sara Garifullina	There is a question from community, sent it in the chat