Scatter plot is crashing on a reasonably large dataset for a user.
- Linux, 5.9.1 and 5.10.0-RC1, remote server (16 ranks).
- Dataset is about 2 million cells, 512 files.
- Server side looks healthy. Crash is on the client side.
Let me start with a simple question. How does scatter plotting work remote server? (i.e., were does the data flow). What I’m wondering is if scatter plot moves all of the data to rank 0, and/or if it moves it to the client?