What does an ElastiCube put in RAM
I heard "An Elasticube, once built and set to the Querying state, is designed to keep its data in memory for fast query access." But I noticed that my cube is 90GB on disk, only 8GB in RAM after I refreshed one dashboard, and the RAM use goes jumped to 21GB when I opened a dashboard that queries a larger quantity of the cube's values. So what's in RAM and what remains on disk? Cached query results? Indexes? Just the rows that have been used since the last build? I'm curious because RAM use of large cubes is sometimes an issue; sometimes we redesign the cube to be smaller and sometimes we get more RAM. The more I understand, the better we can discuss and estimate.
Semantic Layer tables stack up like a deck of cards
Hi All, About once a month, all of the tables in our semantic layer stack on top of one another. I'm not sure why. It takes me about an hour to put them back in their right places. Has anyone else had this problem, and if so, how did you stop it from recurring? Cheers.
