I've a deployed (pio deploy) 3 models on a machine with 8Gib memory and each model is taking 2Gib memory, In future I want to deploy more models.
Here, I need help on memory issue. Suppose, If I setup PredictionIO on AWS EMR so that my Spark cluster can add/remove slave nodes. If I run pio deploy command (when only 1Gib memory is available). My assumption is that as PredictionIO is installed over the Apache Spark and if its low on memory, will pio deploy work? Like is there any way that pio deploy will try to deploy model by adding another slave through Apache Spark? Long story short: When low on memory, can pio deploy command deploy new model? -- Thank you. Regards, Shoaib Hussain
