Today I’m going to share my configuration for running custom Anaconda Python with DGL (Deep Graph Library) and mxnet library, with GPU support via CUDA, running in Spark hosted in EMR. Actually, I have Redshift configuration as well, with support for gensim, tensorflow, keras, theano, pygpu, and cloudpickle. You can also install more libraries if … Continue reading Running Anaconda with DGL and mxnet on CUDA GPU in Spark running in EMR
Tag: Redshift
Running any query in Redshift or JDBC from Spark in EMR
Last time we saw how to connect to Redshift from Spark running in EMR. Provided solution was nice but allowed for reading data only. Sometimes we might want to run any DDL or DML query, not only simple read statements. To do that, we need to connect to Redshift directly over JDBC. I assume you … Continue reading Running any query in Redshift or JDBC from Spark in EMR
Connecting to Redshift from Spark running in EMR
Today I’ll share my configuration for Spark running in EMR to connect to Redshift cluster. First, I assume the cluster is accessible (so configure virtual subnet, allowed IPs and all network stuff before running this). I’m using Zeppelin so I’ll show two interpreters configured for the connection, but the same thing should work with standalone … Continue reading Connecting to Redshift from Spark running in EMR
Aborted transactions in Redshift due to Serializable isolation level
I was working with seemingly simple load job in Redshift. I was generating some data and wanted to store the output in separate table, just to persist it for debugging purposes. However, since the job could be reran if needed, I wanted to remove rows generated previously and add new ones. So I tried the … Continue reading Aborted transactions in Redshift due to Serializable isolation level