If you find that a cluster using
Spark 2.0.0 version takes a longer time to append
data to an existing dataset and in particular, all of Spark jobs have finished, but your
command has not finished, it is because driver node is moving the output files of tasks from the
job temporary directory to the final destination one-by-one, which is slow with cloud storage. To resolve this
mapreduce.fileoutputcommitter.algorithm.version to 2. This issue does not affect overwriting a dataset or writing data to a new location.
Spark 2.0.1-db1, the default value of
mapreduce.fileoutputcommitter.algorithm.version is 2.
If you are using
Spark 2.0.0, manually set this config if you experience this slowness issue.
How to confirm if I am experiencing this issue?
You can confirm if you are experiencing this issue by checking the following things:
All of your Spark jobs have finished and your cell has not finished. The progress bar should look like
The thread dump of the driver (you can find it on the executor page of the Spark UI) shows that there is a thread spending a long time inside the
How do I set
spark.hadoop.mapreduce.fileoutputcommitter.algorithm.version to 2?
You can set this config by using any of the following methods:
- When you launch your cluster, you can put
spark.hadoop.mapreduce.fileoutputcommitter.algorithm.version 2in the Spark config.
- In your notebook, you can run
%sql set mapreduce.fileoutputcommitter.algorithm.version=2or
SparkSessionobject provided with Databricks notebooks).
- When you write data using Dataset API, you can set it in the option, i.e.
What is the cause?
When Spark appends data to an existing dataset, Spark uses
FileOutputCommitter to manage staging output files and final output files. The behavior of
FileOutputCommitter has direct impact on the performance of jobs that write data.
FileOutputCommitter has two methods,
Apache Spark 2.0 and higher versions use Apache Hadoop 2, which uses the value of
to control how
commitJob work. In Hadoop 2, the default value of
mapreduce.fileoutputcommitter.algorithm.version is 1. For this version,
commitTask moves data generated by a task
from the task temporary directory to job temporary directory and when all tasks complete,
commitJob moves data to
from job temporary directory to the final destination . Because the driver is doing the work of
commitJob, for cloud storage, this operation can take a long time. You may often think that your cell is “hanging”.
However, when the value of
mapreduce.fileoutputcommitter.algorithm.version is 2,
commitTask moves data
generated by a task directly to the final destination and
commitJob is basically a no-op.