Support Questions

Find answers, ask questions, and share your expertise
Announcements
Celebrating as our community reaches 100,000 members! Thank you!

Spark-submit Options --jar, --spark-driver-classpath and spark.executor.extraClasspath

avatar
Explorer

Hi,

1- I have confusion between difference between --driver-class-path --driver-library-path.. Please help me in understanding difference between these two.

2- I am bit new to scala. can you please help in understanding difference between class path and library path. At end, both requires jar path to be set.

3- If i add extra dependencies with --jar option, then do i need to separately project jar path with driver-class-path and spark.executor.executorClassPath

1 ACCEPTED SOLUTION

avatar
@Vinitkumar Pandey

--driver-class-path is used to mention "extra" jars to add to the "driver" of the spark job

--driver-library-path is used to "change" the default library path for the jars needed for the spark driver

--driver-class-path will only push the jars to the driver machine. If you want to send the jars to "executors", you need to use --jar

Hope that helps!

View solution in original post

5 REPLIES 5

avatar
@Vinitkumar Pandey

--driver-class-path is used to mention "extra" jars to add to the "driver" of the spark job

--driver-library-path is used to "change" the default library path for the jars needed for the spark driver

--driver-class-path will only push the jars to the driver machine. If you want to send the jars to "executors", you need to use --jar

Hope that helps!

avatar
@Vinitkumar Pandey

Did the answer help in the resolution of your query? Please close the thread by marking the answer as Accepted!

avatar
New Contributor

I had used --driver-class-path and tried with two options:

1. used to mention hdfs:// path .. but it did not work.

2. used to mention local path .. Expected Spark to copy it to slave nodes - this also does not seem to work


The executors are working perfectly where i had mentioned the same HDFS path using --jars option.

However, the driver is not getting the reference to this path.

This path is a directory where some external customizable configurations could be kept by a user who wish to override default settings shipped in our jars..


For #2 - I am planning to copy to all slave nodes of the cluster.. and see if that does the trick.. Shall update here..

avatar
New Contributor

UPDATE:

=======

--driver-class-path - worked when I passed a local path to it.... However, this did not work until I had to copy the path to be available in all the nodes.

Wish Spark fixes this or if there is any other alternative way - please do share ... I wish it either accepted HDFS path .. or at least do the copy automatically like it does for --jars option.

avatar
New Contributor

@Rahul Soni Hi, Currently i am using particular jar as follow : spark-shell --jars abc.jar,

now i am trying build jar out my code, what is the way to add this jar (abc.jar)