Skip to content

Spark Configurations Supported by HDFS Driver

The Trimble Cloud Fileservice HDFS Driver supports the following Spark configurations to enable seamless integration with Trimble Cloud File Service:

Configuration
Value
Description
spark.hadoop.fs.tcfs.impl com.trimble.hadoop.fs.FileServiceFileSystem Specifies the implementation class for the TCFS FileSystem. Required.
spark.hadoop.fs.AbstractFileSystem.tcfs.impl com.trimble.hadoop.fs.Tcfs Specifies the abstract file system implementation for TCFS. Required.
spark.hadoop.fs.defaultFS tcfs://{space id}@{virtual domain} Sets the default file system URI to TCFS. Required.
spark.hadoop.fs.tcfs.client.id {CLIENT-ID} Client ID for authenticating with Trimble Cloud File Service. Required.
spark.hadoop.fs.tcfs.client.secret {CLIENT_SECRET} Client secret for authenticating with Trimble Cloud File Service. Required.
spark.jars s3://{BUCKET}/{PATH}/trimble-hadoop-fs-tcfs-{VERSION}-all.jar Path to the TCFS JAR file. Required.
spark.hadoop.fs.tcfs.http.requestTimeout Timeout in seconds (e.g., 60) HTTP request timeout for TCFS operations. Default: 30 seconds.
spark.hadoop.fs.tcfs.http.retryTimeout Timeout in seconds (e.g., 120) HTTP retry timeout for TCFS operations. Default: 30 seconds.
spark.hadoop.fs.tcfs.tid.token.refresh Interval in minutes (e.g., 50) Token refresh interval. Default: 50 minutes.
spark.hadoop.fs.tcfs.io.readbuffer.size Size in bytes (e.g., 16384 for 16 KB) Buffer size for read operations. Default: 256 MB. Min: 16 KB, Max: 1 GB.
spark.hadoop.fs.tcfs.io.writebuffer.size Size in bytes (e.g., 16384 for 16 KB) Buffer size for write operations. Default: 4 MB. Min: 16 KB, Max: 1 GB.
spark.hadoop.fs.tcfs.io.multipart.size Size in bytes (e.g., 5242880 for 5 MB) Multipart block size for upload operations. Default: 8 MB. Min: 5 MB, Max: 1 GB.
spark.hadoop.fs.tcfs.io.activationTimeout Timeout in seconds (e.g., 60) File activation timeout. Default: 60 seconds.
spark.hadoop.fs.tcfs.io.activationRetryDelay Delay in seconds (e.g., .1f for 100 ms) Delay between activation retries. Default: 0.1 seconds.
spark.hadoop.fs.tcfs.io.writeThreads.max Number of threads (e.g., 8) Maximum number of threads for write operations. Default: 8. Min: 1, Max: 8.
spark.hadoop.fs.tcfs.io.queuedWrites.max Number of queued writes (e.g., 20) Maximum number of queued write operations. Default: 20. Should be at least 2x write threads.
spark.hadoop.fs.tcfs.groupAuth.enabled true or false Enables/disables IAM group authentication. Default: false.