This topic describes how to mount a unified namespace by running Jindo commands to implement cache-based acceleration.

Background information

The JindoFSx storage acceleration system allows you to mount the unified namespace jindo:// to applications. Then, the applications can use the unified namespace and APIs to access different independent storage systems. This way, the applications can communicate with the storage systems based on only the connection to JindoFSx. Unified namespace

Prerequisites

Note In this example, a cluster of EMR V3.42.0 is used.

Run Jindo commands to mount a data source

  • Run the following command to mount a data source:
    jindo admin -mount <path> <realpath>
  • Run the following command to mount the OSS or OSS-HDFS path:
    jindo admin -mount /jindooss oss://<yourBucketName>/<dir>
  • Run the following command to access jindo://emr-header-1:8101/jindooss/:
    hdfs dfs -ls jindo://emr-header-1:8101/jindooss

    This operation is equivalent to the operation of accessing oss://<yourBucketName>/<dir>.

  • Run the following command to mount the HDFS path:
    jindo admin -mount /hdfs hdfs://<ip>:<port>/dir
  • Run the following command to access jindo://emr-header-1:8101/hdfs/:
    hdfs dfs -ls jindo://emr-header-1:8101/hdfs

    The returned information is jindo://emr-header-1:8101/hdfs. This operation is equivalent to the operation of accessing hdfs://<ip>:<port>/dir.

  • Run the following command to mount the NAS path:
    jindo admin -mount /nas local:///mnt/nas

    /mnt/nas in the command is the mount point path of NAS on the physical machine. You must add local:// before /mnt/nas.

  • Run the following command to access jindo://emr-header-1:8101/nas/:
    hdfs dfs -ls jindo://emr-header-1:8101/nas

    The returned information is jindo://emr-header-1:8101/nas. This operation is equivalent to the operation of accessing files in the /mnt/nas directory.