Skip to main content

Use JuiceFS on Hadoop Ecosystem

JuiceFS provides Hadoop-compatible FileSystem by Hadoop Java SDK. Various applications in the Hadoop ecosystem can smoothly use JuiceFS to store data without changing the code.

Requirements

JuiceFS Hadoop Java SDK is compatible with Hadoop 2.x and Hadoop 3.x. As well as variety of components in Hadoop ecosystem.

2. User permissions

JuiceFS use local mapping of user and UID. So, you should sync all the needed users and their UIDs across the whole Hadoop cluster to avoid permission error. You can also specify a global user list and user group file, please refer to the relevant configurations.

3. File system

You should first create at least one JuiceFS file system to provide storage for components related to the Hadoop ecosystem through the JuiceFS Java SDK. When deploying the Java SDK, specify the metadata engine address of the created file system in the configuration file.

To create a file system, please refer to JuiceFS Quick Start Guide.

Note: If you want to use JuiceFS in a distributed environment, when creating a file system, please plan the object storage and database to be used reasonably to ensure that they can be accessed by each node in the cluster.

4. Memory

JuiceFS Hadoop Java SDK need extra 4 * juicefs.memory-size off-heap memory at most. By default, up to 1.2 GB of additional memory is required (depends on write load).

Client compilation

Compilation depends on the following tools:

  • Go 1.16+
  • JDK 8+
  • Maven 3.3+
  • git
  • make
  • GCC 5.4+

Note: If Ceph RADOS is used to store data, you need to install librados-dev and build libjfs.so with -tag ceph.

Linux and macOS

Clone the repository:

$ git clone https://github.com/juicedata/juicefs.git

Enter the directory and compile:

$ cd juicefs/sdk/java
$ make

After the compilation, you can find the compiled JAR file in the sdk/java/target directory, including two versions:

  • Contains third-party dependent packages: juicefs-hadoop-X.Y.Z.jar
  • Does not include third-party dependent packages: original-juicefs-hadoop-X.Y.Z.jar

It is recommended to use a version that includes third-party dependencies.

Windows

The client used in the Windows environment needs to be obtained through cross-compilation on Linux or macOS. The compilation depends on mingw-w64, which needs to be installed first.

The steps are the same as compiling on Linux or macOS. For example, on the Ubuntu system, install the mingw-w64 package first to solve the dependency problem:

$ sudo apt install mingw-w64

Clone and enter the JuiceFS source code directory, execute the following code to compile:

$ cd juicefs/sdk/java
$ make win

Note: No matter which system environment the client is compiled for, the compiled JAR file has the same name and can only be deployed in the matching system environment. For example, when compiled in Linux, it can only be used in the Linux environment. In addition, since the compiled package depends on glibc, it is recommended to compile with a lower version system to ensure better compatibility.

Deploy the client

To enable each component of the Hadoop ecosystem to correctly identify JuiceFS, the following configurations are required:

  1. Place the compiled JAR file and $JAVA_HOME/lib/tools.jar into the classpath of the component. The installation paths of common big data platforms and components are shown in the table below.
  2. Put JuiceFS configurations into the configuration file of each Hadoop ecosystem component (usually core-site.xml), see Client Configurations for details.

It is recommended to place the JAR file in a fixed location, and the other locations are called it through symbolic links.

Big Data Platforms

NameInstalling Paths
CDH/opt/cloudera/parcels/CDH/lib/hadoop/lib
/opt/cloudera/parcels/CDH/spark/jars
/var/lib/impala
HDP/usr/hdp/current/hadoop-client/lib
/usr/hdp/current/hive-client/auxlib
/usr/hdp/current/spark2-client/jars
Amazon EMR/usr/lib/hadoop/lib
/usr/lib/spark/jars
/usr/lib/hive/auxlib
Alibaba Cloud EMR/opt/apps/ecm/service/hadoop/*/package/hadoop*/share/hadoop/common/lib
/opt/apps/ecm/service/spark/*/package/spark*/jars
/opt/apps/ecm/service/presto/*/package/presto*/plugin/hive-hadoop2
/opt/apps/ecm/service/hive/*/package/apache-hive*/lib
/opt/apps/ecm/service/impala/*/package/impala*/lib
Tencent Cloud EMR/usr/local/service/hadoop/share/hadoop/common/lib
/usr/local/service/presto/plugin/hive-hadoop2
/usr/local/service/spark/jars
/usr/local/service/hive/auxlib
UCloud UHadoop/home/hadoop/share/hadoop/common/lib
/home/hadoop/hive/auxlib
/home/hadoop/spark/jars
/home/hadoop/presto/plugin/hive-hadoop2
Baidu Cloud EMR/opt/bmr/hadoop/share/hadoop/common/lib
/opt/bmr/hive/auxlib
/opt/bmr/spark2/jars

Community Components

NameInstalling Paths
Spark${SPARK_HOME}/jars
Presto${PRESTO_HOME}/plugin/hive-hadoop2
Flink${FLINK_HOME}/lib

Client Configurations

Please refer to the following table to set the relevant parameters of the JuiceFS file system and write it into the configuration file, which is generally core-site.xml.

Core Configurations

ConfigurationDefault ValueDescription
fs.jfs.implio.juicefs.JuiceFileSystemSpecify the storage implementation to be used. By default, jfs:// scheme is used. If you want to use different scheme (e.g. cfs://), just modify it to fs.cfs.impl. No matter what sheme you use, it is always access the data in JuiceFS.
fs.AbstractFileSystem.jfs.implio.juicefs.JuiceFSSpecify the storage implementation to be used. By default, jfs:// scheme is used. If you want to use different scheme (e.g. cfs://), just modify it to fs.AbstractFileSystem.cfs.impl. No matter what sheme you use, it is always access the data in JuiceFS.
juicefs.metaSpecify the metadata engine address of the pre-created JuiceFS file system. You can configure multiple file systems for the client at the same time through the format of juicefs.{vol_name}.meta. Refer to "Multiple file systems configuration".

Cache Configurations

ConfigurationDefault ValueDescription
juicefs.cache-dirDirectory paths of local cache. Use colon to separate multiple paths. Also support wildcard in path. It's recommended create these directories manually and set 0777 permission so that different applications could share the cache data.
juicefs.cache-size0Maximum size of local cache in MiB. It's the total size when set multiple cache directories.
juicefs.cache-full-blocktrueWhether cache every read blocks, false means only cache random/small read blocks.
juicefs.free-space0.1Min free space ratio of cache directory
juicefs.attr-cache0Expire of attributes cache in seconds
juicefs.entry-cache0Expire of file entry cache in seconds
juicefs.dir-entry-cache0Expire of directory entry cache in seconds
juicefs.discover-nodes-urlThe URL to discover cluster nodes, refresh every 10 minutes.

YARN: yarn
Spark Standalone: http://spark-master:web-ui-port/json/
Spark ThriftServer: http://thrift-server:4040/api/v1/applications/
Presto: http://coordinator:discovery-uri-port/v1/service/presto/

I/O Configurations

ConfigurationDefault ValueDescription
juicefs.max-uploads20The max number of connections to upload
juicefs.max-deletes2The max number of connections to delete
juicefs.get-timeout5The max number of seconds to download an object
juicefs.put-timeout60The max number of seconds to upload an object
juicefs.memory-size300Total read/write buffering in MiB
juicefs.prefetch1Prefetch N blocks in parallel
juicefs.upload-limit0Bandwidth limit for upload in Mbps
juicefs.download-limit0Bandwidth limit for download in Mbps

Other Configurations

ConfigurationDefault ValueDescription
juicefs.bucketSpecify a different endpoint for object storage
juicefs.debugfalseWhether enable debug log
juicefs.access-logAccess log path. Ensure Hadoop application has write permission, e.g. /tmp/juicefs.access.log. The log file will rotate automatically to keep at most 7 files.
juicefs.superuserhdfsThe super user
juicefs.usersnullThe path of username and UID list file, e.g. jfs://name/etc/users. The file format is <username>:<UID>, one user per line.
juicefs.groupsnullThe path of group name, GID and group members list file, e.g. jfs://name/etc/groups. The file format is <group-name>:<GID>:<username1>,<username2>, one group per line.
juicefs.umasknullThe umask used when creating files and directories (e.g. 0022), default value is fs.permissions.umask-mode.
juicefs.push-gatewayPrometheus Pushgateway address, format is <host>:<port>.
juicefs.push-interval10Prometheus push interval in seconds
juicefs.push-authPrometheus basic auth information, format is <username>:<password>.
juicefs.fast-resolvetrueWhether enable faster metadata lookup using Redis Lua script
juicefs.no-usage-reportfalseWhether disable usage reporting. JuiceFS only collects anonymous usage data (e.g. version number), no user or any sensitive data will be collected.

Multiple file systems configuration

When multiple JuiceFS file systems need to be used at the same time, all the above configuration items can be specified for a specific file system. You only need to put the file system name in the middle of the configuration item, such as jfs1 and jfs2 in the following example:

<property>
<name>juicefs.jfs1.meta</name>
<value>redis://jfs1.host:port/1</value>
</property>
<property>
<name>juicefs.jfs2.meta</name>
<value>redis://jfs2.host:port/1</value>
</property>

Configuration Example

The following is a commonly used configuration example. Please replace the {HOST}, {PORT} and {DB} variables in the juicefs.meta configuration with actual values.

<property>
<name>fs.jfs.impl</name>
<value>io.juicefs.JuiceFileSystem</value>
</property>
<property>
<name>fs.AbstractFileSystem.jfs.impl</name>
<value>io.juicefs.JuiceFS</value>
</property>
<property>
<name>juicefs.meta</name>
<value>redis://{HOST}:{PORT}/{DB}</value>
</property>
<property>
<name>juicefs.cache-dir</name>
<value>/data*/jfs</value>
</property>
<property>
<name>juicefs.cache-size</name>
<value>1024</value>
</property>
<property>
<name>juicefs.access-log</name>
<value>/tmp/juicefs.access.log</value>
</property>

Configuration in Hadoop

Please refer to the aforementioned configuration tables and add configuration parameters to the Hadoop configuration file core-site.xml.

CDH6

If you are using CDH 6, in addition to modifying core-site, you also need to modify mapreduce.application.classpath through the YARN service interface, adding:

$HADOOP_COMMON_HOME/lib/juicefs-hadoop.jar

HDP

In addition to modifying core-site, you also need to modify the configuration mapreduce.application.classpath through the MapReduce2 service interface and add it at the end (variables do not need to be replaced):

/usr/hdp/${hdp.version}/hadoop/lib/juicefs-hadoop.jar

Add configuration parameters to conf/flink-conf.yaml. If you only use JuiceFS in Flink, you don't need to configure JuiceFS in the Hadoop environment, you only need to configure the Flink client.

Hudi

Note: The latest version of Hudi (v0.9.0) does not yet support JuiceFS, you need to compile the latest master branch yourself.

Please refer to "Hudi Official Documentation" to learn how to configure JuiceFS.

Restart Services

When the following components need to access JuiceFS, they should be restarted.

Note: Before restart, you need to confirm JuiceFS related configuration has been written to the configuration file of each component, usually you can find them in core-site.xml on the machine where the service of the component was deployed.

ComponentsServices
HiveHiveServer
Metastore
SparkThriftServer
PrestoCoordinator
Worker
ImpalaCatalog Server
Daemon
HBaseMaster
RegionServer

HDFS, Hue, ZooKeeper and other services don't need to be restarted.

When Class io.juicefs.JuiceFileSystem not found or No FilesSystem for scheme: jfs exceptions was occurred after restart, reference FAQ.

Environmental Verification

After the deployment of the JuiceFS Java SDK, the following methods can be used to verify the success of the deployment.

Hadoop

$ hadoop fs -ls jfs://{JFS_NAME}/

Note: The JFS_NAME is the volume name when you format JuiceFS file system.

Hive

CREATE TABLE IF NOT EXISTS person
(
name STRING,
age INT
) LOCATION 'jfs://{JFS_NAME}/tmp/person';

Monitoring metrics collection

JuiceFS Hadoop Java SDK supports reporting metrics to Prometheus Pushgateway, then you can use Grafana and dashboard template to visualize these metrics.

Enable metrics reporting through following configurations:

<property>
<name>juicefs.push-gateway</name>
<value>host:port</value>
</property>

Note: Each process using JuiceFS Hadoop Java SDK will have a unique metric, and Pushgateway will always remember all the collected metrics, resulting in the continuous accumulation of metrics and taking up too much memory, which will also slow down Prometheus crawling metrics. It is recommended to clean up metrics which job is juicefs on Pushgateway regularly.

It is recommended to use the following command to clean up once every hour. The running Hadoop Java SDK will continue to update after the metrics are cleared, which basically does not affect the use.

$ curl -X DELETE http://host:9091/metrics/job/juicefs

For a description of all monitoring metrics, please refer to JuiceFS Metrics.

Benchmark

Here are a series of methods to use the built-in stress testing tool of the JuiceFS client to test the performance of the client environment that has been successfully deployed.

1. Local Benchmark

Metadata

  • create

    hadoop jar juicefs-hadoop.jar nnbench create -files 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench -local

    This command will create 10000 empty files

  • open

    hadoop jar juicefs-hadoop.jar nnbench open -files 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench -local

    This command will open 10000 files without reading data

  • rename

    hadoop jar juicefs-hadoop.jar nnbench rename -files 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench -local
  • delete

    hadoop jar juicefs-hadoop.jar nnbench delete -files 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench -local
  • For reference

OperationTPSLatency (ms)
create6441.55
open34670.29
rename4832.07
delete5061.97

I/O Performance

  • sequential write

    hadoop jar juicefs-hadoop.jar dfsio -write -size 20000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/DFSIO -local
  • sequential read

    hadoop jar juicefs-hadoop.jar dfsio -read -size 20000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/DFSIO -local

    When run the cmd for the second time, the result may be much better than the first run. It's because the data was cached in memory, just clean the local disk cache.

  • For reference

OperationThroughput (MB/s)
write647
read111

If the network bandwidth of the machine is relatively low, it can generally reach the network bandwidth bottleneck.

2. Distributed Benchmark

The following command will start the MapReduce distributed task to test the metadata and IO performance. During the test, it is necessary to ensure that the cluster has sufficient resources to start the required map tasks.

Computing resources used in this test:

  • Server: 4 cores and 32 GB memory, burst bandwidth 5Gbit/s x 3
  • Database: Alibaba Cloud Redis 5.0 Community 4G Master-Slave Edition

Metadata

  • create

    hadoop jar juicefs-hadoop.jar nnbench create -maps 10 -threads 10 -files 1000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench

    10 map task, each has 10 threads, each thread create 1000 empty file. 100000 files in total

  • open

    hadoop jar juicefs-hadoop.jar nnbench open -maps 10 -threads 10 -files 1000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench

    10 map task, each has 10 threads, each thread open 1000 file. 100000 files in total

  • rename

    hadoop jar juicefs-hadoop.jar nnbench rename -maps 10 -threads 10 -files 1000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench

    10 map task, each has 10 threads, each thread rename 1000 file. 100000 files in total

  • delete

    hadoop jar juicefs-hadoop.jar nnbench delete -maps 10 -threads 10 -files 1000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/NNBench

    10 map task, each has 10 threads, each thread delete 1000 file. 100000 files in total

  • For reference

    • 10 threads
    OperationIOPSLatency (ms)
    create41782.2
    open94070.8
    rename31972.9
    delete30603.0
    • 100 threads
    OperationIOPSLatency (ms)
    create117737.9
    open340832.4
    rename899510.8
    delete719113.6

I/O Performance

  • sequential write

    hadoop jar juicefs-hadoop.jar dfsio -write -maps 10 -size 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/DFSIO

    10 map task, each task write 10000MB random data sequentially

  • sequential read

    hadoop jar juicefs-hadoop.jar dfsio -read -maps 10 -size 10000 -baseDir jfs://{JFS_NAME}/tmp/benchmarks/DFSIO

    10 map task, each task read 10000MB random data sequentially

  • For reference
OperationAverage throughput (MB/s)Total Throughput (MB/s)
write1981835
read1241234

FAQ

1. Class io.juicefs.JuiceFileSystem not found exception

It means JAR file was not loaded, you can verify it by lsof -p {pid} | grep juicefs.

You should check whether the JAR file was located properly, or other users have the read permission.

Some Hadoop distribution also need to modify mapred-site.xml and put the JAR file location path to the end of the parameter mapreduce.application.classpath.

2. No FilesSystem for scheme: jfs exception

It means JuiceFS Hadoop Java SDK was not configured properly, you need to check whether there is JuiceFS related configuration in the core-site.xml of the component configuration.