tencent cloud


Getting Started

Last updated: 2022-08-01 12:06:23

    This document describes how to quickly deploy GooseFS on a local device, perform debugging, and use COS as remote storage.


    Before using GooseFS, you need to:

    1. Create a bucket in COS as remote storage. For detailed directions, please see Getting Started With the Console.
    2. Install Java 8 or a later version.
    3. Install SSH, ensure that you can connect to the LocalHost using SSH, and log in remotely.

    Downloading and Configuring GooseFS

    1. Download the GooseFS installation package from the repository at goosefs-1.3.0-bin.tar.gz
    2. Run the following command to decompress the installation package:
      tar -zxvf goosefs-1.3.0-bin.tar.gz
      cd goosefs-1.3.0

    After the decompression, the home directory of GooseFS goosefs-1.2.0 will be generated. This document uses ${GOOSEFS_HOME} as the absolute path of this home directory.
    3. Create the conf/goosefs-site.properties configuration file in ${GOOSEFS_HOME}/conf. You can use a built-in configuration template.

    $ cp conf/goosefs-site.properties.template conf/goosefs-site.properties
    1. Set goosefs.master.hostname to localhost in conf/goosefs-site.properties.
      $ echo"goosefs.master.hostname=localhost">> conf/goosefs-site.properties

    Running GooseFS

    1. Before running GooseFS, run the following command to check the local system environment to ensure that GooseFS can run properly:

      $ goosefs validateEnv local
    2. Run the command below to format GooseFS before you run it. This command clears the logs of GooseFS and files stored in the worker directory.

      $ goosefs format
    3. Run the command below to run GooseFS. By default, a master and a worker will be enabled.

      $ ./bin/goosefs-start.sh local SudoMount

      After the command above is executed, you can access http://localhost:9201 and http://localhost:9204 to view the running status of the master and the worker, respectively.

    Mounting COS or Tencent Cloud HDFS to GooseFS

    To mount COS or Tencent Cloud HDFS to the root directory of GooseFS, configure the required parameters of COSN/CHDFS (including but not limited to fs.cosn.impl, fs.AbstractFileSystem.cosn.impl, fs.cosn.userinfo.secretId, and fs.cosn.userinfo.secretKey) in conf/core-site.xml, as shown below:

    <!-- COSN related configurations -->
    <!-- CHDFS related configurations -->

    The following describes how to create a namespace to mount COS or CHDFS.

    1. Create a namespace and mount COS:
    $ goosefs ns create myNamespace cosn://bucketName-1250000000/3TB \
    --secret fs.cosn.userinfo.secretId=AKXXXXXXXXXXX \
    --secret fs.cosn.userinfo.secretKey=XXXXXXXXXXXX \
    --attribute fs.cosn.bucket.region=ap-xxx \

    • When creating the namespace that mounts COSN, you must use the –-secret parameter to specify the key, and use --attribute to specify all required parameters of Hadoop-COS (COSN). For the required parameters, please see Hadoop.
    • When you create the namespace, if there is no read/write policy (rPolicy/wPolicy) specified, the read/write type set in the configuration file, or the default value (CACHE/CACHE_THROUGH) will be used.

    Likewise, create a namespace to mount Tencent Cloud HDFS:

    goosefs ns create MyNamespaceCHDFS ofs://xxxxx-xxxx.chdfs.ap-guangzhou.myqcloud.com/3TB \
    --attribute fs.ofs.user.appid=1250000000
    --attribute fs.ofs.tmp.cache.dir=/tmp/chdfs
    1. After the namespaces are created, run the list command to list all namespaces created in the cluster:

      $ goosefs ns ls
      namespace      mountPoint       ufsPath                      creationTime                wPolicy      rPolicy     TTL   ttlAction
      myNamespace    /myNamespace   cosn://bucketName-125xxxxxx/3TB  03-11-2021 11:43:06:239      CACHE_THROUGH   CACHE        -1      DELETE
      myNamespaceCHDFS /myNamespaceCHDFS ofs://xxxxx-xxxx.chdfs.ap-guangzhou.myqcloud.com/3TB 03-11-2021 11:45:12:336 CACHE_THROUGH   CACHE  -1  DELETE
    2. Run the following command to specify the namespace information:

      $ goosefs ns stat myNamespace
      NamespaceStatus{name=myNamespace, path=/myNamespace, ttlTime=-1, ttlAction=DELETE, ufsPath=cosn://bucketName-125xxxxxx/3TB, creationTimeMs=1615434186076, lastModificationTimeMs=1615436308143, lastAccessTimeMs=1615436308143, persistenceState=PERSISTED, mountPoint=true, mountId=4948824396519771065, acl=user::rwx,group::rwx,other::rwx, defaultAcl=, owner=user1, group=user1, mode=511, writePolicy=CACHE_THROUGH, readPolicy=CACHE}

    Information recorded in the metadata is as follows:

    No. Parameter Description
    1 name Name of the namespace
    2 path Path of the namespace in GooseFS
    3 ttlTime TTL period of files and directories in the namespace
    4 ttlAction TTL action to handle files and directories in the namespace. Valid values: FREE (default), DELETE
    5 ufsPath Mount path of the namespace in the UFS
    6 creationTimeMs Time when the namespace is created, in milliseconds
    7 lastModificationTimeMs Time when files or directories in the namespace is last modified, in milliseconds
    8 persistenceState Persistence state of the namespace
    9 mountPoint Whether the namespace is a mount point. The value is fixed to true.
    10 mountId Mount point ID of the namespace
    11 acl ACL of the namespace
    12 defaultAcl Default ACL of the namespace
    13 owner Owner of the namespace
    14 group Group where the namespace owner belongs
    15 mode POSIX permission of the namespace
    16 writePolicy Write policy for the namespace
    17 readPolicy Read policy for the namespace

    Loading Table Data to GooseFS

    1. You can load Hive table data to GooseFS. Before the loading, attach the database to GooseFS using the following command:
      $ goosefs table attachdb --db test_db hive thrift://
    2. 16.16.22:7004 test_for_demo

    Replace thrift in the command with the actual Hive Metastore address.

    1. After the database is attached, run the ls command to view information about the attached database and table:

      $ goosefs table ls test_db web_page
      OWNER: hadoop
      DBNAME.TABLENAME: testdb.web_page (
      wp_web_page_sk bigint,
      wp_web_page_id string,
      wp_rec_start_date string,
      wp_rec_end_date string,
      wp_creation_date_sk bigint,
      wp_access_date_sk bigint,
      wp_autogen_flag string,
      wp_customer_sk bigint,
      wp_url string,
      wp_type string,
      wp_char_count int,
      wp_link_count int,
      wp_image_count int,
      wp_max_ad_count int,
      LOCATION (
      partitionName: web_page
      location: gfs://
    2. Run the load command to load table data:

      $ goosefs table load test_db web_page
      Asynchronous job submitted successfully, jobId: 1615966078836

      The loading of table data is asynchronous. Therefore, a job ID will be returned. You can run the goosefs job stat &lt;Job Id> command to view the loading progress. When the status becomes "COMPLETED", the loading succeeds.

    Using GooseFS for Uploads/Downloads

    1. GooseFS supports most file system−related commands. You can run the following command to view the supported commands:

      $ goosefs fs
    2. Run the ls command to list files in GooseFS. The following example lists all files in the root directory:

      $ goosefs fs ls /
    3. Run the copyFromLocal command to copy a local file to GooseFS:

    $ goosefs fs copyFromLocal LICENSE /LICENSE
    Copied LICENSE to /LICENSE
    $ goosefs fs ls /LICENSE
    -rw-r--r--  hadoop         supergroup               20798       NOT_PERSISTED 03-26-2021 16:49:37:215   0% /LICENSE
    1. Run the cat command to view the file content:
    $ goosefs fs cat /LICENSE                                                                         
    Apache License
    Version 2.0, January 2004
    1. By default, GooseFS uses the local disk as the underlying file system. The default file system path is ./underFSStorage. You can run the persist command to store files to the local system persistently as follows:
    $ goosefs fs persist /LICENSE
    persisted file /LICENSE with size 26847

    Using GooseFS to Accelerate Uploads/Downloads

    1. Check the file status to determine whether a file is cached. The file status PERSISTED indicates that the file is in the memory, and NOT_PERSISTED indicates not.
    $ goosefs fs ls /data/cos/sample_tweets_150m.csv
    -r-x------ staff  staff 157046046 NOT_PERSISTED 01-09-2018 16:35:01:002   0% /data/cos/sample_tweets_150m.csv
    1. Count how many times “tencent” appeared in the file and calculate the time consumed:
    $ time goosefs fs cat /data/s3/sample_tweets_150m.csv | grep-c tencent
    real    0m22.857s
    user    0m7.557s
    sys    0m1.181s
    1. Caching data in memory can effectively speed up queries. An example is as follows:
    $ goosefs fs ls /data/cos/sample_tweets_150m.csv
    -r-x------ staff  staff 157046046 
    ED 01-09-2018 16:35:01:002   0% /data/cos/sample_tweets_150m.csv
    $ time goosefs fs cat /data/s3/sample_tweets_150m.csv | grep-c tencent
    real    0m1.917s
    user    0m2.306s
    sys        0m0.243s

    The data above shows that the system delay is reduced from 1.181s to 0.243s, achieving a 10-times improvement.

    Shutting Down GooseFS

    Run the following command to shut down GooseFS:

    $ ./bin/goosefs-stop.sh local
    Contact Us

    Contact our sales team or business advisors to help your business.

    Technical Support

    Open a ticket if you're looking for further assistance. Our Ticket is 7x24 avaliable.

    7x24 Phone Support