NoSQLBench – Setup and Evaluating Cassandra Performance

• Download the latest binary  nb file using this link https://github.com/nosqlbench/nosqlbench/blob/main/DOWNLOADS.md

scenarios:

  default:

    – run driver=cql tags==phase:schema threads==1 cycles==UNDEF

    – run driver=cql tags==phase:rampup cycles===TEMPLATE(rampup-cycles,100000) threads=auto

    – run driver=cql tags==phase:main cycles===TEMPLATE(main-cycles,1000000 threads=auto

bindings:

  seq_key: Mod(<<keycount:10000>>); ToString() -> String

  seq_value: HashedFileExtractToString(‘data/lorem_ipsum_full.txt’,0,96)

  rw_key: <<keydist:Uniform(0,10000)->int>>; ToString() -> String

#  rw_key: <<keydist:Uniform(0,10000)->int>>; ToString() -> String

  #rw_value: Hash(); <<valdist:Uniform(0,10000)->int>>; ToString() -> String

  rw_value: HashedFileExtractToString(‘data/lorem_ipsum_full.txt’,0,96)

blocks:

  – name: schema

    tags:

      phase: schema

    params:

      prepared: false

    statements:

    – create-keyspace: |

        create keyspace if not exists <<keyspace:amar>>

        WITH replication = {‘class’: ‘SimpleStrategy’, ‘replication_factor’: ‘<<rf:1>>’}

        AND durable_writes = true;

      tags:

         name: create-keyspace

    – create-table: |

        create table if not exists <<keyspace:amar>>.<<table:keyvalue>> (

        key text,

         value text,

         PRIMARY KEY (key)

        )

      tags:

        name: create-table

  – name: rampup

    tags:

      phase: rampup

    params:

      cl: <<write_cl:LOCAL_QUORUM>>

    statements:

      – rampup-insert: |

          insert into <<keyspace:amar>>.<<table:keyvalue>>

          (key, value)

          values ({seq_key},{seq_value});

        tags:

           name: rampup-insert

  – name: verify

    tags:

      phase: verify

      type: read

    params:

      cl: <<read_cl:LOCAL_QUORUM>>

    statements:

      – verify-select: |

          select * from <<keyspace:amar>>.<<table:keyvalue>> where key={seq_key};

        verify-fields: key->seq_key, value->seq_value

        tags:

           name: verify

  – name: main-read

    tags:

      phase: main

      type: read

    params:

      ratio: 80

      cl: <<read_cl:LOCAL_QUORUM>>

    statements:

      – main-select: |

         select * from <<keyspace:amar>>.<<table:keyvalue>> where key={rw_key};

        tags:

          name: main-select

        instrument: true

  – name: main-write

    tags:

      phase: main

      type: write

    params:

      ratio: 20

      cl: <<write_cl:LOCAL_QUORUM>>

    statements:

      – main-insert: |

          insert into <<keyspace:amar>>.<<table:keyvalue>>

          (key, value) values ({rw_key}, {rw_value});

        tags:

           name: main-insert

        instrument: true

• ./nb workload=cql-keyvalue –copy=cql-keyvalue.yaml
• To see the list of workloads , run the below command
  • ./nb –list-workloads
• The yaml file is a list named blocks and  Each block has tags. We use these tags to indicate the phase of the workload. In this example, the phase is schema, which is short for schema creation.
• ./nb run driver=cql workload=cql-keyvalue tags=phase:schema
• when we execute above command it refers to the schema phase in workload. Below is the one example block
  • blocks:  – tags : phase: schema

• ./nb run driver=cql yaml=<Yaml FilePath> tags=phase:schema host=<IP Adress>

• ./nb run driver=cql yaml=<Yaml FilePath> tags=phase:rampup host=<IP Adress> cycles=10000 –progress console:1s

• ./nb run driver=cql yaml=<Yaml FilePath> tags=phase:main,type=read host=<IP Adress> cycles=10000 –progress console:1s

• ./nb run driver=cql yaml=<Yaml FilePath> tags=phase:main host=<IP Adress> cycles=10000 –progress console:1s
• change the read and write ratios in yaml file as per run

•  we can create or add new scenarios in workloads that allow us to automate a sequence of activities.  The default scenario name will be selected automatically if the user doesn’t specify one.We can specify the name of the workload and the named scenario on the command line
• we can run the scenario with following command
  • ./nb ./cql-keyvalue.yaml default

• we can run the individual block with following command
  • ./nb ./cql-keyvalue.yaml ramup
• Here is an example ,ramup block added in worload

• In the blocks section, we have the tag to name the phase
  • Example
    • blocks: – tags: phase: ramup
• we can run individual phase with following command
  • ./nb run yaml=<path of file> tags=phase:ramup cycles=1000 –progress console:1s