fluent-plugin-redshift-kwarter 0.0.3

checksums.yaml

@@ -0,0 +1,7 @@
+---
+SHA1:
+  metadata.gz: 2da95d008425386ffe13ffcbdaf19b2b78982853
+  data.tar.gz: 0729d2617cbeac1f7970c0c56aae1a65502b1e8f
+SHA512:
+  metadata.gz: c5bf1ffc8bfbd8dc5ec9296e817d7f0ff559d797648d5443a2d2024694114546af5d5896f8bafaf8b97dc7ddfa4bcd20f984656c4ef0eef0b55f4e2527f1254f
+  data.tar.gz: 9f1875e617ac4b5cad4b0e13abf9bcb720c8af9c65a9f1250592f56c8bcde7caa31cfcdd0136b6016688de7bc90be72c6716e33a913c4f5aea60daf7f3d13bd7

data/.gitignore

@@ -0,0 +1,19 @@
+*.gem
+*.rbc
+*.swp
+.bundle
+.config
+.yardoc
+Gemfile.lock
+InstalledFiles
+_yardoc
+coverage
+doc/
+lib/bundler/man
+pkg
+rdoc
+spec/reports
+test/tmp
+test/version_tmp
+tmp
+vendor/

data/Gemfile

@@ -0,0 +1,3 @@
+source 'https://rubygems.org'
+
+gemspec

data/README.md

@@ -0,0 +1,154 @@
+Amazon Redshift output plugin for Fluentd
+========
+
+## Overview
+
+Amazon Redshift output plugin uploads event logs to an Amazon Redshift Cluster. Supportted data formats are csv, tsv and json. An S3 bucket and a Redshift Cluster are required to use this plugin.
+
+## Installation
+
+    gem install fluent-plugin-redshift
+
+## Configuration
+
+Format:
+
+    <match my.tag>
+        type redshift
+
+        # s3 (for copying data to redshift)
+        aws_key_id YOUR_AWS_KEY_ID
+        aws_sec_key YOUR_AWS_SECRET_KEY
+        s3_bucket YOUR_S3_BUCKET
+        s3_endpoint YOUR_S3_BUCKET_END_POINT
+        path YOUR_S3_PATH
+        timestamp_key_format year=%Y/month=%m/day=%d/hour=%H/%Y%m%d-%H%M
+
+        # redshift
+        redshift_host YOUR_AMAZON_REDSHIFT_CLUSTER_END_POINT
+        redshift_port YOUR_AMAZON_REDSHIFT_CLUSTER_PORT
+        redshift_dbname YOUR_AMAZON_REDSHIFT_CLUSTER_DATABASE_NAME
+        redshift_user YOUR_AMAZON_REDSHIFT_CLUSTER_USER_NAME
+        redshift_password YOUR_AMAZON_REDSHIFT_CLUSTER_PASSWORD
+        redshift_schemaname YOUR_AMAZON_REDSHIFT_CLUSTER_TARGET_SCHEMA_NAME
+        redshift_tablename YOUR_AMAZON_REDSHIFT_CLUSTER_TARGET_TABLE_NAME
+        file_type [tsv|csv|json|msgpack]
+
+        # buffer
+        buffer_type file
+        buffer_path /var/log/fluent/redshift
+        flush_interval 15m
+        buffer_chunk_limit 1g
+    </match>
+
+Example (watch and upload json formatted apache log):
+
+    <source>
+        type tail
+        path redshift_test.json
+        pos_file redshift_test_json.pos
+        tag redshift.json
+        format /^(?<log>.*)$/
+    </source>
+
+    <match redshift.json>
+        type redshift
+
+        # s3 (for copying data to redshift)
+        aws_key_id YOUR_AWS_KEY_ID
+        aws_sec_key YOUR_AWS_SECRET_KEY
+        s3_bucket hapyrus-example
+        s3_endpoint s3.amazonaws.com
+        path path/on/s3/apache_json_log/
+        timestamp_key_format year=%Y/month=%m/day=%d/hour=%H/%Y%m%d-%H%M
+
+        # redshift
+        redshift_host xxx-yyy-zzz.xxxxxxxxxx.us-east-1.redshift.amazonaws.com
+        redshift_port 5439
+        redshift_dbname fluent-redshift-test
+        redshift_user fluent
+        redshift_password fluent-password
+        redshift_tablename apache_log
+        file_type json
+
+        # buffer
+        buffer_type file
+        buffer_path /var/log/fluent/redshift
+        flush_interval 15m
+        buffer_chunk_limit 1g
+    <match>
+
++ `type` (required) : The value must be `redshift`.
+
++ `aws_key_id` (required) : AWS access key id to access s3 bucket.
+
++ `aws_sec_key` (required) : AWS securet key id to access s3 bucket.
+
++ `s3_bucket` (required) : s3 bucket name. S3 bucket must be same as the region of your Redshift cluster.
+
++ `s3_endpoint` : s3 endpoint.
+
++ `path` (required) : s3 path to input.
+
++ `timestamp_key_format` : The format of the object keys. It can include date-format directives.
+
+  - Default parameter is "year=%Y/month=%m/day=%d/hour=%H/%Y%m%d-%H%M"
+  - For example, the s3 path is as following with the above example configration.
+    <pre>
+  hapyrus-example/apache_json_log/year=2013/month=03/day=05/hour=12/20130305_1215_00.gz
+  hapyrus-example/apache_json_log/year=2013/month=03/day=05/hour=12/20130305_1230_00.gz
+</pre>
+
++ `redshift_host` (required) : the end point(or hostname) of your Amazon Redshift cluster.
+
++ `redshift_port` (required) : port number.
+
++ `redshift_dbname` (required) : database name.
+
++ `redshift_user` (required) : user name.
+
++ `redshift_password` (required) : password for the user name.
+
++ `redshift_tablename` (required) : table name to store data.
+
++ `redshift_schemaname` : schema name to store data. By default, this option is not set and find table without schema as your own search_path.
+
++ `file_type` : file format of the source data.  `csv`, `tsv`, `msgpack` or `json` are available.
+
++ `delimiter` : delimiter of the source data. This option will be ignored if `file_type` is specified.
+
++ `buffer_type` : buffer type.
+
++ `buffer_path` : path prefix of the files to buffer logs.
+
++ `flush_interval` : flush interval.
+
++ `buffer_chunk_limit` : limit buffer size to chunk.
+
++ `utc` : utc time zone. This parameter affects `timestamp_key_format`.
+
+## Logging examples
+```ruby
+# examples by fluent-logger
+require 'fluent-logger'
+log = Fluent::Logger::FluentLogger.new(nil, :host => 'localhost', :port => 24224)
+
+# file_type: csv
+log.post('your.tag', :log => "12345,12345")
+
+# file_type: tsv
+log.post('your.tag', :log => "12345\t12345")
+
+# file_type: json
+require 'json'
+log.post('your.tag', :log => { :user_id => 12345, :data_id => 12345 }.to_json)
+
+# file_type: msgpack
+log.post('your.tag', :user_id => 12345, :data_id => 12345)
+```
+
+## License
+
+Copyright (c) 2013 [Hapyrus Inc](http://hapyrus.com)
+
+[Apache License, Version 2.0](http://www.apache.org/licenses/LICENSE-2.0)

data/Rakefile

@@ -0,0 +1,16 @@
+require "bundler"
+Bundler::GemHelper.install_tasks
+require 'rake/testtask'
+
+Rake::TestTask.new(:test) do |test|
+  test.libs << 'lib' << 'test'
+  test.test_files = FileList['test/plugin/*.rb']
+  test.verbose = true
+end
+
+task :coverage do |t|
+  ENV['COVERAGE'] = '1'
+  Rake::Task["test"].invoke
+end
+
+task :default => [:build]

data/VERSION

@@ -0,0 +1 @@
+0.0.3

data/fluent-plugin-redshift-kwarter.gemspec

@@ -0,0 +1,25 @@
+# -*- encoding: utf-8 -*-
+$:.push File.expand_path('../lib', __FILE__)
+
+Gem::Specification.new do |gem|
+  gem.name          = "fluent-plugin-redshift-kwarter"
+  gem.version       = File.read("VERSION").strip
+  gem.authors       = ["Masashi Miyazaki", "Daniel Koepke"]
+  gem.email         = ["mmasashi@gmail.com", "daniel.koepke@kwarter.com"]
+  gem.description   = %q{Amazon Redshift output plugin for Fluentd (updated by Kwarter)}
+  gem.summary       = gem.description
+  gem.homepage      = "https://github.com/kwarter/fluent-plugin-redshift"
+  gem.has_rdoc      = false
+
+  gem.files         = `git ls-files`.split($/)
+  gem.executables   = gem.files.grep(%r{^bin/}).map{ |f| File.basename(f) }
+  gem.test_files    = gem.files.grep(%r{^(test|spec|features)/})
+  gem.require_paths = ["lib"]
+
+  gem.add_dependency "fluentd", "~> 0.10.0"
+  gem.add_dependency "aws-sdk", ">= 1.6.3"
+  gem.add_dependency "pg", "~> 0.14.0"
+  gem.add_development_dependency "rake"
+  gem.add_development_dependency "simplecov", ">= 0.5.4"
+  gem.add_development_dependency "flexmock", ">= 1.3.1"
+end

data/lib/fluent/plugin/out_redshift.rb

@@ -0,0 +1,288 @@
+module Fluent
+
+
+class RedshiftOutput < BufferedOutput
+  Fluent::Plugin.register_output('redshift', self)
+
+  # ignore load table error. (invalid data format)
+  IGNORE_REDSHIFT_ERROR_REGEXP = /^ERROR:  Load into table '[^']+' failed\./
+
+  def initialize
+    super
+    require 'aws-sdk'
+    require 'zlib'
+    require 'time'
+    require 'tempfile'
+    require 'pg'
+    require 'json'
+    require 'csv'
+  end
+
+  config_param :record_log_tag, :string, :default => 'log'
+  # s3
+  config_param :aws_key_id, :string
+  config_param :aws_sec_key, :string
+  config_param :s3_bucket, :string
+  config_param :s3_endpoint, :string, :default => nil
+  config_param :path, :string, :default => ""
+  config_param :timestamp_key_format, :string, :default => 'year=%Y/month=%m/day=%d/hour=%H/%Y%m%d-%H%M'
+  config_param :utc, :bool, :default => false
+  # redshift
+  config_param :redshift_host, :string
+  config_param :redshift_port, :integer, :default => 5439
+  config_param :redshift_dbname, :string
+  config_param :redshift_user, :string
+  config_param :redshift_password, :string
+  config_param :redshift_tablename, :string
+  config_param :redshift_schemaname, :string, :default => nil
+  config_param :redshift_copy_base_options, :string , :default => "FILLRECORD ACCEPTANYDATE TRUNCATECOLUMNS"
+  config_param :redshift_copy_options, :string , :default => nil
+  # file format
+  config_param :file_type, :string, :default => nil  # json, tsv, csv, msgpack
+  config_param :delimiter, :string, :default => nil
+  # for debug
+  config_param :log_suffix, :string, :default => ''
+
+  def configure(conf)
+    super
+    @path = "#{@path}/" unless @path.end_with?('/') # append last slash
+    @path = @path[1..-1] if @path.start_with?('/')  # remove head slash
+    @utc = true if conf['utc']
+    @db_conf = {
+      host:@redshift_host,
+      port:@redshift_port,
+      dbname:@redshift_dbname,
+      user:@redshift_user,
+      password:@redshift_password
+    }
+    @delimiter = determine_delimiter(@file_type) if @delimiter.nil? or @delimiter.empty?
+    $log.debug format_log("redshift file_type:#{@file_type} delimiter:'#{@delimiter}'")
+    @copy_sql_template = "copy #{table_name_with_schema} from '%s' CREDENTIALS 'aws_access_key_id=#{@aws_key_id};aws_secret_access_key=%s' delimiter '#{@delimiter}' GZIP ESCAPE #{@redshift_copy_base_options} #{@redshift_copy_options};"
+  end
+
+  def start
+    super
+    # init s3 conf
+    options = {
+      :access_key_id     => @aws_key_id,
+      :secret_access_key => @aws_sec_key
+    }
+    options[:s3_endpoint] = @s3_endpoint if @s3_endpoint
+    @s3 = AWS::S3.new(options)
+    @bucket = @s3.buckets[@s3_bucket]
+  end
+
+  def format(tag, time, record)
+    if json?
+      record.to_msgpack
+    elsif msgpack?
+      { @record_log_tag => record }.to_msgpack
+    else
+      "#{record[@record_log_tag]}\n"
+    end
+  end
+
+  def write(chunk)
+    $log.debug format_log("start creating gz.")
+
+    # create a gz file
+    tmp = Tempfile.new("s3-")
+    tmp =
+      if json? || msgpack?
+        create_gz_file_from_structured_data(tmp, chunk, @delimiter)
+      else
+        create_gz_file_from_flat_data(tmp, chunk)
+      end
+
+    # no data -> skip
+    unless tmp
+      $log.debug format_log("received no valid data. ")
+      return false # for debug
+    end
+
+    # create a file path with time format
+    s3path = create_s3path(@bucket, @path)
+
+    # upload gz to s3
+    @bucket.objects[s3path].write(Pathname.new(tmp.path),
+                                  :acl => :bucket_owner_full_control)
+
+    # close temp file
+    tmp.close!
+
+    # copy gz on s3 to redshift
+    s3_uri = "s3://#{@s3_bucket}/#{s3path}"
+    sql = @copy_sql_template % [s3_uri, @aws_sec_key]
+    $log.debug  format_log("start copying. s3_uri=#{s3_uri}")
+    conn = nil
+    begin
+      conn = PG.connect(@db_conf)
+      conn.exec(sql)
+      $log.info format_log("completed copying to redshift. s3_uri=#{s3_uri}")
+    rescue PG::Error => e
+      $log.error format_log("failed to copy data into redshift. s3_uri=#{s3_uri}"), :error=>e.to_s
+      raise e unless e.to_s =~ IGNORE_REDSHIFT_ERROR_REGEXP
+      return false # for debug
+    ensure
+      conn.close rescue nil if conn
+    end
+    true # for debug
+  end
+
+  protected
+  def format_log(message)
+    (@log_suffix and not @log_suffix.empty?) ? "#{message} #{@log_suffix}" : message
+  end
+
+  private
+  def json?
+    @file_type == 'json'
+  end
+
+  def msgpack?
+    @file_type == 'msgpack'
+  end
+
+  def create_gz_file_from_flat_data(dst_file, chunk)
+    gzw = nil
+    begin
+      gzw = Zlib::GzipWriter.new(dst_file)
+      chunk.write_to(gzw)
+    ensure
+      gzw.close rescue nil if gzw
+    end
+    dst_file
+  end
+
+  def create_gz_file_from_structured_data(dst_file, chunk, delimiter)
+    # fetch the table definition from redshift
+    redshift_table_columns = fetch_table_columns
+    if redshift_table_columns == nil
+      raise "failed to fetch the redshift table definition."
+    elsif redshift_table_columns.empty?
+      $log.warn format_log("no table on redshift. table_name=#{table_name_with_schema}")
+      return nil
+    end
+
+    # convert json to tsv format text
+    gzw = nil
+    begin
+      gzw = Zlib::GzipWriter.new(dst_file)
+      chunk.msgpack_each do |record|
+        begin
+          hash = json? ? json_to_hash(record[@record_log_tag]) : record[@record_log_tag]
+          tsv_text = hash_to_table_text(redshift_table_columns, hash, delimiter)
+          gzw.write(tsv_text) if tsv_text and not tsv_text.empty?
+        rescue => e
+          if json?
+            $log.error format_log("failed to create table text from json. text=(#{record[@record_log_tag]})"), :error=>$!.to_s
+          else
+            $log.error format_log("failed to create table text from msgpack. text=(#{record[@record_log_tag]})"), :error=>$!.to_s
+          end
+
+          $log.error_backtrace
+        end
+      end
+      return nil unless gzw.pos > 0
+    ensure
+      gzw.close rescue nil if gzw
+    end
+    dst_file
+  end
+
+  def determine_delimiter(file_type)
+    case file_type
+    when 'json', 'msgpack', 'tsv'
+      "\t"
+    when "csv"
+      ','
+    else
+      raise Fluent::ConfigError, "Invalid file_type:#{file_type}."
+    end
+  end
+
+  def fetch_table_columns
+    conn = PG.connect(@db_conf)
+    begin
+      columns = nil
+      conn.exec(fetch_columns_sql_with_schema) do |result|
+        columns = result.collect{|row| row['column_name']}
+      end
+      columns
+    ensure
+      conn.close rescue nil
+    end
+  end
+
+  def fetch_columns_sql_with_schema
+    @fetch_columns_sql ||= if @redshift_schemaname
+                             "select column_name from INFORMATION_SCHEMA.COLUMNS where table_schema = '#{@redshift_schemaname}' and table_name = '#{@redshift_tablename}' order by ordinal_position;"
+                           else
+                             "select column_name from INFORMATION_SCHEMA.COLUMNS where table_name = '#{@redshift_tablename}' order by ordinal_position;"
+                           end
+  end
+
+  def json_to_hash(json_text)
+    return nil if json_text.to_s.empty?
+
+    JSON.parse(json_text)
+  rescue => e
+    $log.warn format_log("failed to parse json. "), :error => e.to_s
+  end
+
+  def hash_to_table_text(redshift_table_columns, hash, delimiter)
+    return "" unless hash
+
+    # extract values from hash
+    val_list = redshift_table_columns.collect do |cn|
+      val = hash[cn]
+      val = JSON.generate(val) if val.kind_of?(Hash) or val.kind_of?(Array)
+
+      if val.to_s.empty?
+        nil
+      else
+        val.to_s
+      end
+    end
+
+    if val_list.all?{|v| v.nil? or v.empty?}
+      $log.warn format_log("no data match for table columns on redshift. data=#{hash} table_columns=#{redshift_table_columns}")
+      return ""
+    end
+
+    generate_line_with_delimiter(val_list, delimiter)
+  end
+
+  def generate_line_with_delimiter(val_list, delimiter)
+    val_list = val_list.collect do |val|
+      if val.nil? or val.empty?
+        ""
+      else
+        val.gsub(/\\/, "\\\\\\").gsub(/\t/, "\\\t").gsub(/\n/, "\\\n") # escape tab, newline and backslash
+      end
+    end
+    val_list.join(delimiter) + "\n"
+  end
+
+  def create_s3path(bucket, path)
+    timestamp_key = (@utc) ? Time.now.utc.strftime(@timestamp_key_format) : Time.now.strftime(@timestamp_key_format)
+    i = 0
+    begin
+      suffix = "_#{'%02d' % i}"
+      s3path = "#{path}#{timestamp_key}#{suffix}.gz"
+      i += 1
+    end while bucket.objects[s3path].exists?
+    s3path
+  end
+
+  def table_name_with_schema
+    @table_name_with_schema ||= if @redshift_schemaname
+                                  "#{@redshift_schemaname}.#{@redshift_tablename}"
+                                else
+                                  @redshift_tablename
+                                end
+  end
+end
+
+
+end

data/test/plugin/test_out_redshift.rb

@@ -0,0 +1,526 @@
+require 'test_helper'
+
+require 'fluent/test'
+require 'fluent/plugin/out_redshift'
+require 'flexmock/test_unit'
+require 'zlib'
+
+
+class RedshiftOutputTest < Test::Unit::TestCase
+  def setup
+    require 'aws-sdk'
+    require 'pg'
+    require 'csv'
+    Fluent::Test.setup
+  end
+
+  CONFIG_BASE= %[
+    aws_key_id test_key_id
+    aws_sec_key test_sec_key
+    s3_bucket test_bucket
+    path log
+    redshift_host test_host
+    redshift_dbname test_db
+    redshift_user test_user
+    redshift_password test_password
+    redshift_tablename test_table
+    buffer_type memory
+    utc
+    log_suffix id:5 host:localhost
+  ]
+  CONFIG_CSV= %[
+    #{CONFIG_BASE}
+    file_type csv
+  ]
+  CONFIG_TSV= %[
+    #{CONFIG_BASE}
+    file_type tsv
+  ]
+  CONFIG_JSON = %[
+    #{CONFIG_BASE}
+    file_type json
+  ]
+  CONFIG_JSON_WITH_SCHEMA = %[
+    #{CONFIG_BASE}
+    redshift_schemaname test_schema
+    file_type json
+  ]
+  CONFIG_MSGPACK = %[
+    #{CONFIG_BASE}
+    file_type msgpack
+  ]
+  CONFIG_PIPE_DELIMITER= %[
+    #{CONFIG_BASE}
+    delimiter |
+  ]
+  CONFIG_PIPE_DELIMITER_WITH_NAME= %[
+    #{CONFIG_BASE}
+    file_type pipe
+    delimiter |
+  ]
+  CONFIG=CONFIG_CSV
+
+  RECORD_CSV_A = {"log" => %[val_a,val_b,val_c,val_d]}
+  RECORD_CSV_B = {"log" => %[val_e,val_f,val_g,val_h]}
+  RECORD_TSV_A = {"log" => %[val_a\tval_b\tval_c\tval_d]}
+  RECORD_TSV_B = {"log" => %[val_e\tval_f\tval_g\tval_h]}
+  RECORD_JSON_A = {"log" => %[{"key_a" : "val_a", "key_b" : "val_b"}]}
+  RECORD_JSON_B = {"log" => %[{"key_c" : "val_c", "key_d" : "val_d"}]}
+  RECORD_MSGPACK_A = {"key_a" => "val_a", "key_b" => "val_b"}
+  RECORD_MSGPACK_B = {"key_c" => "val_c", "key_d" => "val_d"}
+  DEFAULT_TIME = Time.parse("2013-03-06 12:15:02 UTC").to_i
+
+  def create_driver(conf = CONFIG, tag='test.input')
+    Fluent::Test::BufferedOutputTestDriver.new(Fluent::RedshiftOutput, tag).configure(conf)
+  end
+
+  def create_driver_no_write(conf = CONFIG, tag='test.input')
+    Fluent::Test::BufferedOutputTestDriver.new(Fluent::RedshiftOutput, tag) do
+      def write(chunk)
+        chunk.read
+      end
+    end.configure(conf)
+  end
+
+  def test_configure
+    assert_raise(Fluent::ConfigError) {
+      d = create_driver('')
+    }
+    assert_raise(Fluent::ConfigError) {
+      d = create_driver(CONFIG_BASE)
+    }
+    d = create_driver(CONFIG_CSV)
+    assert_equal "test_key_id", d.instance.aws_key_id
+    assert_equal "test_sec_key", d.instance.aws_sec_key
+    assert_equal "test_bucket", d.instance.s3_bucket
+    assert_equal "log/", d.instance.path
+    assert_equal "test_host", d.instance.redshift_host
+    assert_equal 5439, d.instance.redshift_port
+    assert_equal "test_db", d.instance.redshift_dbname
+    assert_equal "test_user", d.instance.redshift_user
+    assert_equal "test_password", d.instance.redshift_password
+    assert_equal "test_table", d.instance.redshift_tablename
+    assert_equal nil, d.instance.redshift_schemaname
+    assert_equal "FILLRECORD ACCEPTANYDATE TRUNCATECOLUMNS", d.instance.redshift_copy_base_options
+    assert_equal nil, d.instance.redshift_copy_options
+    assert_equal "csv", d.instance.file_type
+    assert_equal ",", d.instance.delimiter
+    assert_equal true, d.instance.utc
+  end
+  def test_configure_with_schemaname
+    d = create_driver(CONFIG_JSON_WITH_SCHEMA)
+    assert_equal "test_schema", d.instance.redshift_schemaname
+  end
+  def test_configure_localtime
+    d = create_driver(CONFIG_CSV.gsub(/ *utc */, ''))
+    assert_equal false, d.instance.utc
+  end
+  def test_configure_no_path
+    d = create_driver(CONFIG_CSV.gsub(/ *path *.+$/, ''))
+    assert_equal "", d.instance.path
+  end
+  def test_configure_root_path
+    d = create_driver(CONFIG_CSV.gsub(/ *path *.+$/, 'path /'))
+    assert_equal "", d.instance.path
+  end
+  def test_configure_path_with_slash
+    d = create_driver(CONFIG_CSV.gsub(/ *path *.+$/, 'path log/'))
+    assert_equal "log/", d.instance.path
+  end
+  def test_configure_path_starts_with_slash
+    d = create_driver(CONFIG_CSV.gsub(/ *path *.+$/, 'path /log/'))
+    assert_equal "log/", d.instance.path
+  end
+  def test_configure_path_starts_with_slash_without_last_slash
+    d = create_driver(CONFIG_CSV.gsub(/ *path *.+$/, 'path /log'))
+    assert_equal "log/", d.instance.path
+  end
+  def test_configure_tsv
+    d1 = create_driver(CONFIG_TSV)
+    assert_equal "tsv", d1.instance.file_type
+    assert_equal "\t", d1.instance.delimiter
+  end
+  def test_configure_json
+    d2 = create_driver(CONFIG_JSON)
+    assert_equal "json", d2.instance.file_type
+    assert_equal "\t", d2.instance.delimiter
+  end
+  def test_configure_msgpack
+    d2 = create_driver(CONFIG_MSGPACK)
+    assert_equal "msgpack", d2.instance.file_type
+    assert_equal "\t", d2.instance.delimiter
+  end
+  def test_configure_original_file_type
+    d3 = create_driver(CONFIG_PIPE_DELIMITER)
+    assert_equal nil, d3.instance.file_type
+    assert_equal "|", d3.instance.delimiter
+
+    d4 = create_driver(CONFIG_PIPE_DELIMITER_WITH_NAME)
+    assert_equal "pipe", d4.instance.file_type
+    assert_equal "|", d4.instance.delimiter
+  end
+  def test_configure_no_log_suffix
+    d = create_driver(CONFIG_CSV.gsub(/ *log_suffix *.+$/, ''))
+    assert_equal "", d.instance.log_suffix
+  end
+
+  def emit_csv(d)
+    d.emit(RECORD_CSV_A, DEFAULT_TIME)
+    d.emit(RECORD_CSV_B, DEFAULT_TIME)
+  end
+  def emit_tsv(d)
+    d.emit(RECORD_TSV_A, DEFAULT_TIME)
+    d.emit(RECORD_TSV_B, DEFAULT_TIME)
+  end
+  def emit_json(d)
+    d.emit(RECORD_JSON_A, DEFAULT_TIME)
+    d.emit(RECORD_JSON_B, DEFAULT_TIME)
+  end
+  def emit_msgpack(d)
+    d.emit(RECORD_MSGPACK_A, DEFAULT_TIME)
+    d.emit(RECORD_MSGPACK_B, DEFAULT_TIME)
+  end
+
+  def test_format_csv
+    d_csv = create_driver_no_write(CONFIG_CSV)
+    emit_csv(d_csv)
+    d_csv.expect_format RECORD_CSV_A['log'] + "\n"
+    d_csv.expect_format RECORD_CSV_B['log'] + "\n"
+    d_csv.run
+  end
+  def test_format_tsv
+    d_tsv = create_driver_no_write(CONFIG_TSV)
+    emit_tsv(d_tsv)
+    d_tsv.expect_format RECORD_TSV_A['log'] + "\n"
+    d_tsv.expect_format RECORD_TSV_B['log'] + "\n"
+    d_tsv.run
+  end
+  def test_format_json
+    d_json = create_driver_no_write(CONFIG_JSON)
+    emit_json(d_json)
+    d_json.expect_format RECORD_JSON_A.to_msgpack
+    d_json.expect_format RECORD_JSON_B.to_msgpack
+    d_json.run
+  end
+
+  def test_format_msgpack
+    d_msgpack = create_driver_no_write(CONFIG_MSGPACK)
+    emit_msgpack(d_msgpack)
+    d_msgpack.expect_format({ 'log' => RECORD_MSGPACK_A }.to_msgpack)
+    d_msgpack.expect_format({ 'log' => RECORD_MSGPACK_B }.to_msgpack)
+    d_msgpack.run
+  end
+
+  class PGConnectionMock
+    def initialize(options = {})
+      @return_keys = options[:return_keys] || ['key_a', 'key_b', 'key_c', 'key_d', 'key_e', 'key_f', 'key_g', 'key_h']
+      @target_schema = options[:schemaname] || nil
+      @target_table = options[:tablename] || 'test_table'
+    end
+
+    def expected_column_list_query
+      if @target_schema
+        /\Aselect column_name from INFORMATION_SCHEMA.COLUMNS where table_schema = '#{@target_schema}' and table_name = '#{@target_table}'/
+      else
+        /\Aselect column_name from INFORMATION_SCHEMA.COLUMNS where table_name = '#{@target_table}'/
+      end
+    end
+
+    def expected_copy_query
+      if @target_schema
+        /\Acopy #{@target_schema}.#{@target_table} from/
+      else
+        /\Acopy #{@target_table} from/
+      end
+    end
+
+    def exec(sql, &block)
+      if block_given?
+        if sql =~ expected_column_list_query
+          yield @return_keys.collect{|key| {'column_name' => key}}
+        else
+          yield []
+        end
+      else
+        unless sql =~ expected_copy_query
+          error = PG::Error.new("ERROR:  Load into table '#{@target_table}' failed.  Check 'stl_load_errors' system table for details.")
+          error.result = "ERROR:  Load into table '#{@target_table}' failed.  Check 'stl_load_errors' system table for details."
+          raise error
+        end
+      end
+    end
+
+    def close
+    end
+  end
+
+  def setup_pg_mock
+    # create mock of PG
+    def PG.connect(dbinfo)
+      return PGConnectionMock.new
+    end
+  end
+
+  def setup_s3_mock(expected_data)
+    current_time = Time.now
+
+    # create mock of s3 object
+    s3obj = flexmock(AWS::S3::S3Object)
+    s3obj.should_receive(:exists?).with_any_args.and_return { false }
+    s3obj.should_receive(:write).with(
+      # pathname
+      on { |pathname|
+        data = nil
+        pathname.open { |f|
+          gz = Zlib::GzipReader.new(f)
+          data = gz.read
+          gz.close
+        }
+        assert_equal expected_data, data
+      },
+      :acl => :bucket_owner_full_control
+    ).and_return { true }
+
+    # create mock of s3 object collection
+    s3obj_col = flexmock(AWS::S3::ObjectCollection)
+    s3obj_col.should_receive(:[]).with(
+      on { |key|
+        expected_key = current_time.utc.strftime("log/year=%Y/month=%m/day=%d/hour=%H/%Y%m%d-%H%M_00.gz")
+        key == expected_key
+      }).
+      and_return {
+        s3obj
+      }
+
+    # create mock of s3 bucket
+    flexmock(AWS::S3::Bucket).new_instances do |bucket|
+      bucket.should_receive(:objects).with_any_args.
+        and_return {
+          s3obj_col
+        }
+    end
+  end
+
+  def setup_tempfile_mock_to_be_closed
+    flexmock(Tempfile).new_instances.should_receive(:close!).at_least.once
+  end
+
+  def setup_mocks(expected_data)
+    setup_pg_mock
+    setup_s3_mock(expected_data) end
+
+  def test_write_with_csv
+    setup_mocks(%[val_a,val_b,val_c,val_d\nval_e,val_f,val_g,val_h\n])
+    setup_tempfile_mock_to_be_closed
+    d_csv = create_driver
+    emit_csv(d_csv)
+    assert_equal true, d_csv.run
+  end
+
+  def test_write_with_json
+    setup_mocks(%[val_a\tval_b\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n])
+    setup_tempfile_mock_to_be_closed
+    d_json = create_driver(CONFIG_JSON)
+    emit_json(d_json)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_json_hash_value
+    setup_mocks("val_a\t{\"foo\":\"var\"}\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit({"log" => %[{"key_a" : "val_a", "key_b" : {"foo" : "var"}}]} , DEFAULT_TIME)
+    d_json.emit(RECORD_JSON_B, DEFAULT_TIME)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_json_array_value
+    setup_mocks("val_a\t[\"foo\",\"var\"]\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit({"log" => %[{"key_a" : "val_a", "key_b" : ["foo", "var"]}]} , DEFAULT_TIME)
+    d_json.emit(RECORD_JSON_B, DEFAULT_TIME)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_json_including_tab_newline_quote
+    setup_mocks("val_a_with_\\\t_tab_\\\n_newline\tval_b_with_\\\\_quote\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit({"log" => %[{"key_a" : "val_a_with_\\t_tab_\\n_newline", "key_b" : "val_b_with_\\\\_quote"}]} , DEFAULT_TIME)
+    d_json.emit(RECORD_JSON_B, DEFAULT_TIME)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_json_no_data
+    setup_mocks("")
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit("", DEFAULT_TIME)
+    d_json.emit("", DEFAULT_TIME)
+    assert_equal false, d_json.run
+  end
+
+  def test_write_with_json_invalid_one_line
+    setup_mocks(%[\t\tval_c\tval_d\t\t\t\t\n])
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit({"log" => %[}}]}, DEFAULT_TIME)
+    d_json.emit(RECORD_JSON_B, DEFAULT_TIME)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_json_no_available_data
+    setup_mocks(%[val_a\tval_b\t\t\t\t\t\t\n])
+    d_json = create_driver(CONFIG_JSON)
+    d_json.emit(RECORD_JSON_A, DEFAULT_TIME)
+    d_json.emit({"log" => %[{"key_o" : "val_o", "key_p" : "val_p"}]}, DEFAULT_TIME)
+    assert_equal true, d_json.run
+  end
+
+  def test_write_with_msgpack
+    setup_mocks(%[val_a\tval_b\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n])
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    emit_msgpack(d_msgpack)
+    assert_equal true, d_msgpack.run
+  end
+
+  def test_write_with_msgpack_hash_value
+    setup_mocks("val_a\t{\"foo\":\"var\"}\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    d_msgpack.emit({"key_a" => "val_a", "key_b" => {"foo" => "var"}} , DEFAULT_TIME)
+    d_msgpack.emit(RECORD_MSGPACK_B, DEFAULT_TIME)
+    assert_equal true, d_msgpack.run
+  end
+
+  def test_write_with_msgpack_array_value
+    setup_mocks("val_a\t[\"foo\",\"var\"]\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    d_msgpack.emit({"key_a" => "val_a", "key_b" => ["foo", "var"]} , DEFAULT_TIME)
+    d_msgpack.emit(RECORD_MSGPACK_B, DEFAULT_TIME)
+    assert_equal true, d_msgpack.run
+  end
+
+  def test_write_with_msgpack_including_tab_newline_quote
+    setup_mocks("val_a_with_\\\t_tab_\\\n_newline\tval_b_with_\\\\_quote\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n")
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    d_msgpack.emit({"key_a" => "val_a_with_\t_tab_\n_newline", "key_b" => "val_b_with_\\_quote"} , DEFAULT_TIME)
+    d_msgpack.emit(RECORD_MSGPACK_B, DEFAULT_TIME)
+    assert_equal true, d_msgpack.run
+  end
+
+  def test_write_with_msgpack_no_data
+    setup_mocks("")
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    d_msgpack.emit({}, DEFAULT_TIME)
+    d_msgpack.emit({}, DEFAULT_TIME)
+    assert_equal false, d_msgpack.run
+  end
+
+  def test_write_with_msgpack_no_available_data
+    setup_mocks(%[val_a\tval_b\t\t\t\t\t\t\n])
+    d_msgpack = create_driver(CONFIG_MSGPACK)
+    d_msgpack.emit(RECORD_MSGPACK_A, DEFAULT_TIME)
+    d_msgpack.emit({"key_o" => "val_o", "key_p" => "val_p"}, DEFAULT_TIME)
+    assert_equal true, d_msgpack.run
+  end
+
+  def test_write_redshift_connection_error
+    def PG.connect(dbinfo)
+      return Class.new do
+        def initialize(return_keys=[]); end
+        def exec(sql)
+          raise PG::Error, "redshift connection error"
+        end
+        def close; end
+      end.new
+    end
+    setup_s3_mock(%[val_a,val_b,val_c,val_d\nval_e,val_f,val_g,val_h\n])
+
+    d_csv = create_driver
+    emit_csv(d_csv)
+    assert_raise(PG::Error) {
+      d_csv.run
+    }
+  end
+
+  def test_write_redshift_load_error
+    PG::Error.module_eval { attr_accessor :result}
+    def PG.connect(dbinfo)
+      return Class.new do
+        def initialize(return_keys=[]); end
+        def exec(sql)
+          error = PG::Error.new("ERROR:  Load into table 'apache_log' failed.  Check 'stl_load_errors' system table for details.")
+          error.result = "ERROR:  Load into table 'apache_log' failed.  Check 'stl_load_errors' system table for details."
+          raise error
+        end
+        def close; end
+      end.new
+    end
+    setup_s3_mock(%[val_a,val_b,val_c,val_d\nval_e,val_f,val_g,val_h\n])
+
+    d_csv = create_driver
+    emit_csv(d_csv)
+    assert_equal false,  d_csv.run
+  end
+
+  def test_write_with_json_redshift_connection_error
+    def PG.connect(dbinfo)
+      return Class.new do
+        def initialize(return_keys=[]); end
+        def exec(sql, &block)
+          error = PG::Error.new("redshift connection error")
+          raise error
+        end
+        def close; end
+      end.new
+    end
+    setup_s3_mock(%[val_a,val_b,val_c,val_d\nval_e,val_f,val_g,val_h\n])
+
+    d_json = create_driver(CONFIG_JSON)
+    emit_json(d_json)
+    assert_raise(PG::Error) {
+      d_json.run
+    }
+  end
+
+  def test_write_with_json_no_table_on_redshift
+    def PG.connect(dbinfo)
+      return Class.new do
+        def initialize(return_keys=[]); end
+        def exec(sql, &block)
+          yield [] if block_given?
+        end
+        def close; end
+      end.new
+    end
+    setup_s3_mock(%[val_a,val_b,val_c,val_d\nval_e,val_f,val_g,val_h\n])
+
+    d_json = create_driver(CONFIG_JSON)
+    emit_json(d_json)
+    assert_equal false, d_json.run
+  end
+
+  def test_write_with_json_failed_to_get_columns
+    def PG.connect(dbinfo)
+      return Class.new do
+        def initialize(return_keys=[]); end
+        def exec(sql, &block)
+        end
+        def close; end
+      end.new
+    end
+    setup_s3_mock("")
+
+    d_json = create_driver(CONFIG_JSON)
+    emit_json(d_json)
+    assert_raise(RuntimeError, "failed to fetch the redshift table definition.") {
+      d_json.run
+    }
+  end
+
+  def test_write_with_json_fetch_column_with_schema
+    def PG.connect(dbinfo)
+      return PGConnectionMock.new(:schemaname => 'test_schema')
+    end
+    setup_s3_mock(%[val_a\tval_b\t\t\t\t\t\t\n\t\tval_c\tval_d\t\t\t\t\n])
+    d_json = create_driver(CONFIG_JSON_WITH_SCHEMA)
+    emit_json(d_json)
+    assert_equal true, d_json.run
+  end
+end

data/test/test_helper.rb

@@ -0,0 +1,8 @@
+if ENV['COVERAGE']
+  require 'simplecov'
+  SimpleCov.start do
+    add_filter 'test/'
+    add_filter 'pkg/'
+    add_filter 'vendor/'
+  end
+end

metadata

@@ -0,0 +1,140 @@
+--- !ruby/object:Gem::Specification
+name: fluent-plugin-redshift-kwarter
+version: !ruby/object:Gem::Version
+  version: 0.0.3
+platform: ruby
+authors:
+- Masashi Miyazaki
+- Daniel Koepke
+autorequire: 
+bindir: bin
+cert_chain: []
+date: 2013-10-17 00:00:00.000000000 Z
+dependencies:
+- !ruby/object:Gem::Dependency
+  name: fluentd
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.10.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.10.0
+- !ruby/object:Gem::Dependency
+  name: aws-sdk
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 1.6.3
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 1.6.3
+- !ruby/object:Gem::Dependency
+  name: pg
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.14.0
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - ~>
+      - !ruby/object:Gem::Version
+        version: 0.14.0
+- !ruby/object:Gem::Dependency
+  name: rake
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: '0'
+- !ruby/object:Gem::Dependency
+  name: simplecov
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 0.5.4
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 0.5.4
+- !ruby/object:Gem::Dependency
+  name: flexmock
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 1.3.1
+  type: :development
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '>='
+      - !ruby/object:Gem::Version
+        version: 1.3.1
+description: Amazon Redshift output plugin for Fluentd (updated by Kwarter)
+email:
+- mmasashi@gmail.com
+- daniel.koepke@kwarter.com
+executables: []
+extensions: []
+extra_rdoc_files: []
+files:
+- .gitignore
+- Gemfile
+- README.md
+- Rakefile
+- VERSION
+- fluent-plugin-redshift-kwarter.gemspec
+- lib/fluent/plugin/out_redshift.rb
+- test/plugin/test_out_redshift.rb
+- test/test_helper.rb
+homepage: https://github.com/kwarter/fluent-plugin-redshift
+licenses: []
+metadata: {}
+post_install_message: 
+rdoc_options: []
+require_paths:
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+required_rubygems_version: !ruby/object:Gem::Requirement
+  requirements:
+  - - '>='
+    - !ruby/object:Gem::Version
+      version: '0'
+requirements: []
+rubyforge_project: 
+rubygems_version: 2.1.9
+signing_key: 
+specification_version: 4
+summary: Amazon Redshift output plugin for Fluentd (updated by Kwarter)
+test_files:
+- test/plugin/test_out_redshift.rb
+- test/test_helper.rb