sphinx覆盖旧索引

换用sphinx后，产品同事发现旧的数据也能搜索出新数据来，查了下官方文档 http://sphinxsearch.com/docs/manual-0.9.9.html#index-merging， 发现更新索引时thinking-sphinx没有把它自定义的sphinx_deleted属性同时更新上去，导致在执行增量索引时用的--merge-dst-range选项无效。现解决方案如下：

１，在config/production.sphinx.conf的searchd部分加上 attr_flush_period = 5 ，让sphinx在更新sphinx_deleted属性后写入到磁盘里。

２，因为sphinx要部署在另外一台独立的机子上，为了方便运维部署和维护，不用安装其他的类似rmagick之类和sphinx无关的软件，就写了一个ruby脚本，用bundler配置安装下gem包，放在cron里定时跑，部分代码如下：

#!/usr/bin/env ruby -rubygems
RAILS_ROOT = File.expand_path(File.dirname(__FILE__)) unless defined?(RAILS_ROOT)
sphinx_config_yml = RAILS_ROOT + '/config/sphinx.yml'
mysql_config_yml = RAILS_ROOT + '/config/database.yml'
production_sphinx_conf = RAILS_ROOT + '/config/production.sphinx.conf'
gem 'rails', '2.3.4'
require 'initializer'
%w[active_record active_support action_view action_controller action_mailer].map {|act| require act}
gem "thinking-sphinx", "1.3.18", :lib => "thinking_sphinx"
%w[yaml riddle thinking_sphinx].map {|lib| require lib}

class Hash
  def symbolize_keys
    inject({}) do |options, (key, value)|
      options[(key.to_sym rescue key) || key] = value
      options
    end
  end
end

sphinx_config = YAML.load_file(sphinx_config_yml)['production'].symbolize_keys
ActiveRecord::Base.establish_connection(YAML.load_file(mysql_config_yml)['production'].symbolize_keys)

class ActiveRecord::Base
  def self.has_attached_file(a, b = {}); end
  def self.validates_attachment_content_type(a, b = {} ); end
end

files = Dir.glob(RAILS_ROOT + "/app/models/*/*.rb") + Dir.glob(RAILS_ROOT + "/app/models/*/*/*.rb")
model_strs = files.map {|path| path.scan(/app\/models\/(.*)\.rb/)[0][0].split('/').map(&:camelize).join('::') }

model_strs.each do |str|
  arr = str.split("::")
  arr.size.times do |x|
    begin
      eval("class #{arr[0..x].join('::')} < ActiveRecord::Base; end")
    rescue TypeError
      # FIX superclass mismatch for class Data (TypeError)
    end
  end
end

files.each do |x|
  begin
    load x
  rescue TypeError
  end
end


client = Riddle::Client.new(sphinx_config[:address], sphinx_config[:port])
indexes = []

models = ThinkingSphinx.context.indexed_models.each do |str|
  prefix = str.split('::').map {|s| s.downcase }.join('_')
  indexes << ( index = ["#{prefix}_core", "#{prefix}_delta"] )
  attrs = {}
  str.constantize.all(:select => "id", :conditions => ["updated_at > ?", Time.now - 3650]).each do |item|
    attrs[item.id * 5] = [1]
  end

  unless attrs.blank?
    puts attrs.inspect
    puts "Updating #{client.update(index[0], ['sphinx_deleted'], attrs )} docs"
  end
end
sleep 10 # 等待写入到磁盘里

# 只允许0出现在最终索引里
system "/usr/local/bin/indexer --rotate --config #{RAILS_ROOT}/config/production.sphinx.conf #{indexes.map {|x| x[1] }.join(' ')}"
sleep 2 # 马上执行以下会导致delta没有更新到main索引里
indexes.each do |index|
  system "/usr/local/bin/indexer --rotate --config #{RAILS_ROOT}/config/production.sphinx.conf --merge #{index.join(' ')} --merge-dst-range sphinx_deleted 0 0"
end