fluentd 1.11.5-x86-mingw32 → 1.12.4-x86-mingw32

data/lib/fluent/plugin/in_monitor_agent.rb

@@ -338,7 +338,7 @@ module Fluent::Plugin
         obj.merge!(pe.statistics['output'] || {})
       end
 
-      obj['retry'] = get_retry_info(pe.retry) if opts[:with_retry] and pe.instance_variable_defined?(:@retry)
+      obj['retry'] = get_retry_info(pe.retry) if opts[:with_retry] && pe.instance_variable_defined?(:@retry)
 
       # include all instance variables if :with_debug_info is set
       if opts[:with_debug_info]

data/lib/fluent/plugin/in_tail.rb

@@ -22,6 +22,7 @@ require 'fluent/event'
 require 'fluent/plugin/buffer'
 require 'fluent/plugin/parser_multiline'
 require 'fluent/variable_store'
+require 'fluent/capability'
 require 'fluent/plugin/in_tail/position_file'
 
 if Fluent.windows?
@@ -104,6 +105,8 @@ module Fluent::Plugin
     config_param :ignore_repeated_permission_error, :bool, default: false
     desc 'Format path with the specified timezone'
     config_param :path_timezone, :string, default: nil
+    desc 'Follow inodes instead of following file names. Guarantees more stable delivery and allows to use * in path pattern with rotating files'
+    config_param :follow_inodes, :bool, default: false
 
     config_section :parse, required: false, multi: true, init: true, param_name: :parser_configs do
       config_argument :usage, :string, default: 'in_tail_parser'
@@ -154,6 +157,9 @@ module Fluent::Plugin
         end
         @variable_store[@pos_file] = self.plugin_id
       else
+        if @follow_inodes
+          raise Fluent::ConfigError, "Can't follow inodes without pos_file configuration parameter"
+        end
         $log.warn "'pos_file PATH' parameter is not set to a 'tail' source."
         $log.warn "this parameter is highly recommended to save the position to resume tailing."
       end
@@ -171,6 +177,7 @@ module Fluent::Plugin
       @dir_perm = system_config.dir_permission || Fluent::DEFAULT_DIR_PERMISSION
       # parser is already created by parser helper
       @parser = parser_create(usage: parser_config['usage'] || @parser_configs.first.usage)
+      @capability = Fluent::Capability.new(:current_process)
     end
 
     def configure_tag
@@ -214,7 +221,7 @@ module Fluent::Plugin
         FileUtils.mkdir_p(pos_file_dir, mode: @dir_perm) unless Dir.exist?(pos_file_dir)
         @pf_file = File.open(@pos_file, File::RDWR|File::CREAT|File::BINARY, @file_perm)
         @pf_file.sync = true
-        @pf = PositionFile.load(@pf_file, logger: log)
+        @pf = PositionFile.load(@pf_file, @follow_inodes, expand_paths, logger: log)
 
         if @pos_file_compaction_interval
           timer_execute(:in_tail_refresh_compact_pos_file, @pos_file_compaction_interval) do
@@ -238,7 +245,7 @@ module Fluent::Plugin
 
     def shutdown
       # during shutdown phase, don't close io. It should be done in close after all threads are stopped. See close.
-      stop_watchers(@tails.keys, immediate: true, remove_watcher: false)
+      stop_watchers(existence_path, immediate: true, remove_watcher: false)
       @pf_file.close if @pf_file
 
       super
@@ -250,6 +257,11 @@ module Fluent::Plugin
       close_watcher_handles
     end
 
+    def have_read_capability?
+      @capability.have_capability?(:effective, :dac_read_search) ||
+        @capability.have_capability?(:effective, :dac_override)
+    end
+
     def expand_paths
       date = Fluent::EventTime.now
       paths = []
@@ -263,7 +275,7 @@ module Fluent::Plugin
           paths += Dir.glob(path).select { |p|
             begin
               is_file = !File.directory?(p)
-              if File.readable?(p) && is_file
+              if (File.readable?(p) || have_read_capability?) && is_file
                 if @limit_recently_modified && File.mtime(p) < (date.to_time - @limit_recently_modified)
                   false
                 else
@@ -296,7 +308,37 @@ module Fluent::Plugin
                end
         path.include?('*') ? Dir.glob(path) : path
       }.flatten.uniq
-      paths.uniq - excluded
+      # filter out non existing files, so in case pattern is without '*' we don't do unnecessary work
+      hash = {}
+      (paths - excluded).select { |path|
+        FileTest.exist?(path)
+      }.each { |path|
+        # Even we just checked for existence, there is a race condition here as
+        # of which stat() might fail with ENOENT. See #3224.
+        begin
+          target_info = TargetInfo.new(path, Fluent::FileWrapper.stat(path).ino)
+          if @follow_inodes
+            hash[target_info.ino] = target_info
+          else
+            hash[target_info.path] = target_info
+          end
+        rescue Errno::ENOENT
+          $log.warn "expand_paths: stat() for #{path} failed with ENOENT. Skip file."
+        end
+      }
+      hash
+    end
+
+    def existence_path
+      hash = {}
+      @tails.each_key {|target_info|
+        if @follow_inodes
+          hash[target_info.ino] = target_info
+        else
+          hash[target_info.path] = target_info
+        end
+      }
+      hash
     end
 
     # in_tail with '*' path doesn't check rotation file equality at refresh phase.
@@ -305,21 +347,21 @@ module Fluent::Plugin
     # In such case, you should separate log directory and specify two paths in path parameter.
     # e.g. path /path/to/dir/*,/path/to/rotated_logs/target_file
     def refresh_watchers
-      target_paths = expand_paths
-      existence_paths = @tails.keys
+      target_paths_hash = expand_paths
+      existence_paths_hash = existence_path
 
       log.debug { "tailing paths: target = #{target_paths.join(",")} | existing = #{existence_paths.join(",")}" }
 
-      unwatched = existence_paths - target_paths
-      added = target_paths - existence_paths
+      unwatched_hash = existence_paths_hash.reject {|key, value| target_paths_hash.key?(key)}
+      added_hash = target_paths_hash.reject {|key, value| existence_paths_hash.key?(key)}
 
-      stop_watchers(unwatched, immediate: false, unwatched: true) unless unwatched.empty?
-      start_watchers(added) unless added.empty?
+      stop_watchers(unwatched_hash, immediate: false, unwatched: true) unless unwatched_hash.empty?
+      start_watchers(added_hash) unless added_hash.empty?
     end
 
-    def setup_watcher(path, pe)
+    def setup_watcher(target_info, pe)
       line_buffer_timer_flusher = @multiline_mode ? TailWatcher::LineBufferTimerFlusher.new(log, @multiline_flush_interval, &method(:flush_buffer)) : nil
-      tw = TailWatcher.new(path, pe, log, @read_from_head, method(:update_watcher), line_buffer_timer_flusher, method(:io_handler))
+      tw = TailWatcher.new(target_info, pe, log, @read_from_head, @follow_inodes, method(:update_watcher), line_buffer_timer_flusher, method(:io_handler))
 
       if @enable_watch_timer
         tt = TimerTrigger.new(1, log) { tw.on_notify }
@@ -350,47 +392,60 @@ module Fluent::Plugin
       raise e
     end
 
-    def start_watchers(paths)
-      paths.each { |path|
+    def start_watchers(targets_info)
+      targets_info.each_value { |target_info|
         pe = nil
         if @pf
-          pe = @pf[path]
+          pe = @pf[target_info]
           if @read_from_head && pe.read_inode.zero?
             begin
-              pe.update(Fluent::FileWrapper.stat(path).ino, 0)
+              pe.update(Fluent::FileWrapper.stat(target_info.path).ino, 0)
             rescue Errno::ENOENT
-              $log.warn "#{path} not found. Continuing without tailing it."
+              $log.warn "#{target_info.path} not found. Continuing without tailing it."
             end
           end
         end
 
         begin
-          tw = setup_watcher(path, pe)
+          tw = setup_watcher(target_info, pe)
         rescue WatcherSetupError => e
-          log.warn "Skip #{path} because unexpected setup error happens: #{e}"
+          log.warn "Skip #{target_info.path} because unexpected setup error happens: #{e}"
           next
         end
-        @tails[path] = tw
+
+        begin
+          target_info = TargetInfo.new(target_info.path, Fluent::FileWrapper.stat(target_info.path).ino)
+          @tails[target_info] = tw
+        rescue Errno::ENOENT
+          $log.warn "stat() for #{target_info.path} failed with ENOENT. Drop tail watcher for now."
+          # explicitly detach and unwatch watcher `tw`.
+          tw.unwatched = true
+          detach_watcher(tw, target_info.ino, false)
+        end
       }
     end
 
-    def stop_watchers(paths, immediate: false, unwatched: false, remove_watcher: true)
-      paths.each { |path|
-        tw = remove_watcher ? @tails.delete(path) : @tails[path]
+    def stop_watchers(targets_info, immediate: false, unwatched: false, remove_watcher: true)
+      targets_info.each_value { |target_info|
+        if remove_watcher
+          tw = @tails.delete(target_info)
+        else
+          tw = @tails[target_info]
+        end
         if tw
           tw.unwatched = unwatched
           if immediate
-            detach_watcher(tw, false)
+            detach_watcher(tw, target_info.ino, false)
           else
-            detach_watcher_after_rotate_wait(tw)
+            detach_watcher_after_rotate_wait(tw, target_info.ino)
           end
         end
       }
     end
 
     def close_watcher_handles
-      @tails.keys.each do |path|
-        tw = @tails.delete(path)
+      @tails.keys.each do |target_info|
+        tw = @tails.delete(target_info)
         if tw
           tw.close
         end
@@ -398,25 +453,39 @@ module Fluent::Plugin
     end
 
     # refresh_watchers calls @tails.keys so we don't use stop_watcher -> start_watcher sequence for safety.
-    def update_watcher(path, pe)
-      log.info("detected rotation of #{path}; waiting #{@rotate_wait} seconds")
+    def update_watcher(target_info, pe)
+      log.info("detected rotation of #{target_info.path}; waiting #{@rotate_wait} seconds")
 
       if @pf
-        unless pe.read_inode == @pf[path].read_inode
+        pe_inode = pe.read_inode
+        target_info_from_position_entry = TargetInfo.new(target_info.path, pe_inode)
+        unless pe_inode == @pf[target_info_from_position_entry].read_inode
           log.debug "Skip update_watcher because watcher has been already updated by other inotify event"
           return
         end
       end
-      rotated_tw = @tails[path]
-      @tails[path] = setup_watcher(path, pe)
-      detach_watcher_after_rotate_wait(rotated_tw) if rotated_tw
+
+      rotated_target_info = TargetInfo.new(target_info.path, pe.read_inode)
+      rotated_tw = @tails[rotated_target_info]
+      new_target_info = target_info.dup
+
+      if @follow_inodes
+        new_position_entry = @pf[target_info]
+
+        if new_position_entry.read_inode == 0
+          @tails[new_target_info] = setup_watcher(new_target_info, new_position_entry)
+        end
+      else
+        @tails[new_target_info] = setup_watcher(new_target_info, pe)
+      end
+      detach_watcher_after_rotate_wait(rotated_tw, pe.read_inode) if rotated_tw
     end
 
     # TailWatcher#close is called by another thread at shutdown phase.
     # It causes 'can't modify string; temporarily locked' error in IOHandler
     # so adding close_io argument to avoid this problem.
     # At shutdown, IOHandler's io will be released automatically after detached the event loop
-    def detach_watcher(tw, close_io = true)
+    def detach_watcher(tw, ino, close_io = true)
       tw.watchers.each do |watcher|
         event_loop_detach(watcher)
       end
@@ -425,15 +494,16 @@ module Fluent::Plugin
       tw.close if close_io
 
       if tw.unwatched && @pf
-        @pf.unwatch(tw.path)
+        target_info = TargetInfo.new(tw.path, ino)
+        @pf.unwatch(target_info)
       end
     end
 
-    def detach_watcher_after_rotate_wait(tw)
+    def detach_watcher_after_rotate_wait(tw, ino)
       # Call event_loop_attach/event_loop_detach is high-cost for short-live object.
       # If this has a problem with large number of files, use @_event_loop directly instead of timer_execute.
       timer_execute(:in_tail_close_watcher, @rotate_wait, repeat: false) do
-        detach_watcher(tw)
+        detach_watcher(tw, ino)
       end
     end
 
@@ -601,10 +671,12 @@ module Fluent::Plugin
     end
 
     class TailWatcher
-      def initialize(path, pe, log, read_from_head, update_watcher, line_buffer_timer_flusher, io_handler_build)
-        @path = path
+      def initialize(target_info, pe, log, read_from_head, follow_inodes, update_watcher, line_buffer_timer_flusher, io_handler_build)
+        @path = target_info.path
+        @ino = target_info.ino
         @pe = pe || MemoryPositionEntry.new
         @read_from_head = read_from_head
+        @follow_inodes = follow_inodes
         @update_watcher = update_watcher
         @log = log
         @rotate_handler = RotateHandler.new(log, &method(:on_rotate))
@@ -614,7 +686,7 @@ module Fluent::Plugin
         @watchers = []
       end
 
-      attr_reader :path
+      attr_reader :path, :ino
       attr_reader :pe
       attr_reader :line_buffer_timer_flusher
       attr_accessor :unwatched  # This is used for removing position entry from PositionFile
@@ -709,7 +781,23 @@ module Fluent::Plugin
           end
 
           if watcher_needs_update
-            @update_watcher.call(@path, swap_state(@pe))
+            if @follow_inodes
+              # No need to update a watcher if stat is nil (file not present), because moving to inodes will create
+              # new watcher, and old watcher will be closed by stop_watcher in refresh_watchers method
+              # don't want to swap state because we need latest read offset in pos file even after rotate_wait
+              if stat
+                target_info = TargetInfo.new(@path, stat)
+                @update_watcher.call(target_info, @pe)
+              end
+            else
+              # Permit to handle if stat is nil (file not present).
+              # If a file is mv-ed and a new file is created during
+              # calling `#refresh_watchers`s, and `#refresh_watchers` won't run `#start_watchers`
+              # and `#stop_watchers()` for the path because `target_paths_hash`
+              # always contains the path.
+              target_info = TargetInfo.new(@path, stat ? stat.ino : nil)
+              @update_watcher.call(target_info, swap_state(@pe))
+            end
           else
             @log.info "detected rotation of #{@path}"
             @io_handler = io_handler

data/lib/fluent/plugin/in_tail/position_file.rb

@@ -22,34 +22,40 @@ module Fluent::Plugin
       UNWATCHED_POSITION = 0xffffffffffffffff
       POSITION_FILE_ENTRY_REGEX = /^([^\t]+)\t([0-9a-fA-F]+)\t([0-9a-fA-F]+)/.freeze
 
-      def self.load(file, logger:)
-        pf = new(file, logger: logger)
+      def self.load(file, follow_inodes, existing_paths, logger:)
+        pf = new(file, follow_inodes, existing_paths, logger: logger)
         pf.load
         pf
       end
 
-      def initialize(file, logger: nil)
+      def initialize(file, follow_inodes, existing_paths, logger: nil)
         @file = file
         @logger = logger
         @file_mutex = Mutex.new
         @map = {}
+        @follow_inodes = follow_inodes
+        @existing_paths = existing_paths
       end
 
-      def [](path)
-        if m = @map[path]
+      def [](target_info)
+        if m = @map[@follow_inodes ? target_info.ino : target_info.path]
           return m
         end
 
         @file_mutex.synchronize {
           @file.seek(0, IO::SEEK_END)
-          seek = @file.pos + path.bytesize + 1
-          @file.write "#{path}\t0000000000000000\t0000000000000000\n"
-          @map[path] = FilePositionEntry.new(@file, @file_mutex, seek, 0, 0)
+          seek = @file.pos + target_info.path.bytesize + 1
+          @file.write "#{target_info.path}\t0000000000000000\t0000000000000000\n"
+          if @follow_inodes
+            @map[target_info.ino] = FilePositionEntry.new(@file, @file_mutex, seek, 0, 0)
+          else
+            @map[target_info.path] = FilePositionEntry.new(@file, @file_mutex, seek, 0, 0)
+          end
         }
       end
 
-      def unwatch(path)
-        if (entry = @map.delete(path))
+      def unwatch(target_info)
+        if (entry = @map.delete(@follow_inodes ? target_info.ino : target_info.path))
           entry.update_pos(UNWATCHED_POSITION)
         end
       end
@@ -69,7 +75,11 @@ module Fluent::Plugin
             pos = m[2].to_i(16)
             ino = m[3].to_i(16)
             seek = @file.pos - line.bytesize + path.bytesize + 1
-            map[path] = FilePositionEntry.new(@file, @file_mutex, seek, pos, ino)
+            if @follow_inodes
+              map[ino] = FilePositionEntry.new(@file, @file_mutex, seek, pos, ino)
+            else
+              map[path] = FilePositionEntry.new(@file, @file_mutex, seek, pos, ino)
+            end
           end
         end
 
@@ -94,8 +104,9 @@ module Fluent::Plugin
             @file.truncate(0)
             @file.write(entries.values.map(&:to_entry_fmt).join)
 
-            entries.each do |path, val|
-              if (m = @map[path])
+            # entry contains path/ino key and value.
+            entries.each do |key, val|
+              if (m = @map[key])
                 m.seek = val.seek
               end
             end
@@ -139,13 +150,25 @@ module Fluent::Plugin
               @logger.warn("#{path} already exists. use latest one: deleted #{entries[path]}") if @logger
             end
 
-            entries[path] = Entry.new(path, pos, ino, file_pos + path.size + 1)
+            if @follow_inodes
+              entries[ino] = Entry.new(path, pos, ino, file_pos + path.size + 1)
+            else
+              entries[path] = Entry.new(path, pos, ino, file_pos + path.size + 1)
+            end
             file_pos += line.size
           end
         end
 
+        entries = remove_deleted_files_entries(entries, @existing_paths) if @follow_inodes
         entries
       end
+
+      def remove_deleted_files_entries(existent_entries, existing_paths)
+        filtered_entries = existent_entries.select {|file_entry|
+          existing_paths.key?(file_entry)
+        }
+        filtered_entries
+      end
     end
 
     Entry = Struct.new(:path, :pos, :ino, :seek) do
@@ -224,5 +247,21 @@ module Fluent::Plugin
         @inode
       end
     end
+
+    TargetInfo = Struct.new(:path, :ino) do
+      def ==(other)
+        return false unless other.is_a?(TargetInfo)
+        self.path == other.path
+      end
+
+      def hash
+        self.path.hash
+      end
+
+      def eql?(other)
+        return false unless other.is_a?(TargetInfo)
+        self.path == other.path
+      end
+    end
   end
 end