logstash-output-s3 4.3.7 → 4.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: 238e7be91fe40e4fcb80736f3e8d62e76b6f8108be35e5a0c054cb19bb239428
-  data.tar.gz: eb0c70181aa21d20b794cd05d2e458b322b10f6e343ae41d96a1e6f49cf85858
+  metadata.gz: 7fe328033b222f10871103a51430bfe6f6a269f15460f70e72e423b6b135927e
+  data.tar.gz: e845c48187f640a948624f7da5bf3b6cec3eee3b25c81f0024014494a71363e5
 SHA512:
-  metadata.gz: 1194c4ee3defe1104fcc1c68d914c20e2ca2548ba6405960c75d4b398a7cd9949287c3410ead913ff82b5183164c4bd5bf858fe836f94f63654c582efab78ab5
-  data.tar.gz: cb135c3b28297db0f5ad90a4f7ab4ba6b8c112574815e38bdccdec685f5caa28ed7636f68569586f71205d651118dad7cc7dd0194786efb59bfb8113fa7d0afb
+  metadata.gz: 9c5a89d3551c5d199b0b289c17b0a36138c4efbb4c4191610167039aab4692588a0c9fc17a158af788ca704d9fbd454a9f2de501b2aa34d17f5aee5820345829
+  data.tar.gz: 989eef2c121767e315199177f0d672bea2b2ff18fcb6a5a9acf21e8b813adadbcd35ab0ce221eff4a633bdaf735e9267b7943a3fbefb788dee42bdc4d1df293a

data/CHANGELOG.md

@@ -1,3 +1,6 @@
+## 4.4.0
+  - Logstash recovers corrupted gzip and uploads to S3 [#249](https://github.com/logstash-plugins/logstash-output-s3/pull/249)
+
 ## 4.3.7
   - Refactor: avoid usage of CHM (JRuby 9.3.4 work-around) [#248](https://github.com/logstash-plugins/logstash-output-s3/pull/248)
 

data/README.md

@@ -19,7 +19,7 @@ Need help? Try #logstash on freenode IRC or the https://discuss.elastic.co/c/log
 
 ## Developing
 
-### 1. Plugin Developement and Testing
+### 1. Plugin Development and Testing
 
 #### Code
 - To get started, you'll need JRuby with the Bundler gem installed.

data/VERSION

@@ -0,0 +1 @@
+4.4.0

data/lib/logstash/outputs/s3/size_rotation_policy.rb

@@ -7,7 +7,7 @@ module LogStash
 
         def initialize(size_file)
           if size_file <= 0
-            raise LogStash::ConfigurationError, "`size_file` need to be greather than 0"
+            raise LogStash::ConfigurationError, "`size_file` need to be greater than 0"
           end
 
           @size_file = size_file

data/lib/logstash/outputs/s3/temporary_file.rb

@@ -2,15 +2,23 @@
 require "thread"
 require "forwardable"
 require "fileutils"
+require "logstash-output-s3_jars"
 
 module LogStash
   module Outputs
     class S3
-      # Wrap the actual file descriptor into an utility classe
-      # It make it more OOP and easier to reason with the paths.
+
+      java_import 'org.logstash.plugins.outputs.s3.GzipUtil'
+
+      # Wrap the actual file descriptor into an utility class
+      # Make it more OOP and easier to reason with the paths.
       class TemporaryFile
         extend Forwardable
 
+        GZIP_EXTENSION = "txt.gz"
+        TXT_EXTENSION = "txt"
+        RECOVERED_FILE_NAME_TAG = "-recovered"
+
         def_delegators :@fd, :path, :write, :close, :fsync
 
         attr_reader :fd
@@ -33,8 +41,10 @@ module LogStash
         def size
           # Use the fd size to get the accurate result,
           # so we dont have to deal with fsync
-          # if the file is close we will use the File::size
+          # if the file is close, fd.size raises an IO exception so we use the File::size
           begin
+            # fd is nil when LS tries to recover gzip file but fails
+            return 0 unless @fd != nil
             @fd.size
           rescue IOError
             ::File.size(path)
@@ -45,7 +55,7 @@ module LogStash
           @key.gsub(/^\//, "")
         end
 
-        # Each temporary file is made inside a directory named with an UUID,
+        # Each temporary file is created inside a directory named with an UUID,
         # instead of deleting the file directly and having the risk of deleting other files
         # we delete the root of the UUID, using a UUID also remove the risk of deleting unwanted file, it acts as
         # a sandbox.
@@ -58,13 +68,46 @@ module LogStash
           size == 0
         end
 
+        # only to cover the case where LS cannot restore corrupted file, file is not exist
+        def recoverable?
+          !@fd.nil?
+        end
+
         def self.create_from_existing_file(file_path, temporary_folder)
           key_parts = Pathname.new(file_path).relative_path_from(temporary_folder).to_s.split(::File::SEPARATOR)
 
+          # recover gzip file and compress back before uploading to S3
+          if file_path.end_with?("." + GZIP_EXTENSION)
+            file_path = self.recover(file_path)
+          end
           TemporaryFile.new(key_parts.slice(1, key_parts.size).join("/"),
-                         ::File.open(file_path, "r"),
+                         ::File.exist?(file_path) ? ::File.open(file_path, "r") : nil, # for the nil case, file size will be 0 and upload will be ignored.
                          ::File.join(temporary_folder, key_parts.slice(0, 1)))
         end
+
+        def self.gzip_extension
+          GZIP_EXTENSION
+        end
+
+        def self.text_extension
+          TXT_EXTENSION
+        end
+
+        def self.recovery_file_name_tag
+          RECOVERED_FILE_NAME_TAG
+        end
+
+        private
+        def self.recover(file_path)
+          full_gzip_extension = "." + GZIP_EXTENSION
+          recovered_txt_file_path = file_path.gsub(full_gzip_extension, RECOVERED_FILE_NAME_TAG + "." + TXT_EXTENSION)
+          recovered_gzip_file_path = file_path.gsub(full_gzip_extension, RECOVERED_FILE_NAME_TAG + full_gzip_extension)
+          GzipUtil.recover(file_path, recovered_txt_file_path)
+          if ::File.exist?(recovered_txt_file_path) && !::File.zero?(recovered_txt_file_path)
+            GzipUtil.compress(recovered_txt_file_path, recovered_gzip_file_path)
+          end
+          recovered_gzip_file_path
+        end
       end
     end
   end

data/lib/logstash/outputs/s3/temporary_file_factory.rb

@@ -19,9 +19,6 @@ module LogStash
       # I do not have to mess around to check if the other directory have file in it before destroying them.
       class TemporaryFileFactory
         FILE_MODE = "a"
-        GZIP_ENCODING = "gzip"
-        GZIP_EXTENSION = "txt.gz"
-        TXT_EXTENSION = "txt"
         STRFTIME = "%Y-%m-%dT%H.%M"
 
         attr_accessor :counter, :tags, :prefix, :encoding, :temporary_directory, :current
@@ -48,7 +45,7 @@ module LogStash
 
         private
         def extension
-          gzip? ? GZIP_EXTENSION : TXT_EXTENSION
+          gzip? ? TemporaryFile.gzip_extension : TemporaryFile.text_extension
         end
 
         def gzip?

data/lib/logstash/outputs/s3/uploader.rb

@@ -31,6 +31,7 @@ module LogStash
           end
         end
 
+        # uploads a TemporaryFile to S3
         def upload(file, options = {})
           upload_options = options.fetch(:upload_options, {})
 
@@ -68,6 +69,7 @@ module LogStash
           @workers_pool.shutdown
           @workers_pool.wait_for_termination(nil) # block until its done
         end
+
       end
     end
   end

data/lib/logstash/outputs/s3.rb

@@ -97,6 +97,7 @@ class LogStash::Outputs::S3 < LogStash::Outputs::Base
                                                                    :fallback_policy => :caller_runs
                                                                  })
 
+  GZIP_ENCODING = "gzip"
 
   config_name "s3"
   default :codec, "line"
@@ -181,7 +182,7 @@ class LogStash::Outputs::S3 < LogStash::Outputs::Base
   config :tags, :validate => :array, :default => []
 
   # Specify the content encoding. Supports ("gzip"). Defaults to "none"
-  config :encoding, :validate => ["none", "gzip"], :default => "none"
+  config :encoding, :validate => ["none", GZIP_ENCODING], :default => "none"
 
   # Define the strategy to use to decide when we need to rotate the file and push it to S3,
   # The default strategy is to check for both size and time, the first one to match will rotate the file.
@@ -315,7 +316,7 @@ class LogStash::Outputs::S3 < LogStash::Outputs::Base
       :server_side_encryption => @server_side_encryption ? @server_side_encryption_algorithm : nil,
       :ssekms_key_id => @server_side_encryption_algorithm == "aws:kms" ? @ssekms_key_id : nil,
       :storage_class => @storage_class,
-      :content_encoding => @encoding == "gzip" ? "gzip" : nil,
+      :content_encoding => @encoding == GZIP_ENCODING ? GZIP_ENCODING : nil,
       :multipart_threshold => @upload_multipart_threshold
     }
   end
@@ -397,16 +398,48 @@ class LogStash::Outputs::S3 < LogStash::Outputs::Base
     @crash_uploader = Uploader.new(bucket_resource, @logger, CRASH_RECOVERY_THREADPOOL)
 
     temp_folder_path = Pathname.new(@temporary_directory)
-    Dir.glob(::File.join(@temporary_directory, "**/*"))
-      .select { |file| ::File.file?(file) }
-      .each do |file|
-      temp_file = TemporaryFile.create_from_existing_file(file, temp_folder_path)
-      if temp_file.size > 0
-        @logger.debug? && @logger.debug("Recovering from crash and uploading", :path => temp_file.path)
-        @crash_uploader.upload_async(temp_file, :on_complete => method(:clean_temporary_file), :upload_options => upload_options)
+    files = Dir.glob(::File.join(@temporary_directory, "**/*"))
+               .select { |file_path| ::File.file?(file_path) }
+    under_recovery_files = get_under_recovery_files(files)
+
+    files.each do |file_path|
+      # when encoding is GZIP, if file is already recovering or recovered and uploading to S3, log and skip
+      if under_recovery_files.include?(file_path)
+        unless file_path.include?(TemporaryFile.gzip_extension)
+          @logger.warn("The #{file_path} file either under recover process or failed to recover before.")
+        end
       else
-        clean_temporary_file(temp_file)
+        temp_file = TemporaryFile.create_from_existing_file(file_path, temp_folder_path)
+        # do not remove or upload if Logstash tries to recover file but fails
+        if temp_file.recoverable?
+          if temp_file.size > 0
+            @logger.debug? && @logger.debug("Recovering from crash and uploading", :path => temp_file.path)
+            @crash_uploader.upload_async(temp_file,
+                                         :on_complete => method(:clean_temporary_file),
+                                         :upload_options => upload_options)
+          else
+            clean_temporary_file(temp_file)
+          end
+        end
+      end
+    end
+  end
+
+  # figures out the recovering files and
+  # creates a skip list to ignore for the rest of processes
+  def get_under_recovery_files(files)
+    skip_files = Set.new
+    return skip_files unless @encoding == GZIP_ENCODING
+
+    files.each do |file_path|
+      if file_path.include?(TemporaryFile.recovery_file_name_tag)
+        skip_files << file_path
+        if file_path.include?(TemporaryFile.gzip_extension)
+          # also include the original corrupted gzip file
+          skip_files << file_path.gsub(TemporaryFile.recovery_file_name_tag, "")
+        end
       end
     end
+    skip_files
   end
 end

data/lib/logstash-output-s3_jars.rb

@@ -0,0 +1,4 @@
+# AUTOGENERATED BY THE GRADLE SCRIPT. DO NOT EDIT.
+
+require 'jar_dependencies'
+require_jar('org.logstash.plugins.outputs.s3', 'logstash-output-s3', '4.4.0')

data/lib/tasks/build.rake

@@ -0,0 +1,15 @@
+# encoding: utf-8
+require "jars/installer"
+require "fileutils"
+
+task :vendor do
+  exit(1) unless system './gradlew vendor'
+  version = File.read("VERSION").strip
+end
+
+desc "clean"
+task :clean do
+  ["build", "vendor/jar-dependencies", "Gemfile.lock"].each do |p|
+    FileUtils.rm_rf(p)
+  end
+end

data/logstash-output-s3.gemspec

@@ -1,13 +1,13 @@
 Gem::Specification.new do |s|
   s.name            = 'logstash-output-s3'
-  s.version         = '4.3.7'
+  s.version         = '4.4.0'
   s.licenses        = ['Apache-2.0']
   s.summary         = "Sends Logstash events to the Amazon Simple Storage Service"
   s.description     = "This gem is a Logstash plugin required to be installed on top of the Logstash core pipeline using $LS_HOME/bin/logstash-plugin install gemname. This gem is not a stand-alone program"
   s.authors         = ["Elastic"]
   s.email           = 'info@elastic.co'
   s.homepage        = "http://www.elastic.co/guide/en/logstash/current/index.html"
-  s.require_paths = ["lib"]
+  s.require_paths = ["lib", "vendor/jar-dependencies"]
 
   # Files
   s.files = Dir["lib/**/*","spec/**/*","*.gemspec","*.md","CONTRIBUTORS","Gemfile","LICENSE","NOTICE.TXT", "vendor/jar-dependencies/**/*.jar", "vendor/jar-dependencies/**/*.rb", "VERSION", "docs/**/*"]

data/spec/integration/restore_from_crash_spec.rb

@@ -7,18 +7,17 @@ require "stud/temporary"
 describe "Restore from crash", :integration => true do
   include_context "setup plugin"
 
-  let(:options) { main_options.merge({ "restore" => true, "canned_acl" => "public-read-write" }) }
-
   let(:number_of_files) { 5 }
   let(:dummy_content) { "foobar\n" * 100 }
-  let(:factory) { LogStash::Outputs::S3::TemporaryFileFactory.new(prefix, tags, "none", temporary_directory)}
 
   before do
     clean_remote_files(prefix)
   end
 
-
   context 'with a non-empty tempfile' do
+    let(:options) { main_options.merge({ "restore" => true, "canned_acl" => "public-read-write" }) }
+    let(:factory) { LogStash::Outputs::S3::TemporaryFileFactory.new(prefix, tags, "none", temporary_directory)}
+
     before do
       # Creating a factory always create a file
       factory.current.write(dummy_content)
@@ -41,6 +40,9 @@ describe "Restore from crash", :integration => true do
   end
 
   context 'with an empty tempfile' do
+    let(:options) { main_options.merge({ "restore" => true, "canned_acl" => "public-read-write" }) }
+    let(:factory) { LogStash::Outputs::S3::TemporaryFileFactory.new(prefix, tags, "none", temporary_directory)}
+
     before do
       factory.current
       factory.rotate!
@@ -63,5 +65,68 @@ describe "Restore from crash", :integration => true do
       expect(bucket_resource.objects(:prefix => prefix).count).to eq(0)
     end
   end
+
+  context "#gzip encoding" do
+    let(:options) { main_options.merge({ "restore" => true, "canned_acl" => "public-read-write", "encoding" => "gzip" }) }
+    let(:factory) { LogStash::Outputs::S3::TemporaryFileFactory.new(prefix, tags, "gzip", temporary_directory)}
+    describe "with empty recovered file" do
+      before do
+        # Creating a factory always create a file
+        factory.current.write('')
+        factory.current.fsync
+        factory.current.close
+      end
+
+      it 'should not upload and not remove temp file' do
+        subject.register
+        try(20) do
+          expect(bucket_resource.objects(:prefix => prefix).count).to eq(0)
+          expect(Dir.glob(File.join(temporary_directory, "*")).size).to eq(1)
+        end
+      end
+    end
+
+    describe "with healthy recovered, size is greater than zero file" do
+      before do
+        # Creating a factory always create a file
+        factory.current.write(dummy_content)
+        factory.current.fsync
+        factory.current.close
+
+        (number_of_files - 1).times do
+          factory.rotate!
+          factory.current.write(dummy_content)
+          factory.current.fsync
+          factory.current.close
+        end
+      end
+
+      it 'should recover, upload to S3 and remove temp file' do
+        subject.register
+        try(20) do
+          expect(bucket_resource.objects(:prefix => prefix).count).to eq(number_of_files)
+          expect(Dir.glob(File.join(temporary_directory, "*")).size).to eq(0)
+          expect(bucket_resource.objects(:prefix => prefix).first.acl.grants.collect(&:permission)).to include("READ", "WRITE")
+        end
+      end
+    end
+
+    describe "with failure when recovering" do
+      before do
+        # Creating a factory always create a file
+        factory.current.write(dummy_content)
+        factory.current.fsync
+      end
+
+      it 'should not upload to S3 and not remove temp file' do
+        subject.register
+        try(20) do
+          expect(bucket_resource.objects(:prefix => prefix).count).to eq(0)
+          expect(Dir.glob(File.join(temporary_directory, "*")).size).to eq(1)
+        end
+      end
+    end
+  end
+
 end
 

data/spec/outputs/s3/size_rotation_policy_spec.rb

@@ -25,11 +25,11 @@ describe LogStash::Outputs::S3::SizeRotationPolicy do
   end
 
   it "raises an exception if the `size_file` is 0" do
-    expect { described_class.new(0) }.to raise_error(LogStash::ConfigurationError, /need to be greather than 0/)
+    expect { described_class.new(0) }.to raise_error(LogStash::ConfigurationError, /need to be greater than 0/)
   end
 
   it "raises an exception if the `size_file` is < 0" do
-    expect { described_class.new(-100) }.to raise_error(LogStash::ConfigurationError, /need to be greather than 0/)
+    expect { described_class.new(-100) }.to raise_error(LogStash::ConfigurationError, /need to be greater than 0/)
   end
 
   context "#needs_periodic?" do

data/spec/supports/helpers.rb

@@ -5,6 +5,7 @@ shared_context "setup plugin" do
   let(:bucket) { ENV["AWS_LOGSTASH_TEST_BUCKET"] }
   let(:access_key_id) {  ENV["AWS_ACCESS_KEY_ID"] }
   let(:secret_access_key) { ENV["AWS_SECRET_ACCESS_KEY"] }
+  let(:session_token) { ENV["AWS_SESSION_TOKEN"] }
   let(:size_file) { 100 }
   let(:time_file) { 100 }
   let(:tags) { [] }
@@ -18,6 +19,7 @@ shared_context "setup plugin" do
       "temporary_directory" => temporary_directory,
       "access_key_id" => access_key_id,
       "secret_access_key" => secret_access_key,
+      "session_token" => session_token,
       "size_file" => size_file,
       "time_file" => time_file,
       "region" => region,
@@ -25,7 +27,7 @@ shared_context "setup plugin" do
     }
   end
 
-  let(:client_credentials) { Aws::Credentials.new(access_key_id, secret_access_key) }
+  let(:client_credentials) { Aws::Credentials.new(access_key_id, secret_access_key, session_token) }
   let(:bucket_resource) { Aws::S3::Bucket.new(bucket, { :credentials => client_credentials, :region => region }) }
 
   subject { LogStash::Outputs::S3.new(options) }

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: logstash-output-s3
 version: !ruby/object:Gem::Version
-  version: 4.3.7
+  version: 4.4.0
 platform: ruby
 authors:
 - Elastic
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2022-06-16 00:00:00.000000000 Z
+date: 2022-07-19 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   requirement: !ruby/object:Gem::Requirement
@@ -129,7 +129,9 @@ files:
 - LICENSE
 - NOTICE.TXT
 - README.md
+- VERSION
 - docs/index.asciidoc
+- lib/logstash-output-s3_jars.rb
 - lib/logstash/outputs/s3.rb
 - lib/logstash/outputs/s3/file_repository.rb
 - lib/logstash/outputs/s3/patch.rb
@@ -142,6 +144,7 @@ files:
 - lib/logstash/outputs/s3/uploader.rb
 - lib/logstash/outputs/s3/writable_directory_validator.rb
 - lib/logstash/outputs/s3/write_bucket_permission_validator.rb
+- lib/tasks/build.rake
 - logstash-output-s3.gemspec
 - spec/integration/dynamic_prefix_spec.rb
 - spec/integration/gzip_file_spec.rb
@@ -164,6 +167,7 @@ files:
 - spec/outputs/s3_spec.rb
 - spec/spec_helper.rb
 - spec/supports/helpers.rb
+- vendor/jar-dependencies/org/logstash/plugins/outputs/s3/logstash-output-s3/4.4.0/logstash-output-s3-4.4.0.jar
 homepage: http://www.elastic.co/guide/en/logstash/current/index.html
 licenses:
 - Apache-2.0
@@ -174,6 +178,7 @@ post_install_message:
 rdoc_options: []
 require_paths:
 - lib
+- vendor/jar-dependencies
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
   - - ">="