db2s3 0.2.6 → 0.3.0

data/HISTORY

@@ -1,4 +1,7 @@
-0.2.6
+0.3.0 (6 Dec 2009)
+  - Added db2s3:backup:clean task to delete old backups
+  - Added dependency on activesupport for the clean task
+0.2.6 (6 Dec 2009)
   - Remove metrics task, since it was far too inaccurate
   - Add statistics task to show you the size of your tables
   - Only add username to mysql command line if provided in database.yml

data/README

@@ -27,11 +27,9 @@ Usage:
   rake db2s3:backup:incremental # Unimplemented
 
   # Handy tasks
-  rake db2s3:metrics         # Estimated costs
+  rake db2s3:statistics      # Shows you the size of your DB
   rake db2s3:backup:restore  # You should be testing this regularly
-
-Caveats:
-  Currently does not clean up old back ups
+  rake db2s3:backup:clear    # Clean up old backups - cron this
 
 Kudos:
   http://github.com/pauldowman/blog_code_examples/tree/master/mysql_s3_backup

data/VERSION

@@ -1 +1 @@
-0.2.6
+0.3.0

data/lib/db2s3.rb

@@ -1,3 +1,4 @@
+require 'activesupport'
 require 'aws/s3'
 require 'tempfile'
 
@@ -20,6 +21,38 @@ class DB2S3
     run "gunzip -c #{file.path} | mysql #{mysql_options}"
   end
 
+  # TODO: This method really needs specs
+  def clean
+    to_keep = []
+    filelist = store.list
+    files = filelist.reject {|file| file.ends_with?(most_recent_dump_file_name) }.collect do |file|
+      {
+        :path => file,
+        :date => Time.parse(file.split('-').last.split('.').first)
+      }
+    end
+    # Keep all backups from the past day
+    files.select {|x| x[:date] >= 1.day.ago }.each do |backup_for_day|
+      to_keep << backup_for_day
+    end
+
+    # Keep one backup per day from the last week
+    files.select {|x| x[:date] >= 1.week.ago }.group_by {|x| x[:date].strftime("%Y%m%d") }.values.each do |backups_for_last_week|
+      to_keep << backups_for_last_week.sort_by{|x| x[:date].strftime("%Y%m%d") }.first
+    end
+
+    # Keep one backup per week since forever
+    files.group_by {|x| x[:date].strftime("%Y%W") }.values.each do |backups_for_week|
+      to_keep << backups_for_week.sort_by{|x| x[:date].strftime("%Y%m%d") }.first
+    end
+
+    to_destroy = filelist - to_keep.uniq.collect {|x| x[:path] }
+    to_destroy.delete_if {|x| x.ends_with?(most_recent_dump_file_name) }
+    to_destroy.each do |file|
+      store.delete(file.split('/').last)
+    end
+  end
+
   private
 
   def dump_db
@@ -88,6 +121,17 @@ class DB2S3
       file
     end
 
+    def list
+      ensure_connected
+      AWS::S3::Bucket.find(bucket).objects.collect {|x| x.path }
+    end
+
+    def delete(file_name)
+      if object = AWS::S3::S3Object.find(file_name, bucket)
+        object.delete
+      end
+    end
+
     private
 
     def bucket

data/tasks/tasks.rake

@@ -10,26 +10,31 @@ namespace :db2s3 do
       DB2S3.new.restore
     end
 
-    desc "Show table sizes for your database"
-    task :statistics => :environment do
-      # From http://mysqlpreacher.com/wordpress/tag/table-size/
-      results = ActiveRecord::Base.connection.execute(<<-EOS)
-      SELECT
-        engine,
-        ROUND(data_length/1024/1024,2) total_size_mb,
-        ROUND(index_length/1024/1024,2) total_index_size_mb,
-        table_rows,
-        table_name article_attachment
-        FROM information_schema.tables
-        WHERE table_schema = 'rhnh_production'
-        ORDER BY 3 desc;
-      EOS
-
-      rows = []
-      header = [["Type", "Data MB", "Index", "Rows", "Name"], []]
-      results.each {|x| rows << x.to_a }
-      rows.sort_by {|x| -x[3].to_i }
-      puts (header + rows).collect {|x| x.join("\t") }
+    desc "Keep all backups for the last day, one per day for the last week, and one per week before that. Delete the rest."
+    task :clean => :environment do
+      DB2S3.new.clean
     end
   end
+
+  desc "Show table sizes for your database"
+  task :statistics => :environment do
+    # From http://mysqlpreacher.com/wordpress/tag/table-size/
+    results = ActiveRecord::Base.connection.execute(<<-EOS)
+    SELECT
+      engine,
+      ROUND(data_length/1024/1024,2) total_size_mb,
+      ROUND(index_length/1024/1024,2) total_index_size_mb,
+      table_rows,
+      table_name article_attachment
+      FROM information_schema.tables
+      WHERE table_schema = 'rhnh_production'
+      ORDER BY 3 desc;
+    EOS
+
+    rows = []
+    header = [["Type", "Data MB", "Index", "Rows", "Name"], []]
+    results.each {|x| rows << x.to_a }
+    rows.sort_by {|x| -x[3].to_i }
+    puts (header + rows).collect {|x| x.join("\t") }
+  end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification 
 name: db2s3
 version: !ruby/object:Gem::Version 
-  version: 0.2.6
+  version: 0.3.0
 platform: ruby
 authors: 
 - Xavier Shay
@@ -61,7 +61,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: 
-rubygems_version: 1.3.4
+rubygems_version: 1.3.5
 signing_key: 
 specification_version: 3
 summary: Summarize your gem