about summary refs log tree commit diff
path: root/app
diff options
context:
space:
mode:
authormultiple creatures <dev@multiple-creature.party>2019-08-09 19:00:58 -0500
committermultiple creatures <dev@multiple-creature.party>2019-08-09 19:00:58 -0500
commite466f9c2ceea2dad79f516256d989d797a1f780d (patch)
treef2bc0f4bc3446d02e77f3b1dc4a4ec98c65f7fe7 /app
parentd8156acb0641d5e52b62fff5c323148ad1b6eaf8 (diff)
janitor - add db & media pruning job
Diffstat (limited to 'app')
-rw-r--r--app/workers/scheduler/janitor_scheduler.rb40
1 files changed, 40 insertions, 0 deletions
diff --git a/app/workers/scheduler/janitor_scheduler.rb b/app/workers/scheduler/janitor_scheduler.rb
index 25a67a9a3..4ebed8fa2 100644
--- a/app/workers/scheduler/janitor_scheduler.rb
+++ b/app/workers/scheduler/janitor_scheduler.rb
@@ -4,6 +4,7 @@ class Scheduler::JanitorScheduler
   include Sidekiq::Worker
   include BlocklistHelper
   include ModerationHelper
+  include Redisable
 
   MIN_POSTS = 6
 
@@ -24,6 +25,7 @@ class Scheduler::JanitorScheduler
     import_blocklists!
     export_suspensions!
     export_activityrelay_config!
+    prune_database! unless redis.exists('janitor:pune_database')
   end
 
   private
@@ -32,6 +34,44 @@ class Scheduler::JanitorScheduler
     Account.local.where.not(suspended_at: nil).destroy_all
   end
 
+  def prune_database!
+    suspended_accounts = Account.where.not(suspended_at: nil).select(:id)
+    suspended_domains = DomainBlock.suspend.select(:domain)
+
+    # remove statuses from suspended accounts missed by SuspendStatusService
+    # if its sidekiq job crashed
+    Status.where(account_id: suspended_accounts).in_batches do |status|
+      BatchedRemoveStatusService.new.call(status)
+    end
+
+    # prune leaves of threads that lost their context after a suspension
+    # keeping these around eats a pretty good amount of storage
+    deleted_mentions = Mention.where(account_id: suspended_accounts).select(:status_id)
+    Status.remote.where(account_id: deleted_mentions).in_batches do |status|
+      BatchedRemoveStatusService.new.call(status)
+    end
+
+    # remove mention entries that have no status or account attached to them
+    Mention.where(account_id: nil).in_batches.destroy_all
+    Mention.where(status_id: nil).in_batches.destroy_all
+
+    # remove media attachments that don't belong to any status
+    MediaAttachment.where(status_id: nil).in_batches.destroy_all
+
+    # remove custom emoji from suspended domains missed by SuspendAccountService
+    CustomEmoji.where(domain: suspended_domains).in_batches.destroy_all
+
+    # prune empty tags
+    Tag.all.find_each { |tag| tag.destroy unless tag.statuses.exists? }
+
+    # remove audit log entries with missing context
+    # we already use LOG_USER to avoid that problem
+    Admin::ActionLog.where.not(target_id: Account.select(:id)).in_batches.destroy_all
+    Admin::ActionLog.where.not(account_id: Account.local.select(:id)).in_batches.destroy_all
+
+    redis.setex('janitor:prune_database', 1.day, 1)
+  end
+
   def suspend_abandoned_accounts!
     reason = "Appears to be abandoned. Freeing up the username for someone else."
     abandoned_accounts.find_each do |account|