BulkSpamCleanup

From Yosemite Manifesto
Jump to: navigation, search

This page is only for administrative use.

To delete a single user, see Special:UserMerge.

This is how I (DBooth) did a bulk cleanup of wiki spam:

1. Enable the extensions for batch deletion in wiki/LocalSettings.php on the server:

################# Extensions for batch deletion ####################
require_once "$IP/extensions/Nuke/Nuke.php";
require_once "$IP/extensions/DeleteBatch/DeleteBatch.php";
require_once "$IP/extensions/DynamicPageList/DynamicPageList.php";
require_once "$IP/extensions/NukeDPL/NukeDPL.php";

2. Check the list of PagesNotCreatedByAdmin

3. Nuke the pages not created by Admin. If it looked like they should all or almost all be deleted, go to Special:NukeDPL and paste into the form:

  notmodifiedby = Admin
  nottitlematch = Main\_Page

and submit.

Then uncheck any that should NOT be deleted, and submit.

4. Delete all revision history. WARNING: this applies to *all* pages! No 'undo' will be possible for any page after this! See http://www.mediawiki.org/wiki/Manual:DeleteArchivedRevisions.php

cd wiki/maintenance
php deleteArchivedRevisions.php
# And if the number of revisions to delete looks correct, do:
php deleteArchivedRevisions.php --delete

5. Delete unused accounts (see http://www.mediawiki.org/wiki/Manual:RemoveUnusedAccounts.php ):

cd wiki/maintenance
php removeUnusedAccounts.php
# And if the result looks correct, do:
php removeUnusedAccounts.php --delete

6. Manually delete other spam users by merging them to Anonymous account and selecting the delete option:

7. Disable the extensions for batch deletion in wiki/LocalSettings.php on the server, because they are too dangerous to leave enabled:

################# Extensions for batch deletion ####################
# require_once "$IP/extensions/Nuke/Nuke.php";
# require_once "$IP/extensions/DeleteBatch/DeleteBatch.php";
# require_once "$IP/extensions/DynamicPageList/DynamicPageList.php";
# require_once "$IP/extensions/NukeDPL/NukeDPL.php";

Background Info

These are pages that I used to figure out how to do the above.