Skip to main content
Repairs the two columns search reads for “Upload HTML” content: a File’s extracted indexed_content and the source_file_id reverse index that lets a Page embedding that File be found by the File’s text. A clean upgrade needs nothing from this command — both columns are backfilled once by the migrations that introduce them. Run it instead when that one-shot backfill could not have covered every row, for example:
  • A prior migration run partially failed on an unreadable attachment.
  • Content was restored from a backup taken before the columns existed.
  • Rows need to be re-derived after the text-extraction logic itself changes.
The command is safe to run repeatedly: each step only writes a row whose current value disagrees with what a fresh save would produce, so a second run reports nothing changed. Options: Examples:
Example output from a dry run against an installation with one “Upload HTML” page whose columns had been cleared:
After a real run, trigger a search reindex (POST /portal/reindexSearch) so external search engines pick up the backfilled content.