Note: This is currently an outline, and should be scripted later.

  1. Stop the Thalia IME front end and the Alfresco repository server.
  2. Perform a dump of the MySQL database, using map-backup.py.
  3. Use sed to replace the declaratives for Latin1 to UTF8 in the sql file.
    sed s/latin1/utf8/g CMS-PROTO-TH3.MIT.EDU-db-thaliaproto-2008-09-03-15-37.sql > CMS-PROTO-TH3.MIT.EDU-db-thaliaunicode-2008-09-03-15-37.sql.temp
    
  4. Replace the database name with the name of the new database.
    sed -i s/thaliaproto/thaliaunicode/g CMS-PROTO-TH3.MIT.EDU-db-thaliaunicode-2008-09-03-15-37.sql.temp
    
  5. Use pluto to change the Latin1 and non-standard text to Unicode encoding.
    cat UTF-8 CMS-PROTO-TH3.MIT.EDU-db-thaliaunicode-2008-09-03-15-37.sql.temp | pluto -c > CMS-PROTO-TH3.MIT.EDU-db-thaliaunicode-2008-09-03-15-37.sql
    
    • Pluto is not install by default on server ops systems. It is available on Trogdor, and will need to be built, suggested to do this on a office desktop, and scp the sql files there and back again.
  6. Remove the temporary file.
    rm CMS-PROTO-TH3.MIT.EDU-db-thaliaunicode-2008-09-03-15-37.sql.temp
    
  • No labels