Converting Parallels H-Sphere Database To UNICODE

The system database must be in UNICODE (UTF-8).

To convert your database to Unicode:

1. Stop the control panel

   Log in as root and stop the control panel:

   For Linux:

   /etc/rc.d/init.d/httpdcp stop
   killall -9 java

   For FreeBSD:

   /usr/local/etc/rc.d/apachecp.sh stop
   killall -9 java

2. Find out your current database encoding

   Type:

   su -l cpanel -c 'psql hsphere'
   hsphere# \encoding

   If the encoding is UNICODE (UTP-8), you have found what you need. If not, the next step is to dump Parallels H-Sphere system database.

3. Dump Parallels H-Sphere system database
   1. Create and enter backup directory:

      mkdir pg_backup
      cd pg_backup

   2. Get the password for wwwuser. You'll need it to query the database:

      cat ~cpanel/shiva/psoft_config/hsphere.properties | grep PASS

   3. Dump Parallels H-Sphere system database.

      Export schema:

      pg_dump -u -s -f schema.db hsphere
      chmod 600 schema.db
      cp -p schema.db schema_backup.db

      Export data:

      pg_dump -u -a -f data.db hsphere
      chmod 600 data.db
      cp -p data.db data_backup.db

      Notes:

      1. If your system database is large, the dump can take several hours to complete. You can speed it up by setting

         fsync=off

         in postgresql.conf. When you are done, unset this option back for safety reasons.

      2. The dump file is created with 644 permissions by default; you need to set more secure 600 permissions to prevent the data from being read by other users.
   4. For additional security, you may disallow access to the backup directory for all other users:

      chmod 700

4. Convert the dump to UNICODE.

   Convert the dump into Unicode with the iconv utility.

   Linux:

   iconv --from-code=<REGIONAL_ENCODING> --to-code=UTF-8 -o utf_data.db data.db
   mv utf_data.db data.db

   FreeBSD:

   iconv -f <REGIONAL_ENCODING> -t UTF-8 data.db > utf_data.db
   mv utf_data.db data.db

   If your dump file exceeds 2GB:

   1. Split it into smaller files, 1GB each:

      split -b 1024m data.db data_db

   2. Run iconvfor for each of these files to convert them to UNICODE:

      iconv --from-code=<REGIONAL_ENCODING>--to-code=UTF-8 -o utf_data_db.aa data_db.aa
      iconv --from-code=<REGIONAL_ENCODING>--to-code=UTF-8 -o utf_data_db.ab data_db.ab
      ...

   3. Join them back into data.db:

      cat utf_data_db.aa utf_data_db.ab utf_data_db.ac ... > data.db

      
      Here, <REGIONAL_ENCODING> is the source encoding. For example, for native US English encoding:

      Linux:

      iconv --from-code=ISO-8859-1 --to-code=UTF-8 -o utf_data.db data.db

      FreeBSD:

      iconv -f ISO-8859-1 -t UTF-8 data.db > utf_data.db

      The resulting data.db file will contain the data converted to Unicode.

      For better security, run the following command:

      chmod 600 data.db

5. Save the postgres directory in a backup location.
   1. Stop the database:

      For Linux:

      /etc/rc.d/init.d/postgresql stop

      For FreeBSD:

      /usr/local/etc/rc.d/010.pgsql.sh stop

   2. Save the postgres directory:

      For Linux:

      cp -pR ~postgres/data ./

      For FreeBSD:

      cp -pR ~pgsql/data ./

   3. Start the database:

      For Linux:

      /etc/rc.d/init.d/postgresql start

      For FreeBSD:

      /usr/local/etc/rc.d/010.pgsql.sh start

6. Recreate Parallels H-Sphere database.
   1. Delete old Parallels H-Sphere database:

      # su -l cpanel
      $ dropdb hsphere

   2. Create database:

      createdb -E UNICODE -U wwwuser hsphere

   3. Create Parallels H-Sphere DB schema:

      psql -q -U wwwuser -f schema.db hsphere

   4. Import Parallels H-Sphere system data:

      psql -q -U wwwuser -f data.db hsphere

      Note: If you face problems with importing data, please see the Troubleshooting section in CP Acceleration guide.

   5. If you added

      fsync=off

      to postgresql.conf, don't forget to delete it.

   6. Start the Control Panel.