Sequence database migration

If you had only a few, mostly public databases active on your old server, it is easiest to select them from the predefined database definitions in Mascot Database Manager on the new server. Doing this means that the configurations will be kept up-to-date automatically, and makes it easy to schedule automatic updating of the database files.

If you had a large number of active databases on your old server that you had configured manually, you may wish to copy across the files and configuration information. This can only be done before running Database Manager for the first time on the new server.

If the old server was Mascot 2.4 or later but Database Manager had never been run, or if the old server was Mascot 2.3 or earlier, follow Procedure 1.

If the old server was Mascot 2.4 or later and Database Manager had been run at least once, follow Procedure 2.

Procedure 1

To be used when Database Manager has never been run on the old server.

  1. Copy the sequence database directories and files to the new server.
  2. If any databases had taxonomy defined, and the taxonomy files (in the taxonomy directory) are newer than those on the new server, copy these files across.
  3. In a text editor, open the old and new mascot.dat files, side-by-side. Make sure you are working with the latest file from the old server, and not some backup (e.g. mascot.dat.34).
  4. Copy the following sections from the old file to the new, replacing the corresponding sections in the new file:
    • Databases
    • PARSE
    • WWW
    Each section starts with the keyword above and ends with word end. Lines starting with # are comments.
  5. If the paths to either the program files or the database files have changed, use search and replace to update the paths. Make sure you don’t accidentally change text in some other part of mascot.dat
  6. If you have defined any additional taxonomy sections in the old mascot.dat that are not present in the new file, copy these across. (These sections start with Taxonomy_n, where n is an index from 1 upwards). If the index is already being used in the new file, you’ll need to choose a new one and change the references in the Databases section.
  7. Double check your changes and save the new file.
  8. Carefully read the help page for Mascot Database Manager, then start it for the first time.
  9. For most of your old database definitions, select Keep as Custom. In some cases, Database Manager will recognise that an old definition is very similar to one of the predefined database definitions, and suggest that it be synchronised. If the match is correct, then best to select Synchronise.
  10. Before you can enable automatic updating of the migrated database files, you will need to edit their definition(s) in Database Manager to add download URL(s).

Procedure 2

To be used when Database Manager has been run at least once on the old server.

  1. Copy the sequence database directories and files to the new server.
  2. If any databases had taxonomy defined, and the taxonomy files (in the taxonomy directory) are newer than those on the new server, copy these files across.
  3. Copy these files from under the Mascot directory on the old server:
    • config/db_manager/public plus all the files it contains
    • config/db_manager/global.conf (will only exist if you have changed certain defaults such as proxy settings)
    • config/db_manager/configuration.xml
  4. If the paths to either the program files or the database files have changed, you’ll need to edit the following configuration files. Before editing a file, make a backup in some other directory.
    • If you have defined database update schedules, copy the entries in the Cron section in the old mascot.dat file to the new mascot.dat file, and fix the paths to dbman_add_task.pl and dbman_process_tasks.pl to point to the corresponding scripts in the new server.
    • In global.conf, change the setting sequence directory under [default directories] to point to the new default sequence directory.
    • In configuration.xml, do a global search and replace from the old sequence directory to the new sequence directory. The elements to target are msgd:current_path and msgd:incoming_path under msgd:databases. If you’ve set up sequence databases outside the Mascot sequence directory, you need to manually inspect the paths under msgd:databases and edit as necessary.
    • In configuration.xml, do a global search and replace from the old UniGene directory to the new UniGene directory. The elements to target are msgd:path under msgd:unigene_entries.
    • In configuration.xml, delete <msgd:mascot_dat_md5>, </msgd:mascot_dat_md5>, and everything in between.
  5. Run Database Manager. It will detect that configuration.xml doesn’t match mascot.dat anymore. Click on Fix mascot.dat.