DSpace Import, Export & Backup Mukesh A Pund Scientist NISCAIR

Backup Vs. Export/Import  Backup is meant for guarding the data from disk crash, virus attack, hacking or any calamity  Export/Import is meant for exchange of digital objects across repositories

Hardware Required for Backup  Any one of the following  CD-ROM/DVD-ROM  DAT Drive  External Hard disk  Another system on the LAN

DSpace Directory Structure  /dspace/assetstore  /dspace/assetstore (bitstreams – most important)  /dspace/bin (commands to be used at command line, can always be generated from dspace-source files  /dspace/config (you might have customized it, one time backup is good enough)  /dspace/handle-server

DSpace Directory Structure  /dspace/lib (can always be generated from dspace source)  /dspace/logs (essential to generate statistical reports and bug tracing)  /dspace/reports (can be generated from Log files)  /dspace/search (can be regenerated using index-all command)

Where DSpace stores data  /dspace/assetstore directory will have all the  Bitstreams and licenses  PostgreSQL databases contains information on  Communities  Collections  e-groups  E-persons, thier passwords  Host of other information

What should be Backedup  Your DSpace postgreSQL database  /dspace/assetstore (minimum backup)  /dspace (entire directory)

Creating Backup Directory  Create one directory where backup files will be stored  Eg  #mkdir /dspacebkp  #chmod 777 /dspacebkp

tar Command (compress)  To back up /dspace directory  $tar zcvf /dspacebkp/dspace150508.tar.gz /dspace  To back up only /dspace/assetstore  $tar zcvf /dspace/asset150508.tar.gz /dspace/assetstore

Untar (uncompress)  To untar and unzip the tar.gz file, you may use the following command  $tar zxvf /dspacebkp/dspace150508.tar.gz



WARNING:

The safer approach is to use the above command in temp directory and copy it to dspace directory only after successfully untaring the file

Backup of database The following commands are for Postgresql database backup  Run pg_dump as dspace user  Ex: $ su - /dspace  /usr/local/pgsql/bin/pg_dump dspace > /dspacebkp/dspace_db_150508

Backup of database Where  dspace is name of the database  /dspacebkp/dspace_db_150508 file is backup file in which all the table definitions and contents will be stored

Restoring the backup data One can use any of the following commands:  psql command OR  pg_restore

Restoring the Database 

WARNING

: You do not need to restore, unless your data got corrupted.  Not to be used as a routine  Of course backup should be done periodically

Using psql to Restore  $ psql -d dspace -f /dspacebkp/dspace_db_150508  Where dspace is the name of database  dspace_db_150508 is the backup file taken on 15 th May 2008.

Using pg_restore to Restore  pg_restore -d dspace /dspacebkp/dspace_db_150508 More options of pg_restore can be explored by  $ man pg_restore

Export/Import in Dspace  Not to be used as a backup mechanism  Export and import deal only with bitstreams, metadata, license and handles  You can export or Import  An item  A collection or

Export dsrun org.dspace.app.itemexport.ItemExport \ --type=COLLECTION \ --id=collID \ --dest=dest_dir \ --number=seq_num \ Where --type can have either the value COLLECTION or ITEM --id is the handle/collection_or_Item_Id ex: 1849/2 (or 123456789/2 in case you do not have handle) --dest is destination directory, if does note exist create --number is sequence number it can be just 1

Importing dsrun org.dspace.app.itemimport.ItemImport \ --add \ [email protected] \ --collection=collectionID \ --source=items_dir --mapfile=mapfile Where --add or --replace or --remove --mapfile can be used later to remove uploaded items

What is exported  The following files will be created for every item  dublin_core.xml ( metadata)  Handle ( one line having the handle number)  license.txt

 Actual file ( bitstream: could be pdf or doc or an image file)  Contents (with two lines – license file name, and actual bitstream name)