Frequently Asked Questions

From EPrints Documentation
Revision as of 12:36, 21 January 2022 by (talk | contribs) (No longer obsolete but certainly out of date)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

This all needs sorting out, it's just being grabbed from the old wiki for now.

What operating systems can we use?

EPrints should work on any Linux operating system. We use Ubuntu (Debian) and Redhat Enterprise Linux.

Running EPrints on Microsoft Windows is possible but strongly not recommended. It can be very difficult to get all the necessary dependencies in place. If you want to get EPrints running but do not have as computer running a Linux operating system. You could install VirtualBox and then create a Ubuntu or Redhat "virtual machine" (VM) to run on top of your Windows operating system. Alternatively, you could install Docker Desktop for Windows and install EPrints using Docker. However, both these methods are only recommended for testing or demoing EPrints. If you intend to run EPrints long-term you should install on a dedicated Linux server or virtual machine.

What computer do we need?

Any modern physical PC (server or desktop) or a virtual machine should be more than capable of running EPrints. A suggested minimum specification for running a live service:

  • 2GB RAM
  • 15GB disk space
  • 100Mb/s network speed

How much will it cost to set up?

Most of the costs are staff time. Equipment costs are a PC, an internet connection and a BACKUP STRATEGY. Please remember to budget for backups.

EPrints, and all the other software required to make it work, are available for no cost. At some point in the future EPrints may offer some pay-services, but the core software will remain zero cost and freely available to all.

See How much will it cost?

How much disk space will we need?

The main usage of disk space will be the uploaded files for the publications, (e.g. PDF, Word documents, etc.). On average an eprint item will have associated files (e.g. uploaded PDF and thumbnail and preview images of that PDF) that take up 2MB of disk space. So for every extra 500 eprint items you will need an extra 1GB of disk. However, if you intend to upload lots of videos, images, are large documents, then this requirement will be higher. On top of the space needed for uploaded documents and associated files you will need approximately 5GB for installing the operating system. The disk space required by the database is difficult to determine, as this is dependent on a number of different factors but 1MB per eprint item.

Example usage requirements

  • Small repository: 1000 eprint items
    • Operating system: 5GB
    • Upload and associated files: 2MB x 1000 = 2GB
    • Database: 1MB x 1000 = 1GB
    • TOTAL: 8GB (15GB to give room for expansion)
  • Medium-sized repository: 10,000 eprint items
    • Operating system: 5GB
    • Upload and associated files: 2MB x 10,000 = 20GB
    • Database: 1MB x 10,000 = 10GB
    • TOTAL: 35GB (50GB to give room for expansion)
  • Large repository: 100,000 eprint items
    • Operating system: 5GB
    • Upload and associated files: 2MB x 100,000 = 200GB
    • Database: 1MB x 100,000 = 100GB
    • TOTAL: 305GB (350-400GB to give room for expansion)

Is it possible when depositing a document, to just point to an "alternative location" rather than have the full text copied and held in the eprint archive?

Yes, and to suppress the normal warning about no documents attached look in [archive_root]/cfg/cfg.d/

I need to run apache as a user other than "eprints", what do I do to make EPrints work in this situation?

Example, apache is running as user "apache".

  • Make all the eprints files owned by "apache" instead of "eprints"
  • Edit SystemSettings?.pm to tell eprints to run as user "apache"
  • You'll need to run all command line scripts as user "apache"
  • All eprints cron jobs should be owned by user "apache"

If you are installing a new copy of eprints, you can specify the user and group to use when you run 'configure'. Do

./configure --help

for details.

How do I get the body HTML of a page without the template around it?

This is handy for dynamically linking EPrints content into other sites.

For "view" pages you need to add the option include=>1 to the view configuration. This will cause generate_views to make a .include page in addition to the .html page. The .include page will have no template around it.

For dynamic pages, those under /cgi/, you can add the cgi parameter mainonly=yes. E.g.

How do I get statistics on number of deposits per month?

This rather grim bit of SQL should work, although datestamp is the last modified date, not the submission OR creation date, it should still give a good indication.

select count(*), year(datestamp), month(datestamp)
from archive
group by year(datestamp),month(datestamp)
order by year(datestamp),month(datestamp);

I've edited the template (or other config file) but nothing seems to have changed - why?

While EPrints tries to automatically reload when a configuration file is changed you may need to reload the Web server for changes to take effect. See also the epadmin reload, the generate_static and generate_views commands.

How to I add a metadata field to a live repository?

See Adding a Field to a Live Repository.

Installation Related Questions

When running a script I get the error; "Insecure dependency in mkdir while running with -T switch"

This usually indicates you are running an eprints script as root. Don't do that; become user 'eprints' instead.

title and fulltext search returns no results, but date search does

The indexer daemon is probably not running or is not working correctly, see API:bin/indexer.

I don't want to give epadmin my mysql root password. What is the alternative?

 mysql> CREATE DATABASE myrepo;
 Query OK, 1 row affected (0.06 sec)

 mysql> GRANT ALL PRIVILEGES ON myrepo.* TO myrepo@localhost IDENTIFIED BY 'secret';
 Query OK, 0 rows affected (0.52 sec)

Then provide the database name, user name and password to the epadmin create command.

How do I get a value for a field of an eprint (without using any SQL)?

(assuming the eprint is in the main archive, and has eprintid number 23)

my $ds = $repo->dataset( "eprint" );
my $eprint = $ds->dataobj( 23 );
my $value = $eprint->value( "editors" );

See API:EPrints/DataObj.

How can I get a utf8 string of the name of a subject, given its subjectid?

sub get_subject_name_string
	my( $session, $subjectid ) = @_;
	my $subj = EPrints::Subject->new( $session, $subjectid ); 
	if( !defined $subj ) 
		return "errer, unknown subject: $subjectid";
	return EPrints::Utils::tree_to_utf8( $subj->render_description() );