Hello!

 

So, I’m getting to the stage where I’m needing to implement this and I’ve taken a look at a record to see how I’m going to need to rewrite the export XML to accommodate moving the documents.

 

Take this example:

 

<document id='http://eprints.lincoln.ac.uk/id/document/57053'>

        <docid>57053</docid>

        <rev_number>2</rev_number>

        <files>

          <file id='http://eprints.lincoln.ac.uk/id/file/321720'>

            <fileid>321720</fileid>

            <datasetid>document</datasetid>

            <objectid>57053</objectid>

            <filename>27039 Simon Burton_ Nowhere Men - a-n The Artists Information Company.pdf</filename>

            <mime_type>application/pdf</mime_type>

            <hash>939078bc226712a5e8d640ced5df31b3</hash>

            <hash_type>MD5</hash_type>

           <filesize>270630</filesize>

            <mtime>2017-05-10 09:42:34</mtime>

            <url>http://eprints.lincoln.ac.uk/27039/1/27039%20Simon%20Burton_%20Nowhere%20Men%20-%20a-n%20The%20Artists%20Information%20Company.pdf</url>

          </file>

        </files>

        <eprintid>27039</eprintid>

        <pos>1</pos>

        <placement>1</placement>

        <mime_type>application/pdf</mime_type>

        <format>application/pdf</format>

        <language>en</language>

        <security>staffonly</security>

        <main>27039 Simon Burton_ Nowhere Men - a-n The Artists Information Company.pdf</main>

        <content>whole_document</content>

        <stage>published</stage>

      </document>

 

Going by the EPrints logic, I’ve located the file in <eprints_root>/archives/<myarchive>/documents/disc0/00/02/70/39/01/<document>

 

That’s cool, I can find that. So…

 

Do I pull the whole documents folder over to my new server and put it in the new <eprints_root>/archives/<myarchive>/documents folder? Or do I put it elsewhere?

 

Which part of the xml needs rewriting to tell the import where to look for the file? When I’m setting the new site up I’m going to be using eprints.lincoln.ac.uk and faking the host for development purposes so that we can retain domain integrity on the move.

 

 

 

 

No, the subdirectories will be the same – you’ll just need to replace the bit that leads up to the documents directory.

 

 

Ah, that makes sense. So will I need to factor that in when I’m doing an XML rewrite?

 

 

it’s the eprint id, padded with zeros and then broken up into pairs of digits to make directory names.  That way there’s only ever 100 directories in each directory.

 

--

Adam

 

 

You know me all too well ;)

 

So that should work? I was only wondering because I know that under the Documents folder the structure seems rather ambiguous to me, 00 with 00, 01 and 02 and further numbers underneath that. Is there a method to that?

 

 

Hi Andrew

 

                If it were anyone but you, I’d recommend doing a mysqldump and keeping all paths the same as the simplest way to migrate a repository.  However, I’m sure you’ll counter that with “…but my repository is non-standard and I’m trying to make it standard”

 

                When I’ve had to monkey around with paths in XML files, I’ve usually done it with vim and find/replace commands.  Move the documents directory to the new server into a temporary directory, then compare the path in the XML file to the path on disk.  This will help you understand what string replacement you need to do.  The command will be something like:

 

:%s/\/usr\/share\/eprint3\/archives\/foo\/documents/\/home\/anbeeken\/migration\/documents/g

 

                ...it may take some time to run.

 

--

Adam

 

 

 

Hi all,

 

I’m looking into the best options for migrating EPrints to a new server and investigating the possibility of pulling our 43.6Gb worth of documents across so as not to embed them in XML and create large files. I know that I can bring the Documents folder over, however I’m not sure how to interpret the folder structure to rewrite the URL’s in the XML export. Any thoughts?

 

Andrew

*** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech *** Archive: http://www.eprints.org/tech.php/ *** EPrints community wiki: http://wiki.eprints.org/ *** EPrints developers Forum: http://forum.eprints.org/

*** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech *** Archive: http://www.eprints.org/tech.php/ *** EPrints community wiki: http://wiki.eprints.org/ *** EPrints developers Forum: http://forum.eprints.org/

*** Options: http://mailman.ecs.soton.ac.uk/mailman/listinfo/eprints-tech *** Archive: http://www.eprints.org/tech.php/ *** EPrints community wiki: http://wiki.eprints.org/ *** EPrints developers Forum: http://forum.eprints.org/