Removing an EAD

From UA Libraries Digital Services Planning and Documentation
Revision as of 11:12, 1 September 2016 by Jlderidder (talk | contribs) (FIRST)

The primary considerations in the workflow are whether the collection has already been released in LOCKSS for preservation copying, and whether there is any digitized content for the collection. Please be careful with the commands below, as they are irreversible, and if you type in the wrong collection number (or worse), we will not be able to restore the deletion.

The examples below are dependent upon the collection number being "u0003_0003696" or "u0003_0003714", and the gaining collection (which it's been combined with) being "u0003_0003748". Change the numbers to what is CORRECT for the entry you are working with.

The process to do this is as follows:


*Check to see if the content is in LOCKSS, and if it has digital content:

  1. On libcontent, Log into the database InfoTrack: select * from inLOCKSS where id_2009 like "u0003_0003696"; (if null, it is not in LOCKSS yet.)
  2. To check for digital content, there are two parts:
    1. in InfoTrack: select * from allColls where id_2009 like "u0003_0003696" and AnalogOrDigital like "D"; (if null, there still might be some in progress)
    2. on the Share drive, look through the Completed folder, the in Progress folder and the Waiting folder for collections with this collection number.


  1. ASK DONNELLY IF THE ITEMS NEED TO GO TO THE/AN OTHER COLLECTION. If so, this is a completely separate process. Do NOT delete those directories or content yet until this is sorted out, either online or in the archive.


  1. Create a problems.txt in /srv/archive/u0003/0003696/Documentation/ and indicate that "Donnelly Lancaster told us on x date to take the EAD down, that the collection was combined with collection u0003_0003748" (or whatever reason was given; substitute the name as is appropriate). The process for this step is as follows:
    1. cd /srv/archive/u0003/0003696/Documentation (change these numbers to get to the correct collection)
    2. vi u0003_0003696.problems.txt (again, change the numbers)
    3. type in "i" for insert
    4. type the message
    5. type in escape to stop writing
    6. type in ":wq" to exit and save what you wrote.


*Clean up the databases and document

  1. Modify this entry appropriately, then use it in the InfoTrack database:
 update allColls set online=NULL, inAcumen=NULL, listLevel=NULL, notes = "MSS 3696 Guide to the Arley E. Hughes Sr. Letters was added to    collection MSS 3748 on 7/21/14 per Donnelly Lancaster, then taken offline" where id_2009 like "u0003_0003748" and AnalogOrDigital like "A";
  1. `rm /srv/archive/u0003/0003696/Metadata/*ead*xml’ (this gets versioned ones as well)
  2. `rm /srv/www/htdocs/content/u0003/0003696/Metadata/*ead.xml’
  3. Modify the manifest to remove EAD links: vi /srv/archive/u0003/0003696/Documentation/Manifest.html

*If NO items are in the collection:

  1. Make a note of the title for the next set of steps, and then:
  2. `rm –r /srv/www/htdocs/content/u0003/0003696`
  3. Retain the collection directory and Documentation directory for the problems.txt file.
  4. And remove from database: delete from allColls where id_2009 like “u0003_0003696”
  5. And remove the flat file for checksums: rm /srv/scripts/md5/cya/checksums/u0003/0003696
  6. In either case, remove the EAD listing from the database: delete from allColls where id_2009 like “u0003_0003748” and AnalogOrDigital like “A”;

*If in LOCKSS:

  1. Remove the live version: `rm /srv/www/htdocs/content/u0003/0003696/Metadata/*ead.xml’

6. Whether in LOCKSS or not, remove alternative versions of the EAD: `rm /srv/deposits/EADs/*/u0003_0003696.ead.xml`

NOTE: If you delete the manifest (which happens if you delete the archival directories) you MUST remove the link to the manifest(s) that are in the holder level directory manifest as well. For example, if you remove /srv/archive/u0003/0003696/ then you need to edit the /srv/archive/u0003/Documentation/Manifest.html and remove the link(s) to the manifests for the u0003_0003696 collection.


  1. Since this impacted another collection, check the title both in the database and in the Manifest in the archive (and the collection xml).
    1. view /srv/archive/u0003/0003748/Documentation/u0003_0003748.xml (there is none, since there’s no items online yet).
    2. vi /srv/archive/u0003/0003748/Documentation/Manifest.html (changed the title to “Hughes family papers” on both <head> and <title> lines.
  2. In the InfoTrack database:
    1. select * from allColls where id_2009 like "u0003_0003748"; (the title was in this case updated by the upload of the new EAD; if it was not, must modify it. )
    2. Also in this database, add a note:
 update allColls set notes = “MSS 3696 Guide to the Arley E. Hughes Sr. Letters was added into this collection 7/21/14 per Donnelly Lancaster; that collection was removed.” where id_2009 like “u0003_0003748”;


  1. If the collection for which an EAD was retracted contains items, check to see if the items go to another collection. If they do, that’s a whole ‘nother process, that impacts checksums too…
  2. Delete the PURL for the EAD from the database so that when we have to update those, we won’t be confused.
 `delete from lookup where id_2009 like “u0003_0003696” and history like “%EAD%”;

In InfoTrack we also have a table for deleted collections (this has clearly happened a few times).

3. If the entire collection was deleted (as this was), then we need an entry here that indicates the identifier, whether it was analog (the EAD) or digital, the date pulled, who requested this, and any notes.

So, for this one, I logged into InfoTrack and typed in:

insert into removedColls values(NULL,"u0003_0003696","A","2014-07-21","Donnelly Lancaster Walton", "combined with collection u0003_0003748");

To see the results, type in “select * from removedColls” and you’ll see the whole list.


  1. Change directory to /srv/deposits/EADs/
  2. do an ls for the finding aid for any directory in that area, for example: ls */u0003_0003714.ead.xml
  3. remove each of those files, for example: rm */u0003_0003714.ead.xml