We have modified the SOLR logging code to post additions/updates etc to both
SOLR and ElasticSearch so this is possible or you could configure Dspace to
send all SOLR logging from each Dspace instance to an external SOLR server. The
problem with this approach is that you would have to ensure that the handles
used by each Dspace instance are unique as this is used as the unique key in
SOLR or by default you would just update a SOLR document if the unique key
already existed.
Gary, are you intending to build some interface to search over the aggregate
collection of Dspace search indices? If so one approach that we have used when
doing this sort of work is to use SOLJ to write applications that grab data
from one or more SOLR instances and index them in another - you then have full
control over what the unique key is and the ability to record what Dspace
instance the search data came from (if you need to link back to the resource in
the Dspace instance). The main issue in this approach is dealing with deletions
and having some mechanism to control what data is grabbed from each Dspace SOLR
index (if you are using Discovery there is a last indexed field that could be
used in some way).
If your requirements are not strict on the accuracy (in terms of how often the
aggregate index is updated) then harvesting into one Dspace instance would be
the simplest option.
Regards,
Ben
------------------------------------------------------------------
Dr Ben Ryan
Jorum Technical Manager
5.12 Roscoe Building
The University of Manchester
Oxford Road
Manchester
M13 9PL
Tel: 0160 275 6039
E-mail: [email protected]
------------------------------------------------------------------
-----Original Message-----
From: helix84 [mailto:[email protected]]
Sent: 29 July 2013 09:59
To: Gary Browne
Cc: dspace-tech Tech
Subject: Re: [Dspace-tech] Federating DSpace instances
One more real-time option. I'm not sure if this would be easy, because my
Solr-fu is not that strong.
You could somehow channel the content of the Solr indices of all the DSpace
instances into a single Solr index and do searches on top of that. The open
question here is how this would handle updates. I didn't do any research on
that, but you could start here [1].
You might notice that this is Similar to the VuFind solution, because VuFind is
just a thin UI on top of Solr. The difference is in the index schema. DSpace
Solr (search core) and VuFind Solr (biblio core) schemas differ, that's why
with VuFind you need import. You could skip that step here and maybe even use a
DSpace instance as the UI (with modifications).
[1] http://wiki.apache.org/solr/MergingSolrIndexes
Regards,
~~helix84
Compulsory reading: DSpace Mailing List Etiquette
https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics Get end-to-end
visibility with application monitoring from AppDynamics Isolate bottlenecks and
diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
DSpace-tech mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-tech
List Etiquette: https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette
------------------------------------------------------------------------------
See everything from the browser to the database with AppDynamics
Get end-to-end visibility with application monitoring from AppDynamics
Isolate bottlenecks and diagnose root cause in seconds.
Start your free trial of AppDynamics Pro today!
http://pubads.g.doubleclick.net/gampad/clk?id=48808831&iu=/4140/ostg.clktrk
_______________________________________________
DSpace-tech mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/dspace-tech
List Etiquette: https://wiki.duraspace.org/display/DSPACE/Mailing+List+Etiquette
