Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.
Comment: code link to github instead of fisheye

DSpace 1.6 and newer versions uses the Apache SOLR application underlying the statistics. SOLR enables performant searching and adding to vast amounts of (usage) data.
Unlike previous versions, enabling statistics in DSpace does not require additional installation or customization. All the necessary software is included.

Table of Contents
minLevel2
outlinetrue
stylenone

What is exactly being logged ?

...

Definition of which fields are to be stored happens in the file dspace/solr/statistics/conf/schema.xml.
The fields, stored in a usage event by default are:

Code Block

<field name="type" type="integer" indexed="true" stored="true" required="true" />
<field name="id" type="integer" indexed="true" stored="true" required="true" />
<field name="ip" type="string" indexed="true" stored="true" required="false" />
<field name="time" type="date" indexed="true" stored="true" required="true" />
<field name="epersonid" type="integer" indexed="true" stored="true" required="false" />
<field name="continent" type="string" indexed="true" stored="true" required="false"/>
<field name="country" type="string" indexed="true" stored="true" required="false"/>
<field name="countryCode" type="string" indexed="true" stored="true" required="false"/>
<field name="city" type="string" indexed="true" stored="true" required="false"/>
<field name="longitude" type="float" indexed="true" stored="true" required="false"/>
<field name="latitude" type="float" indexed="true" stored="true" required="false"/>
<field name="owningComm" type="integer" indexed="true" stored="true" required="false" multiValued="true"/>
<field name="owningColl" type="integer" indexed="true" stored="true" required="false" multiValued="true"/>
<field name="owningItem" type="integer" indexed="true" stored="true" required="false" multiValued="true"/>
<field name="dns" type="string" indexed="true" stored="true" required="false"/>
<field name="userAgent" type="string" indexed="true" stored="true" required="false"/>
<field name="isBot" type="boolean" indexed="true" stored="true" required="false"/>
<field name="bundleName" type="string" indexed="true" stored="true" required="false" multiValued="true" />

...

Property:

server

Example Value:

server = http://127.0.0.1/solr/statistics

Informational Note:

Is used by the SolrLogger Client class to connect to the Solr server over http and perform updates and queries. In most cases, this can (and should) be set to localhost (or 127.0.0.1).

To determine the correct path, you can use a tool like wget to see where Solr is responding on your server. For example, you'd want to send a query to Solr like the following:

Code Block
wget http://127.0.0.1/solr/statistics/select?q=*:*

Assuming you get an HTTP 200 OK response, then you should set solr.log.server to the '/statistics' URL of 'http://127.0.0.1/solr/statistics' (essentially removing the "/select?q=:" query off the end of the responding URL.)

Property:

spiderips.urls

Example Value:

spiderips.urls =

Code Block

http://iplists.com/google.txt, \
http://iplists.com/inktomi.txt, \
http://iplists.com/lycos.txt, \
http://iplists.com/infoseek.txt, \
http://iplists.com/altavista.txt, \
http://iplists.com/excite.txt, \
http://iplists.com/misc.txt, \
http://iplists.com/non_engines.txt

Informational Note:

List of URLs to download spiders files into [dspace]/config/spiders. These files contain lists of known spider IPs and are utilized by the SolrLogger to flag usage events with an "isBot" field, or ignore them entirely.

The "stats-util" command can be used to force an update of spider files, regenerate "isBot" fields on indexed events, and delete spiders from the index. For usage, run:

Code Block

dspace stats-util -h

from your [dspace]/bin directory

Property:

dbfile

Example Value:

dbfile = ${dspace.dir}/config/GeoLiteCity.dat

Informational Note:

The following referes to the GeoLiteCity database file utilized by the LocationUtils to calculate the location of client requests based on IP address. During the Ant build process (both fresh_install and update) this file will be downloaded from http://www.maxmind.com/app/geolitecity if a new version has been published or it is absent from your [dspace]/config directory.

Property:

resolver.timeout

Example Value:

resolver.timeout = 200

Informational Note:

Timeout in milliseconds for DNS resolution of origin hosts/IPs. Setting this value too high may result in solr exhausting your connection pool.

Property:

useProxies

Example Value:

useProxies = true

Informational Note:

Will cause Statistics logging to look for X-Forward URI to detect clients IP that have accessed it through a Proxy service (e.g. the Apache mod_proxy).  Allows detection of client IP when accessing DSpace. [Note: This setting is found in the DSpace Logging section of dspace.cfg]

Property:

statistics.item.authorization.admin

Example Value:

statistics.item.authorization.admin = true

Informational Note:

When set to true, only general administrators, collection and community administrators are able to access the statistics from the web user interface. As a result, the links to access statistics are hidden for non logged-in admin users. Setting this property to "false" will display the links to access statistics to anyone, making them publicly available.

Property:

solr.statistics.logBots

Example Value:

solr.statistics.logBots = true

Informational Note:

When this property is set to false, and IP is detected as a spider, the event is not logged.
When this property is set to true, the event will be logged with the "isBot" field set to true.
(see solr.statistics.query.filter.* for query filter options)

Property:

solr.statistics.query.filter.spiderIp

Example Value:

solr.statistics.query.filter.spiderIp = false

Informational Note:

If true, statistics queries will filter out spider IPs -- use with caution, as this often results in extremely long query strings.

Property:

solr.statistics.query.filter.isBot

Example Value:

solr.statistics.query.filter.isBot = true

Informational Note:

If true, statistics queries will filter out events flagged with the "isBot" field. This is the recommended method of filtering spiders from statistics.

Property:

query.filter.bundles

Example
Value:

query.filter.bundles=ORIGINAL

Informational
Note:

A comma seperated list that contains the bundles for which the file statistics will be displayed.

Upgrade Process for Statistics

...

First approach the traditional DSpace build process for updating

Code Block

cd [dspace-source]/dspace
 mvn package
 cd [dspace-source]/dspace/target/dspace-<version>-build.dir
 ant -Dconfig=[dspace]/config/dspace.cfg update
 cp -R [dspace]/webapps/* [TOMCAT]/webapps

The last step is only used if you do not follow the recommended practice of configuring [dspace]/webapps as location for webapps in your servlet container (Tomcat, Resin or Jetty). If you only need to build the statistics, and don't make any changes to other web applications, you can replace the copy step above with:

Code Block

cp -R dspace/webapps/solr TOMCAT/webapps

...

The following Dspace.cfg fields are only applicable to the older statistics solution.

Code Block

###### Statistical Report Configuration Settings ######

 # should the stats be publicly available?  should be set to false if you only
 # want administrators to access the stats, or you do not intend to generate
 # any
 report.public = false

 # directory where live reports are stored
 report.dir = ${dspace.dir}/reports/

...

When a backup has been made start the Tomcat/Jetty/Resin server program.
The update script has one optional command which will if given not only update the broken file statistics but also delete file statistics for files that where removed from the system (if this option isn't active these statistics will receive the "BITSTREAM_DELETED" bundle name).

Code Block

#The -r is optional
[dspace]/bin/dspace stats-util -b -r

...

If required, the solr server can be optimized by running

Code Block

{dspace.dir}/bin/stats-util -o

...

In DSpace 1.6.x, each solr event was committed to the solr server individually. For high load DSpace installations, this would result in a huge load of small solr commits resulting in a very high load on the solr server.
This has been resolved in dspace 1.7 by only committing usage events to the solr server every 15 minutes. This will result in a delay of the storage of a usage event of maximum 15 minutes. If required, this value can be altered by changing the maxTime property in the

Code Block

{dspace.dir}/solr/statistics/conf/solrconfig.xml

...

Modify line 178 in the StatisticsTransformer.java file

https://github.com/DSpace/DSpace/blob/dspace-1_8_x/dspace-xmlui/dspace-xmlui-api/src/main/java/org/dspace/app/xmlui/aspect/statistics/StatisticsTransformer.javajava#L178

-6 is the default setting, displaying the past 6 months of statistics. When reducing this to a smaller natural number, less months are being displayed.

...

Top downloaded items by a specific user

Query:

Code Block

http://localhost:8080/solr/statistics/select?indent=on&version=2.2&start=0&rows=10&fl=*%2Cscore&qt=standard&wt=standard&explainOther=&hl.fl=&facet=true&facet.field=epersonid&q=type:0

...

facet.field=epersonid — You want to group by epersonid, which is the user id.
type:0 — Interested in bitstreams only

Code Block

<lst name="facet_counts">
    <lst name="facet_fields">
        <lst name="epersonid">
            <int name="66">1167</int>

<int name="117">251</int>

<int name="52">42</int>

<int name="19">36</int>

<int name="88">20</int>

<int name="112">18</int>

<int name="110">9</int>

<int name="96">0</int>

</lst>
    </lst>
</lst>

...

You have two options to install/update this file:

  1. Attempt to re-run the automatic installer from your DSpace Source Directory ([dspace-source]). This will attempt to automatically download the database file, unzip it and install it into the proper location:

    Code Block
    ant update_geolite
    • NOTE: If the location of the GeoLite Database file is known to have changed, you can also run this auto-installer by passing it the new URL of the GeoLite Database File: ant -Dgeolite=[full-URL-of-geolite] update_geolite
  2. OR, you can manually install the file by performing these steps yourself: