2. FAQ

How can I find out how old my Goobi viewer is?

In the source code of the page there are various meta tags at the very beginning, which provide information about the age and code base of the version used:

<meta name="generator" content="Goobi viewer" />
<meta name="version" content="21.01" />
<meta name="build date" content="2021-01-10 05:44" />
<meta name="git-revision" content="dd0bea0" />

"Goobi viewer" is the product name, "21.01" the version number. "2021-01-10 05:44" is the date when the application was compiled. "dd0bea0" are the first seven characters of the git commit hash used to compile the application.

The date is a good first indicator. The git-commit hash can be searched for on GitHub, for example, to find out exactly what state the application is at.

How can I re-index my entire database?

For the re-indexing of the entire database, the contents of the indexed_mets folder must be copied into the hotfolder:

cp /opt/digiverso/viewer/indexed_mets/*.xml /opt/digiverso/viewer/hotfolder/

Before copying, make sure that the hotfolder is empty so that no potentially corrected records are overwritten with an older version from the indexed_mets folder.

Which web browsers are supported by the Goobi viewer?

The Goobi viewer supports all so-called "evergreen browsers". These are the current versions of Google Chrome, Firefox, Safari and Microsoft Edge.

Older versions, especially Internet Explorer 11, have not been officially supported since mid-July 2018.

Which databases are supported by the Goobi viewer?

The Goobi viewer supports the MySQL or MariaDB databases for productive use and H2 for the Goobi to go environment.

Oracle Database, Microsoft SQL Server and PostgreSQL databases are currently not possible.

Which file formats for 3D objects are supported by the Goobi viewer?

The Goobi viewer has full support for OBJ (Wavefront Object), GLTF/GLB, X3D and partial support for PLY and STL.

Partial support means that only the mesh can be displayed but no surface properties. The models are then displayed in one colour.

What server specifications and operating system is recommended for an installation?

We recommend installing the Goobi viewer in a virtual machine in order to be able to add or remove resources as required. The following parameters have proven to be a good basis:

  • 4 CPUs

  • 8GB RAM

  • 25GB hard disk for the operating system

Storage is also needed for the digitised files. If possible, this should be connected directly to the machine and not as a network drive. An LVM is then created within the machine in order to be able to increase or decrease the memory if necessary. If a network drive is used, we recommend connecting it via NFS. In this case, however, an additional local device should be available for the Solr search index and the image cache. Otherwise performance problems may occur.

We also recommend installing the Goobi viewer in the current Ubuntu Linux Server LTS distribution.

The file system with the Goobi viewer data is full, can I delete any of it?

If the file system on which the data of the Goobi viewer is stored is full, usually /opt/digiverso/, then the following folders can be emptied without problems:

  • updated_mets/

  • deleted_mets/

  • error_mets/

  • success/

For the meaning of the directories see Chapter 3.2.

The following folders can also be checked:

  • abbyy/ The files in this folder have been automatically converted to ALTO by the Goobi viewer Indexer since October 2017. The Goobi viewer itself only uses the ALTO. If the data was imported via Goobi workflow, it is available there and can be deleted in the Goobi viewer.

  • pdf/ The PDF documents stored in this folder are delivered with a PDF download from the Goobi viewer. If there are no files here, (if configured) a PDF with optional text is generated from the image and optional ALTO file and delivered. The result is the same, only the delivery time may take longer. For the PDF files generated in this way, there is a cache under cache/pdf.data, so that PDF documents that have been generated once are stored there up to a cache size of 100GB.

How can I compile the Goobi viewer?

The theme is compiled and the core is included as a dependency. Therefore, the following steps describe how to check out the Reference theme and compile a viewer.war:

  • If not already done, maven must be installed:

sudo apt install maven
  • Then check out the theme:

mkdir -p /opt/digiverso/viewer/themes
cd /opt/digiverso/viewer/themes
git clone https://github.com/intranda/goobi-viewer-theme-reference.git
  • Now compile the application:

cd /opt/digiverso/viewer/themes(goobi-viewer-theme-reference/goobi-viewer-theme-reference/
mvn package

At the end, the compiled Goobi viewer is located in the target/viewer.war folder.

How's working with Git and Branches?

There is a Master branch that is always stable and installable.

There is also a so-called Develop branch, which contains new functions but cannot be completely stable yet.

The Develop branch is used for feature_ branches in which new functionality is developed. As soon as these are finished, they are merged into the Develop branch and from time to time the developments flow back from the Develop branch into the stable Master branch.

How does versioning work in the Goobi viewer?

The Goobi viewer and its components have a version number consisting of potentially three digits, for example 21.01.3.

The first digit of the version number indicates the year in which the version was released. In 2021, for example, this is 21.

The second digit indicates the month in which the version was released. Months are always indicated with two digits and, in case of doubt, filled with a leading zero. In January, for example, this is 01.

The third digit is optional and describes the patch level. It is only given if there is also a bugfix release. In the example above it is 3.

In addition to the stable versions, there are also development versions. These always have the version number of the following month without patch level and end with -SNAPSHOT. Based on the example above, the next development version would be 21.02-SNAPSHOT.

How can I delete orphaned anchor documents without volumes?

Should it ever happen that the Solr index contains empty anchor documents (i.e. multi-volume works without volumes) due to faulty imports, these can be removed from the index in one fell swoop. The following command can be used for this purpose:

cd /opt/digiverso/indexer
java -jar solrIndexer.jar solr_indexerconfig.xml -cleanupGrievingAnchors

How can I empty the entire Solr search index?

Sometimes during development it may be necessary to clear the entire Solr search index. To do this, all works must be deleted from the Goobi viewer without leaving any residues. Finally, the Solr index has to be optimised. On the Linux command line this task can be solved as follows:

cd /opt/digiverso/viewer/indexed_mets/
for i in *.xml; do echo touch ../hotfolder/${i/.xml/.purge}; done
curl "http://localhost:8983/solr/collection1/update?optimize=true&waitFlush=false"

How can I change the Access Denied image?

If the image display in the Goobi viewer is prevented by an access licence, a substitute image is displayed.

The path to this image is configured in the messages file using the following key:

noImage_accessDenied

The default entry is as follows:

noImage_accessDenied=<img src\=“/viewer/resources/images/access_denied.png” />

The key can be overwritten in the local messages files. This allows individual images or texts to be displayed that are different for each language.