Installing Elasticsearch
Elasticsearch is integrated in yuuvis® RAD for searching.
It includes a linguistic plugin for search optimization in exactly one of the following languages: English, French, German, Spanish, or Italian. The language is specified during installation of yuuvis® RAD service-manager. It can be changed.
Ensure that the System Requirements are fulfilled before installation.
To install Elasticsearch, follow these steps:
-
Copy the \elasticsearch\ installation directory to the file system of the computer on which you want to install Elasticsearch.
- Run elasticsearch_setup.exe from the installation directory.
-
Specify the following parameters:
-
Europe
Elasticsearch is configured for European languages.
-
Asia
Elasticsearch is configured for Asian languages.
This option is only available for Kyocera Enterprise Information Manager.
-
Change the service parameters before you start the service:
Launch elasticsearch-service-x64w.exe from the \bin\ directory.
Service parameters:
Log On > Account Account for login to the service
Elasticsearch requires an administrative account.
Logging > Level Optional: Logging level for service logging Logging > Path Optional: Path for service logs Logging > Stdout Optional: Redirect the standard output to a file Logging > Stderror Optional: Redirect the standard error output to a file -
Other optional changes:
- Modify index directory
- Windows Service Control
- Logging
- Search: AND/OR
- Search: Fuzzy search
- No. of shards
- Call the URL: http://<service-admin-IP>:7291/swagger-ui.html#!/api-controller/reindex
- Click Try it out! click
- CSV for exporting hit lists
-
Autocomplete
- Call the URL: http://<service-admin-IP>:7291/swagger-ui.html#!/api-controller/completeAutocompleteIndexUsingPOST
- Optional: Enter the time period
- Click Try it out! click
- Index cleanup
- Call the URL: http://<service-admin-IP>:7291/swagger-ui.html#!/api-controller/cleanUpUsingPOST
- Optional: Enter the object type
- Click Try it out! click
-
Tables
Tables with more than 20,000 rows lead to errors when reindexing. Reindexing may also be necessary for Elasticsearch updates.
This limit is set via the osfts.tableRowSize parameter of the index-prod.yml file located in the \config\ directory of yuuvis® RAD service-manager. The value must not be increased.
- Troubleshooting
- Linguistic plugin
- Quality criteria for the hit lists
- Structure tree/structure service
The index directory can be modified in the elasticsearch.yml configuration file located in the \config\ Elasticsearch directory. Parameter: path.data
Configure recovery in the event of service errors via the Windows Service Control feature: Select Restart service for all errors.
The Elasticsearch log directory can be modified in the elasticsearch.yml configuration file located in the \config\ Elasticsearch directory. Parameter: path.logs
For the full text search, several terms are linked by the Boolean AND if the user does not specify an operator. The default operator can be modified in the search-prod.yml configuration file located in the \config directory of yuuvis® RAD service-manager. Parameter: osfts.default-operator: OR
The fuzzy search can be enabled in the search-prod.yml configuration file located in the \config\ directory of yuuvis® RAD service-manager.
Parameter: osfts.fuzzy-search: true
In a fuzzy search, combinations of terms are only found if the combination is present in the metadata or content.
Default: one shard per index.
The number of shards can be specified in the index-prod.yml configuration file located in the \config directory of yuuvis® RAD service-manager. Parameter: osfts.number-of-shards
Reindexing via yuuvis® RAD services-admin is required when changes are made:
Reindexing starts immediately.
The separator for columns in exported CSV files can be entered in the search-prod.yml configuration file located in the \config directory of yuuvis® RAD service-manager. Parameter: csv.delimiter: ';'. Default: semicolon.
CSV files are always output with UTF-8 encoding. Without BOM is the default. The csv.writeBOM: true parameter is used to output the file with BOM. For a correct display of umlauts and special characters, the output with a BOM should be set in Excel, for example.
Organization reference fields and object reference fields can be resolved in order to export hit lists. In this case, names are output instead of IDs.
This requires the following entry in the extend.json file:
"search": {
"resolveReferencesCSV": true,
"resolveReferencesCSVThreshold": 5
},The resolutions can cause the export to take significantly more time. The resolveReferencesCSVThreshold parameter sets the number of resolutions at which point a message will be shown to the user. The user can then cancel the export. Default: 10.000
The maximum number of suggestions for fields with the relevant property can be specified in the index-prod.yml configuration file located in the \config directory of yuuvis® RAD service-manager. Parameter: osfts.autocomplete.suggestions-size: 500. Default: 500.
Auto-completion can be disabled by entering the following entry in the application-es.yml configuration file located in the \config\ directory of yuuvis® RAD service-manager:
osfts.autocomplete.enabled: false
The autocomplete index can be created again using yuuvis® RAD services-admin:
Indexing starts promptly.
The full-text index can be cleaned up via yuuvis® RAD services-admin after deleting large numbers of documents.
Cleanup starts promptly.
The cleanup is done in batches. The batch size can be configured via the index-prod.yml file located in the \config\ directory of yuuvis® RAD service-manager:
osfts.cleanup.bulkSize: 1000
If errors occur during indexing, three new attempts, or retries, are initiated at intervals of one hour.
This setting can be changed by modifying the configuration in the index-prod.yml file located in the \config\ directory of yuuvis® RAD service-manager:
routing:
errorHandling:
retry: 3
periodInMinutes: 60
startPolicy: "0 0/1 * ? * * *"Parameters:
retry Number of attempts periodInMinutes Interval between attempts startPolicy Schedule for attempts in CRON notation A language for the Linguistic plugin is specified during the installation of yuuvis® RAD service-manager. The index-prod.yml configuration file located in the \config directory of yuuvis® RAD service-manager can be used to activate another language:
osfts:
languages: 'de'Enter the abbreviation of the desired language: English (en), French (fr), German (de), Spanish (es), or Italian (it). Licenses can be acquired from OPTIMAL SYSTEMS and integrated here to unlock multiple and additional languages.
As a default, along with the standard Elasticsearch criteria, yuuvis® RAD uses additional quality criteria for sorting hit lists. For example, documents that users have edited themselves and the number of versions are included and have a higher weighting.
You can limit this scoring to Elasticsearch’s default criteria via the search-prod.yml configuration file:
osfts:
modify-score: falseOr you can adjust the scoring via parameters instead:
osfts:
boost_mode: sum
boost: 1.5boost_mode: Default: avg.
boost: Default: 1.
For information about the values, refer to the documentation of Elasticsearch.
Complex configurations of the structure tree via the 'structureservice' service can lead to errors.
If necessary, add the following entry to the elasticsearch.yml configuration file located in the \config\ Elasticsearch directory:
indices.query.bool.max_nested_depth: 100
- Start the Elasticsearch service.
- Configure Elasticsearch authentication:
- Run the elasticsearch-set-initial-passwords.bat batch file located in the \bin\ Elasticsearch directory with administrative rights.
- Open built-in.usr located in the \config\ Elasticsearch directory.
- Copy the password of the user named 'elastic'.
- Provide the password when installing yuuvis® RAD service-manager.
The installation wizard will start. Follow the steps described by the installation wizard.
HTTP port | Elasticsearch port. Default: 9200 |
Index directory |
Index directory for Elasticsearch. The index directory should be created on a fast and large data medium, if possible, and not in the installation directory. |
Maximum RAM | Maximum working memory for the service (in MB) |
Language pack |
|
Reindexing
Full-text operations are available for reindexing. By default, the full-text operation 'Technical update' is set up to repeat every five minutes.
Clusters
When installing Elasticsearch clusters, please note the following:
- Do not start any further installations until the configuration files have been modified.
- Before launching the service, enter all Elasticsearch addresses in the elasticsearch.yml configuration files of each Elasticsearch installation:
- discovery.seed_hosts: ["host1", "host2"]
- Before launching the service, enter unique names for nodes in the elasticsearch.yml configuration files of each Elasticsearch installation: node.name: nodename1
- Before launching the service, enter all the nodes in the elasticsearch.yml configuration files of each Elasticsearch installation:
- cluster.initial_master_nodes: ["node-1", "node-2"]
- Generate the password using the elasticsearch-set-initial-passwords.bat batch file while the service is running only for the first Elasticsearch installation.
- Start the first Elasticsearch installation and then each of the other installations.
For more information on configuring clusters, refer to the corresponding documentation of Elasticsearch.
Uninstalling
Use the uninstall program to uninstall Elasticsearch from the installation directory.
Updates
For information on updating components, see Release Information.