Object Service Indexing

One of the primary jobs for the 3LC Object Service is indexing 3LC objects to be accessed through the Dashboard. Indexing is done in the background until the Object Service is terminated.

Locations that should be indexed are specified in the configuration settings. In particular, the following settings are relevant:

• indexing.project-root-url: Primary location writing new projects.

• indexing.project-scan-urls: Extra locations to scan for projects.

Tables and runs may be written to a particular location by a training notebook or script with 3LC integration, or by the Object Service itself when saving a revision made in the Dashboard. Those locations are usually specified using the *-root-url values.

Additional locations to index are specified using the *-scan-urls values. Note that each *-root-url is effectively added to the corresponding *-scan-urls value, so there is no need to specify the same location in both places.

The URLs that are part of these configuration options may be local file paths or cloud storage locations. See the URL Details page for more details.

The indexing system employs several strategies to ensure seamless operation. Firstly, it automatically creates scan-URL directories if they don’t already exist. Additionally, it bypasses any scan-URLs that are inaccessible. All indexing issues are recorded by the 3LC logging and can be inspected at log location.

Note

It is necessary to restart the object-service, which also restarts the indexer, whenever there are modifications to the configuration files or relevant changes to the network environment.