See on GitHub

    Migration: Server Breaking Changes

    License keys

    License keys for versions lower than 6.0 are not supported by RavenDB 6.0.
    If you own a valid license key for RavenDB 5.x or lower, please upgrade it using the quick online interface described here.

    RavenDB for Docker

    RavenDB now applies an improved security model, and uses a dedicated user rather than root.
    Read more about this change here.

    Unsupported sharding features

    RavenDB 6.0 introduces sharding. Server and client features that are currently unavailable in a sharded database (but remain available in regular databases) are listed here.

    Graph queries

    Graph Queries support, available in RavenDB versions 4.2 to 5.x, has been removed from the RavenDB 6.x server and client API.

    SQL ETL

    SQL ETL no longer tolerates errors on Load, load errors are thrown immediately.
    This is done to distinguish partial load errors that are used in SQL ETL from, for example, commit errors that may happen during load. (Prior to this change, the ETL would just advance instead of retrying.)

    DateOnly & TimeOnly

    DateOnly and TimeOnly types are now supported for every new auto index.

    Full-text search with wildcards

    Starting with 6.0 we have changed how the Search method handles wildcards when they are included in search terms:

    Behavior for versions lower than 6.0:

    After the analyzer stripped wildcards from the search term, RavenDB would attempt to restore the * to their original positions before sending the term to the search engine (Lucene or Corax).

    Behavior for 6.0 and up:

    Once wildcards are stripped by the analyzer, we no longer add them back before sending the term to the search engine. The search terms sent to the search engine are solely based on the transformations applied by the analyzer used in the index.

    Note the different behavior in the following cases:

    • When using RavenStandardAnalyzer or StandardAnalyzer or NGramAnalyzer:
      The queried terms in the Search method are processed with the LowerCaseKeywordAnalyzer before being sent to the search engine.
    • When using a custom analyzer:
      The queried terms in the Search method are processed according to the custom analyzer's logic.
    • When using the Exact analyzer:
      The queried terms in the Search method remain untouched as produced by the exact analyzer.

    When using Corax as the search engine,
    this behavior will only apply to indexes that are newly created or have been reset.

    See detailed examples in: Searching with wildcards.

    Home / Docs / Migration / Server / Server Breaking Changes

    Do you need any help with Migration?

    Yes, contact support