[ https://issues.apache.org/jira/browse/FLINK-28012?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17553249#comment-17553249 ]
Aitozi commented on FLINK-28012: -------------------------------- I think the gcs should work with the filesystem based fetcher, have you tried by adding the flink-gs-fs-hadoop dependency ? > Add built-in support for fetching jar from GCS (Google Cloud Storage) > --------------------------------------------------------------------- > > Key: FLINK-28012 > URL: https://issues.apache.org/jira/browse/FLINK-28012 > Project: Flink > Issue Type: Improvement > Components: Kubernetes Operator > Reporter: Xin Hao > Priority: Minor > > I think it's easy and meaningful to add built-in support in session job mode > for downloading jar from GCS now. > The logic should look like the code below > {code:java} > if ("http".equals(uri.getScheme()) || "https".equals(uri.getScheme())) { > return HttpArtifactFetcher.INSTANCE.fetch(jarURI, flinkConfiguration, > targetDir); > } else if ("gs".equals(uri.getScheme())) { > return GcsArtifactFetcher.INSTANCE.fetch(jarURI, flinkConfiguration, > targetDir); > } else { > return FileSystemBasedArtifactFetcher.INSTANCE.fetch( > jarURI, flinkConfiguration, targetDir); > } {code} > We only need to extend the ArtifactManager and implement the new > GcsArtifactFetcher. > Also, the users can set up the GCS authorization credentials by ENV and > ConfigMap so that no need for any change inner the operator. -- This message was sent by Atlassian Jira (v8.20.7#820007)