Skip to content

Conversation

Peter-Searby
Copy link

This doesn't fix all of the issues faced when using cluster_run (some of which may or may not just be caused by flakiness in the tests), but does cover the following:

  • Some tests depend on a fixed mgmt port, which can't be met by a cluster_run. To fix this http bootstrapping support is added, so that a http:// connection string can be used to make the tests compatible with a cluster_run instance
  • Any tests depending on fixed ports for other services are switched to using ns_server's /_p/<service>/ proxy API, so that the mgmt port can be used for all such requests (note, this is not needed for any tests depending on any auto-discovery of ports)
  • The get_terse_cluster_config function was incorrectly using a streaming API, causing it to hang when used for bootstrapping, this has been switched to the correct API

@Peter-Searby Peter-Searby force-pushed the RSCBC-147 branch 2 times, most recently from 2d09eaf to 612ec3b Compare September 4, 2025 10:37
@Peter-Searby Peter-Searby force-pushed the RSCBC-147 branch 2 times, most recently from 9afdb85 to 6ca92c0 Compare September 4, 2025 11:22
Fix test_json_block_read in tests/http.rs to use the http port, not
just assume 8091.
The streaming API is being used in a non-streaming context, so it just
hangs.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

1 participant