mingmxu opened a new issue, #10952: URL: https://github.com/apache/pinot/issues/10952
# what's the problem? When one server instance is lost abruptly for reasons, query failures are observed for short period in two scenarios: * The running queries would fail after timeout; * Some queries are sent to the server before the issue is detected; # What's the proposal? Applied to both v1 and v2 engine, by introducing broadcast mode in `AdaptiveServerSelector`, one push-down subquery is sent to all healthy servers which hold a replica of the segment, the first response or the most refreshed response(or other policies) is chosen for the next step. In this way, query failures could be mitigated if lost server number(M) is less than segment replica number(N). # Risk By broadcasting subqueries, server instance has x times more workload; -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org.apache.org For queries about this service, please contact Infrastructure at: us...@infra.apache.org --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org For additional commands, e-mail: commits-h...@pinot.apache.org
