Hi all, I would like to propose to reopen discussion on FLIP-505 (Flink History Server Scalability Improvements : Remote Data Store Fetch and Per-Job Fetch) <https://cwiki.apache.org/confluence/spaces/FLINK/pages/337678729/FLIP+505+Flink+History+Server+Scability+Improvements+Remote+Data+Store+Fetch+and+Per+Job+Fetch> <https://cwiki.apache.org/confluence/spaces/FLINK/pages/337678729/FLIP+505+Flink+History+Server+Scability+Improvements+Remote+Data+Store+Fetch+and+Per+Job+Fetch> .
Given the recent FLIP-584 (Pluggable Storage Backend) <https://cwiki.apache.org/confluence/spaces/FLINK/pages/430408093/FLIP-584+Support+Pluggable+Storage+Backend+for+HistoryServer> and FLIP-585 (Lazy Loading for HistoryServer) <https://cwiki.apache.org/confluence/spaces/FLINK/pages/430408096/FLIP-585+Introduce+Lazy+Loading+for+HistoryServer> merges, I checked whether FLIP-505 is still needed. The problem it addresses still persists as FlinkHistoryServer local retention limit still deletes archives at their original/remote location once exceeded, coupling local cache size to remote retention. Neither FLIP-584 <https://cwiki.apache.org/confluence/spaces/FLINK/pages/430408093/FLIP-584+Support+Pluggable+Storage+Backend+for+HistoryServer> nor FLIP-585 <https://cwiki.apache.org/confluence/spaces/FLINK/pages/430408096/FLIP-585+Introduce+Lazy+Loading+for+HistoryServer> changes this behavior. Open to any thoughts FLINK-37155 <https://issues.apache.org/jira/browse/FLINK-37155> (PR <https://github.com/apache/flink/pull/26878>) before moving forward. -- Yours Sincerely Archit Goyal Email: [email protected]
