You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@flink.apache.org by "Yanfei Lei (Jira)" <ji...@apache.org> on 2023/04/28 10:44:00 UTC

[jira] [Created] (FLINK-31969) Remove all local recovery files immediately when local recovery fail

Yanfei Lei created FLINK-31969:
----------------------------------

             Summary: Remove all local recovery files immediately when local recovery fail
                 Key: FLINK-31969
                 URL: https://issues.apache.org/jira/browse/FLINK-31969
             Project: Flink
          Issue Type: Improvement
          Components: Runtime / State Backends
    Affects Versions: 1.18.0
            Reporter: Yanfei Lei


In BackendRestorerProcedure, when local recovery fail, state backend will try to recover from the remote state, at this time all local recovery files are useless, they needs to wait until the next checkpoint to be cleared, deleting local files immediately can better avoid file leakage, especially when the local space is insufficient, deleting useless local recovery files can free up space to download the remote state file.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)