[jira] [Commented] (FLINK-3103) Accessing closed in FsStateBackend#FsCheckpointStateOutputStream#flush() should be synchronized

["Stephan Ewen (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/FLINK-3103?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15056541#comment-15056541 ] 

Stephan Ewen commented on FLINK-3103:
-------------------------------------

The class is not really designed for concurrent use.

But I see your point, it is actually inconsistent with the fact that the {{close()}} methods use the lock.

I would actually suggest to not use a lock in the {{close()}} methods either. In a single threaded use, there is really no need for these locks.

> Accessing closed in FsStateBackend#FsCheckpointStateOutputStream#flush() should be synchronized
> -----------------------------------------------------------------------------------------------
>
>                 Key: FLINK-3103
>                 URL: https://issues.apache.org/jira/browse/FLINK-3103
>             Project: Flink
>          Issue Type: Bug
>            Reporter: Ted Yu
>
> Here is related code:
> {code}
>     public void flush() throws IOException {
>       if (!closed) {
> {code}
> Lock "FsCheckpointStateOutputStream.this" should be held.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)