[jira] [Updated] (CASSANDRA-13962) should java.io.OutputStream.flush() be called on Commit log operations ?

["C. Scott Andreas (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CASSANDRA-13962?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

C. Scott Andreas updated CASSANDRA-13962:
-----------------------------------------
    Component/s: Local Write-Read Paths

> should java.io.OutputStream.flush() be called on Commit log operations ?
> ------------------------------------------------------------------------
>
>                 Key: CASSANDRA-13962
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-13962
>             Project: Cassandra
>          Issue Type: Bug
>          Components: Local Write-Read Paths
>         Environment: cassandra 2.2.8
>            Reporter: Ilya Shipitsin
>            Priority: Major
>
> we run high loaded cassandra cluster on 2.2.8
> when we reboot node, very often we observe broken commit log, like described here
> https://stackoverflow.com/questions/33304367/cassandra-exiting-due-to-error-while-processing-commit-log-during-initializatio
> I guess, that output is not flushed upon commit log write, I mean  java.io.OutputStream.flush()
> google also say, the same issue occur to sstables (they are written in more rare fashion, so we did not observe that practically)
> any idea why output is not flushed ? is it done in purpose ?



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org