You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@flume.apache.org by "Johny Rufus (JIRA)" <ji...@apache.org> on 2015/03/01 14:32:04 UTC
[jira] [Assigned] (FLUME-2518) 0 byte .flumespool-main.meta
interrupts flume reading out of source spooling directory
[ https://issues.apache.org/jira/browse/FLUME-2518?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Johny Rufus reassigned FLUME-2518:
----------------------------------
Assignee: Johny Rufus
> 0 byte .flumespool-main.meta interrupts flume reading out of source spooling directory
> --------------------------------------------------------------------------------------
>
> Key: FLUME-2518
> URL: https://issues.apache.org/jira/browse/FLUME-2518
> Project: Flume
> Issue Type: Bug
> Reporter: Asim Zafir
> Assignee: Johny Rufus
>
> due to "n" reason undefined, we have seen flume stopped committing to commit hdfs sink and the spooling count in source spooling directory keeps increasing - we have found that happens when .flumespool-main.meta size reduces to 0, this behavior will kick in - the solution is to delete .flumespool-main.meta and restart flume process (or the collector agent) this will recreate this file which is going to be hidden and greater then 0 bytes - after this we flume continues to flush into hdfs - is there a fix on this already ?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)