Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Plain View
Flume, mail # user - About file channel


+
Yanzhi.liu 2012-09-03, 08:46
Copy link to this message
-
Re: About file channel
Jeff Lord 2012-09-05, 16:30
Yanzhi,

It looks like the size of the data folder used by file channel does not
reduce even after all the data has been sent to the target.

The file channel is implemented as a write-ahead log and will continue to
grow the log files. Eventually (depending upon the configured maximum size
of the log files, or due to restart of the channel) the log files will get
rolled, but the old files may still be kept around until all events in them
are dereferenced completely. Only then will the old files be deleted and
space freed up on the system.

-Jeff

On Mon, Sep 3, 2012 at 1:46 AM, Yanzhi.liu <[EMAIL PROTECTED]> wrote:

> hello everybody:
>     I use the file channel to transfer files, which helps document
> security.
> However, the file channel data directory contains log is very easy to
> accumulate very much. And although I know that the the channel generated
> log file will be deleted, but I found that only the first log-1 of the
> deleted file is the fastest, and the rest of the files I waited more than
> 10 minutes, did not remove a. Therefore, I would like to know the file you
> want to delete a file channel data directory, log how long.
> Thanks very much!
> My Name
> Yanzhi Liu
>
>
+
Yanzhi.liu 2012-09-06, 01:41