hadoop-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Wojciech Langiewicz <wlangiew...@gmail.com>
Subject Re: cleaning up 'hadoop.tmp.dir' ?
Date Tue, 09 Nov 2010 14:29:25 GMT
W dniu 09.11.2010 15:18, Harsh J wrote:
> On Tue, Nov 9, 2010 at 7:34 PM, Wojciech Langiewicz
> <wlangiewicz@gmail.com>  wrote:
>> In tmp dir I have another 2: dfs and mapred
>> Directory 'dfs' seems to contain block from the HDFS, is it safe to delete
>> them?
> NO! You'll lose your blocks if you delete that dfs/data directory (If
> it is in use and I suppose that's what is taking up much space).
> In production, the dfs.data.dir property should not point any of its
> directories to a /tmp subdir which may be wiped upon boot.

Thank you very much for answer, this was as I suspected.

So in fact those files are real data inside HDFS ?

What might be causing situation where I have about 5TB in HDFS and 
hadoop tmp dirs have about 16TB in total?

Wojciech Langiewicz

View raw message