hadoop-common-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Edward Capriolo <edlinuxg...@gmail.com>
Subject Re: lost+found files prevent DataNode formatting
Date Tue, 29 Sep 2009 16:11:36 GMT
On Tue, Sep 29, 2009 at 5:17 AM, Stas Oskin <stas.oskin@gmail.com> wrote:
> Hi.
>
> Question - will DataNode ever try to format again the directory after the
> initial format?
>
> Common sense says no, so if I erased them once and they ever come back, it
> should not impact DataNode in any way?
>
> Thanks again.
>
> 2009/9/29 Anthony Urso <anthony.urso@gmail.com>
>
>> Those are created by fsck and will come back.  They belong to the
>> filesystem and you shouldn't delete them.
>>
>> Instead, create subdirectories on those mount points and use them as
>> DFS directories.  E.g. create and use /mnt/dfs instead of /mnt
>>
>> Cheers,
>> Anthony
>>
>> On Mon, Sep 28, 2009 at 6:33 PM, Stas Oskin <stas.oskin@gmail.com> wrote:
>> > Hi.
>> >
>> > I'm trying to spread DataNode files over separate block devices, but
>> these
>> > have lost+found directories created.
>> >
>> > DataNode initialization fails because it can't erase them (probably
>> because
>> > it running under hadoop account).
>> >
>> > Any idea how to solve this? I can erase them manually, but what if they
>> > appear again in future?
>> >
>> > Thanks.
>> >
>>
>

The DataNode does not format. "hadoop namenode -format" formats the
storage on the name node. The Datanode conducts blockreports and
compares those to the namenode metadata and deletes unneeded  blocks.

Mime
View raw message