hadoop-common-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Dhruba Borthakur <dhr...@gmail.com>
Subject Re: HDFS addBlock()
Date Mon, 23 Feb 2009 18:30:52 GMT
Hi Sangmin,

your assumtion is right. addBlock() allocates an ew block only if the
penultimate block has reached the minimum replication factor. The
blockreceived messages from the immediately previous block might take a
while to be sent by the datanode and processed by the namenode.

thanks,
dhruba


On Mon, Feb 23, 2009 at 4:50 AM, Sangmin Lee <sangmin.dev@gmail.com> wrote:

> Hi,
>
> I have a question regarding NameNode.addBlock() function.
> I see that it makes sure that the previous block is replicated before you
> assign an additional block.
> I wonder why it checks penultimate block rather than ultimate block ?
> Is it because it would take some time for the immediately previous block to
> be replicated?
>
> Thanks,
> Sangmin
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message