FAQ
Hi folks,



We had a weird thing where one of our data nodes was 100% disk full (all
hdfs data) and the other nodes were uniformly 20% space utilizated.



Just wondering if this is a bug or whether we made an operator error.
The node in question is not one of primary or secondary namenodes. It's
never been the case that we have been running with any significant
number of nodes dead (never been worse than 2).



The only thing is that I have often done 'dfs -put' of large files (from
a NFS mount) from this node. Would this case local storage to be
allocated by HDFS?



Thanks,



Joydeep

Search Discussions

  • Roddy Lindsay at Aug 17, 2007 at 11:07 pm
    actually, checking out hadoop008 (where I have been doing dfs -puts) the
    disk free drops off sharply, then goes back up. Seems the other nodes are
    at a continuous downward slope.


    On 8/17/07 3:25 PM, "Joydeep Sen Sarma" wrote:

    Hi folks,



    We had a weird thing where one of our data nodes was 100% disk full (all
    hdfs data) and the other nodes were uniformly 20% space utilizated.



    Just wondering if this is a bug or whether we made an operator error.
    The node in question is not one of primary or secondary namenodes. It's
    never been the case that we have been running with any significant
    number of nodes dead (never been worse than 2).



    The only thing is that I have often done 'dfs -put' of large files (from
    a NFS mount) from this node. Would this case local storage to be
    allocated by HDFS?



    Thanks,



    Joydeep
  • Roddy Lindsay at Aug 17, 2007 at 11:10 pm
    meant to send to joydeep and didn't check the reply-to.

    anyways, hello to you hadoop folks :)

    Roddy
  • Raghu Angadi at Aug 17, 2007 at 11:13 pm

    Joydeep Sen Sarma wrote:


    The only thing is that I have often done 'dfs -put' of large files (from
    a NFS mount) from this node. Would this case local storage to be
    allocated by HDFS?
    Yes. Local node stores one copy as long as it has space, if it is also
    part of the cluster.

    Raghu.
  • Joydeep Sen Sarma at Aug 17, 2007 at 11:48 pm
    Wow - interesting.

    I guess this is an optimization in case people are doing parallel loads
    eh? We'll have to watch out ..

    -----Original Message-----
    From: Raghu Angadi
    Sent: Friday, August 17, 2007 4:13 PM
    To: hadoop-user@lucene.apache.org
    Subject: Re: data nodes imbalanced

    Joydeep Sen Sarma wrote:
    The only thing is that I have often done 'dfs -put' of large files (from
    a NFS mount) from this node. Would this case local storage to be
    allocated by HDFS?
    Yes. Local node stores one copy as long as it has space, if it is also
    part of the cluster.

    Raghu.
  • Allen Wittenauer at Aug 18, 2007 at 12:25 am

    On 8/17/07 4:49 PM, "Joydeep Sen Sarma" wrote:

    Wow - interesting.

    I guess this is an optimization in case people are doing parallel loads
    eh? We'll have to watch out ..
    We basically do our data loads from a non-HDFS machine. This will cause
    the write to be done on random nodes.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupcommon-user @
categorieshadoop
postedAug 17, '07 at 10:24p
activeAug 18, '07 at 12:25a
posts6
users4
websitehadoop.apache.org...
irc#hadoop

People

Translate

site design / logo © 2021 Grokbase