FAQ
Hello,

Our cluster has encounterd underreplicated blocks.

We understand the hdfs wil balance and restore the underreplicated blocks.

But what i would like to know , is are there references.

for example. 10 TB HDFS with 30% underreplicated blocks = ETA 5 hours
before HDFS has fixed the problem.

Are there any examples for this ?


To unsubscribe from this group and stop receiving emails from it, send an email to scm-users+unsubscribe@cloudera.org.

Search Discussions

  • Ge ko at Oct 4, 2013 at 12:50 pm
    Hi Bert,

    from my point of view it is hard to provide general numbers, since it
    heavily depends on
    a) the type of network interconnection between the nodes in your cluster
    (10/100/1000Mbit Ethernet, FibreChannel, ...)
    b) the available network bandwidth, since there are many other sources of
    network traffic e.g. MapReduce jobs, and
    c) the replication factor itself

    br...Gerd...

    On 4 October 2013 11:42, Bert Cauwelier wrote:


    Hello,

    Our cluster has encounterd underreplicated blocks.

    We understand the hdfs wil balance and restore the underreplicated blocks.

    But what i would like to know , is are there references.

    for example. 10 TB HDFS with 30% underreplicated blocks = ETA 5 hours
    before HDFS has fixed the problem.

    Are there any examples for this ?


    To unsubscribe from this group and stop receiving emails from it, send an
    email to scm-users+unsubscribe@cloudera.org.
    To unsubscribe from this group and stop receiving emails from it, send an email to scm-users+unsubscribe@cloudera.org.
  • Bert Cauwelier at Oct 7, 2013 at 7:16 am
    Lets say 1000Mbit Ethernet
    High available network bandwidth
    and a replication of 3.
    No running mapreduce jobs
    and casual sata disk's


    2013/10/4 ge ko <koenig.ulm@gmail.com>
    Hi Bert,

    from my point of view it is hard to provide general numbers, since it
    heavily depends on
    a) the type of network interconnection between the nodes in your cluster
    (10/100/1000Mbit Ethernet, FibreChannel, ...)
    b) the available network bandwidth, since there are many other sources of
    network traffic e.g. MapReduce jobs, and
    c) the replication factor itself

    br...Gerd...

    On 4 October 2013 11:42, Bert Cauwelier wrote:


    Hello,

    Our cluster has encounterd underreplicated blocks.

    We understand the hdfs wil balance and restore the underreplicated blocks.

    But what i would like to know , is are there references.

    for example. 10 TB HDFS with 30% underreplicated blocks = ETA 5 hours
    before HDFS has fixed the problem.

    Are there any examples for this ?


    To unsubscribe from this group and stop receiving emails from it, send
    an email to scm-users+unsubscribe@cloudera.org.
    To unsubscribe from this group and stop receiving emails from it, send an email to scm-users+unsubscribe@cloudera.org.

Related Discussions

Discussion Navigation
viewthread | post
Discussion Overview
groupscm-users @
categorieshadoop
postedOct 4, '13 at 9:44a
activeOct 7, '13 at 7:16a
posts3
users3
websitecloudera.com
irc#hadoop

People

Translate

site design / logo © 2022 Grokbase