You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@hadoop.apache.org by George Kousiouris <gk...@mail.ntua.gr> on 2012/09/20 16:39:58 UTC
any way to set hadoop daemons periodicity?
Hi all,
I have noticed that sometimes the hdfs takes too long to act upon a
specific change in the configuration e.g. the replication factor. While
the under/over replicated state of a file is immediately detected, the
corrective action may take some time, even if the file sizes are very
small and the cluster is idle.
Is there any way to make this process somehow predictable?E.g. through
some configuration setting regarding the time intervals of the daemons?
Thanks,
George
--
---------------------------
Re: any way to set hadoop daemons periodicity?
Posted by Robert Molina <rm...@hortonworks.com>.
Hi George,
I don't believe there is a setting to do what you are requesting. But
maybe the following may help regarding replication:
https://issues.apache.org/jira/browse/HDFS-3814
HTH
On Thu, Sep 20, 2012 at 7:39 AM, George Kousiouris <gk...@mail.ntua.gr>wrote:
>
> Hi all,
>
> I have noticed that sometimes the hdfs takes too long to act upon a
> specific change in the configuration e.g. the replication factor. While the
> under/over replicated state of a file is immediately detected, the
> corrective action may take some time, even if the file sizes are very small
> and the cluster is idle.
>
> Is there any way to make this process somehow predictable?E.g. through
> some configuration setting regarding the time intervals of the daemons?
>
> Thanks,
> George
>
> --
>
> ---------------------------
>
>
>
Re: any way to set hadoop daemons periodicity?
Posted by Robert Molina <rm...@hortonworks.com>.
Hi George,
I don't believe there is a setting to do what you are requesting. But
maybe the following may help regarding replication:
https://issues.apache.org/jira/browse/HDFS-3814
HTH
On Thu, Sep 20, 2012 at 7:39 AM, George Kousiouris <gk...@mail.ntua.gr>wrote:
>
> Hi all,
>
> I have noticed that sometimes the hdfs takes too long to act upon a
> specific change in the configuration e.g. the replication factor. While the
> under/over replicated state of a file is immediately detected, the
> corrective action may take some time, even if the file sizes are very small
> and the cluster is idle.
>
> Is there any way to make this process somehow predictable?E.g. through
> some configuration setting regarding the time intervals of the daemons?
>
> Thanks,
> George
>
> --
>
> ---------------------------
>
>
>
Re: any way to set hadoop daemons periodicity?
Posted by Robert Molina <rm...@hortonworks.com>.
Hi George,
I don't believe there is a setting to do what you are requesting. But
maybe the following may help regarding replication:
https://issues.apache.org/jira/browse/HDFS-3814
HTH
On Thu, Sep 20, 2012 at 7:39 AM, George Kousiouris <gk...@mail.ntua.gr>wrote:
>
> Hi all,
>
> I have noticed that sometimes the hdfs takes too long to act upon a
> specific change in the configuration e.g. the replication factor. While the
> under/over replicated state of a file is immediately detected, the
> corrective action may take some time, even if the file sizes are very small
> and the cluster is idle.
>
> Is there any way to make this process somehow predictable?E.g. through
> some configuration setting regarding the time intervals of the daemons?
>
> Thanks,
> George
>
> --
>
> ---------------------------
>
>
>
Re: any way to set hadoop daemons periodicity?
Posted by Robert Molina <rm...@hortonworks.com>.
Hi George,
I don't believe there is a setting to do what you are requesting. But
maybe the following may help regarding replication:
https://issues.apache.org/jira/browse/HDFS-3814
HTH
On Thu, Sep 20, 2012 at 7:39 AM, George Kousiouris <gk...@mail.ntua.gr>wrote:
>
> Hi all,
>
> I have noticed that sometimes the hdfs takes too long to act upon a
> specific change in the configuration e.g. the replication factor. While the
> under/over replicated state of a file is immediately detected, the
> corrective action may take some time, even if the file sizes are very small
> and the cluster is idle.
>
> Is there any way to make this process somehow predictable?E.g. through
> some configuration setting regarding the time intervals of the daemons?
>
> Thanks,
> George
>
> --
>
> ---------------------------
>
>
>