You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ambari.apache.org by "Adaryl \"Bob\" Wakefield, MBA" <ad...@hotmail.com> on 2015/11/03 08:47:12 UTC

HDFS disk usage

On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.
B.

Re: HDFS disk usage

Posted by Olivier Renault <or...@hortonworks.com>.
Hi Bob,

Ambari is filtering some mount point has being not valid places and /home is one of them. BTW, I’ve notice that you are use LVM – it’s recommended to not use LVM for hadoop data disk ( it adds a unnecessary overhead ) not a big deal if it’s just a play cluster.

Given that you are using LVM, you could take a backup of everything under /home destroy the LV and create a smaller one for home and a big one for hadoop.

Stop your cluster

# tar cvf home_backup.tar /home/*
# umount /home
# lvremove /dev/mapper/centos-home
# lvcreate –L +10G –n home centos
# mkfs.ext4 /dev/mapper/centos-home
# lvcreate –L +2.5T –n hadoop centos
# mkfs.ext4 /dev/mapper/centos-hadoop
# mount /dev/mapper/centos-home /home
# mount /dev/mapper/centos-hadoop /mnt
# cp –ax /hadoop/* /mnt/
# umount /mnt
# mount /dev/mapper/centos-hadoop /hadoop
# tar xvf home_backup.tar –C /home

Restart your cluster, check that it’s working – if it does, you should umount /hadoop and clean the folder underneath and then remount. Do not forget to also edit your /etc/fstab to get /hadoop mounted automatically at boot.

Thanks
Olivier

From: "Adaryl Wakefield, MBA Bob" <ad...@hotmail.com>>
Reply-To: "user@ambari.apache.org<ma...@ambari.apache.org>" <us...@ambari.apache.org>>
Date: Wednesday, 4 November 2015 at 23:02
To: "user@ambari.apache.org<ma...@ambari.apache.org>" <us...@ambari.apache.org>>
Subject: Re: HDFS disk usage

I’ve been advised to place the data in a directory that is in /home. I attempted to do that an got an error that said dfs.datanode.data.dir can’t be set to anything in the /home directory.

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Adaryl "Bob" Wakefield, MBA<ma...@hotmail.com>
Sent: Wednesday, November 04, 2015 2:18 PM
To: user@ambari.apache.org<ma...@ambari.apache.org>
Subject: Re: HDFS disk usage

I dug into this a little bit more. Apparently the default is 50GB.

On my datanodes, I ran df –hl. This is the output
Filesystem      Size    Used    Avail   Use%    Mounted on
/dev/mapper/centos-root 50G     12G     39G     23%     /
devtmpfs        16G     0       16G     0%      /dev
tmpfs   16G     0       16G     0%      /dev/shm
tmpfs   16G     1.4G    15G     9%      /run
tmpfs   16G     0       16G     0%      /sys/fs/cgroup
/dev/sda2       494M    123M    372M    25%     /boot
/dev/mapper/centos-home 2.7T    33M     2.7T    1%      /home

If I’m reading this right it HAS to be mounted correctly because no other mount points have enough space correct?


Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Adaryl "Bob" Wakefield, MBA<ma...@hotmail.com>
Sent: Tuesday, November 03, 2015 10:32 AM
To: user@ambari.apache.org<ma...@ambari.apache.org>
Subject: Re: HDFS disk usage

How would I go about doing that?

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Olivier Renault<ma...@hortonworks.com>
Sent: Tuesday, November 03, 2015 3:49 AM
To: user@ambari.apache.org<ma...@ambari.apache.org> ; user@ambari.apache.org<ma...@ambari.apache.org>
Subject: Re: HDFS disk usage


Could you double check that your datanodes are using the correct mount point?

Thanks,
Olivier
------
Olivier Renault
Solution Engineer
Mobile: +44 7500 933 036



On Tue, Nov 3, 2015 at 1:45 AM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com>> wrote:

Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault<ma...@hortonworks.com>
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org<ma...@ambari.apache.org> ; user@ambari.apache.org<ma...@ambari.apache.org>
Subject: Re: HDFS disk usage


It reports the space available for HDFS.

Thanks,
Olivier



On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com>> wrote:

On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.

B.

Re: HDFS disk usage

Posted by "Adaryl \"Bob\" Wakefield, MBA" <ad...@hotmail.com>.
I’ve been advised to place the data in a directory that is in /home. I attempted to do that an got an error that said dfs.datanode.data.dir can’t be set to anything in the /home directory.

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Adaryl "Bob" Wakefield, MBA 
Sent: Wednesday, November 04, 2015 2:18 PM
To: user@ambari.apache.org 
Subject: Re: HDFS disk usage

I dug into this a little bit more. Apparently the default is 50GB.

On my datanodes, I ran df –hl. This is the output
      Filesystem Size Used Avail Use% Mounted on 
      /dev/mapper/centos-root 50G 12G 39G 23% / 
      devtmpfs 16G 0 16G 0% /dev 
      tmpfs 16G 0 16G 0% /dev/shm 
      tmpfs 16G 1.4G 15G 9% /run 
      tmpfs 16G 0 16G 0% /sys/fs/cgroup 
      /dev/sda2 494M 123M 372M 25% /boot 
      /dev/mapper/centos-home 2.7T 33M 2.7T 1% /home 


If I’m reading this right it HAS to be mounted correctly because no other mount points have enough space correct?


Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Adaryl "Bob" Wakefield, MBA 
Sent: Tuesday, November 03, 2015 10:32 AM
To: user@ambari.apache.org 
Subject: Re: HDFS disk usage

How would I go about doing that?

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:49 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

Could you double check that your datanodes are using the correct mount point? 

Thanks,
Olivier
------
Olivier Renault
Solution Engineer
Mobile: +44 7500 933 036






On Tue, Nov 3, 2015 at 1:45 AM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

It reports the space available for HDFS. 

Thanks,
Olivier





On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.
B.

Re: HDFS disk usage

Posted by "Adaryl \"Bob\" Wakefield, MBA" <ad...@hotmail.com>.
I dug into this a little bit more. Apparently the default is 50GB.

On my datanodes, I ran df –hl. This is the output
      Filesystem Size Used Avail Use% Mounted on 
      /dev/mapper/centos-root 50G 12G 39G 23% / 
      devtmpfs 16G 0 16G 0% /dev 
      tmpfs 16G 0 16G 0% /dev/shm 
      tmpfs 16G 1.4G 15G 9% /run 
      tmpfs 16G 0 16G 0% /sys/fs/cgroup 
      /dev/sda2 494M 123M 372M 25% /boot 
      /dev/mapper/centos-home 2.7T 33M 2.7T 1% /home 


If I’m reading this right it HAS to be mounted correctly because no other mount points have enough space correct?


Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Adaryl "Bob" Wakefield, MBA 
Sent: Tuesday, November 03, 2015 10:32 AM
To: user@ambari.apache.org 
Subject: Re: HDFS disk usage

How would I go about doing that?

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:49 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

Could you double check that your datanodes are using the correct mount point? 

Thanks,
Olivier
------
Olivier Renault
Solution Engineer
Mobile: +44 7500 933 036






On Tue, Nov 3, 2015 at 1:45 AM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

It reports the space available for HDFS. 

Thanks,
Olivier





On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.
B.

Re: HDFS disk usage

Posted by "Adaryl \"Bob\" Wakefield, MBA" <ad...@hotmail.com>.
How would I go about doing that?

Adaryl "Bob" Wakefield, MBA
Principal
Mass Street Analytics, LLC
913.938.6685
www.linkedin.com/in/bobwakefieldmba
Twitter: @BobLovesData

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:49 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

Could you double check that your datanodes are using the correct mount point? 

Thanks,
Olivier
------
Olivier Renault
Solution Engineer
Mobile: +44 7500 933 036






On Tue, Nov 3, 2015 at 1:45 AM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

It reports the space available for HDFS. 

Thanks,
Olivier





On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.
B.

Re: HDFS disk usage

Posted by Olivier Renault <or...@hortonworks.com>.
Could you double check that your datanodes are using the correct mount point?

Thanks,
Olivier
------
Olivier Renault
Solution Engineer
Mobile: +44 7500 933 036



On Tue, Nov 3, 2015 at 1:45 AM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com>> wrote:

Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault<ma...@hortonworks.com>
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org<ma...@ambari.apache.org> ; user@ambari.apache.org<ma...@ambari.apache.org>
Subject: Re: HDFS disk usage


It reports the space available for HDFS.

Thanks,
Olivier



On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com>> wrote:

On the dashboard, what exactly is HDFS disk usage reporting? The numbers I'm seeing are WAY less than the total disk space on my cluster.

B.

Re: HDFS disk usage

Posted by "Adaryl \"Bob\" Wakefield, MBA" <ad...@hotmail.com>.
Why is there such a large discrepancy between what is reported and my actual disk size?

B.

From: Olivier Renault 
Sent: Tuesday, November 03, 2015 3:01 AM
To: user@ambari.apache.org ; user@ambari.apache.org 
Subject: Re: HDFS disk usage

It reports the space available for HDFS. 

Thanks,
Olivier





On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com> wrote:


On the dashboard, what exactly is HDFS disk usage reporting? The numbers I’m seeing are WAY less than the total disk space on my cluster.
B.

Re: HDFS disk usage

Posted by Olivier Renault <or...@hortonworks.com>.
It reports the space available for HDFS.

Thanks,
Olivier



On Mon, Nov 2, 2015 at 11:57 PM -0800, "Adaryl "Bob" Wakefield, MBA" <ad...@hotmail.com>> wrote:

On the dashboard, what exactly is HDFS disk usage reporting? The numbers I'm seeing are WAY less than the total disk space on my cluster.

B.