You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@ambari.apache.org by Martin Serrano <ma...@attivio.com> on 2016/05/31 14:58:32 UTC

how is the data directory prefix determined?

Hello,

I'm trying to automate a cluster install.  Ambari gets installed and the
server comes up fine.  But the data directories that are defaulted are
all prefixed with '/dev/console'.  And at times I see '/dev/tty1', etc.
Where does this prefix setting come from?  How can I change it?  
I saw that there was a 'recommendations' REST API at some point,
but it is now gone.

This is preventing the cluster from starting services since this is not a real directory.

My environment:
  CentOS 6 running within LXC.
  Ambari 2.2

Thanks,
Martin Serrano

Re: how is the data directory prefix determined?

Posted by Martin Serrano <ma...@attivio.com>.
Alejandro,

Thank you for the reply. 

I am using Ambari-2.2.2.0 and HDP 2.4.2.0.  My mount points for the host
running the ambari-server are:

[root@foo-0 /]# df -h
Filesystem            Size  Used Avail Use% Mounted on
rootfs                 50G   23G   28G  45% /
/dev/mapper/centos-root
                       50G   23G   28G  45% /
devtmpfs               16G  4.0K   16G   1% /dev/full
devtmpfs               16G  4.0K   16G   1% /dev/null
devtmpfs               16G  4.0K   16G   1% /dev/random
devtmpfs               16G  4.0K   16G   1% /dev/tty
devtmpfs               16G  4.0K   16G   1% /dev/urandom
devtmpfs               16G  4.0K   16G   1% /dev/zero
tmpfs                  16G     0   16G   0% /dev/shm

-Martin

On 05/31/2016 12:55 PM, Alejandro Fernandez wrote:
> Hi Martin,
>
> That code exists in Stack Advisor. If data dir is not specified, it should
> default to /hadoop/hdfs/data
> Make sure to check your mount points.
> Which stack and version are you using?
>
> Thanks,
> Alejandro
>
> On 5/31/16, 7:58 AM, "Martin Serrano" <ma...@attivio.com> wrote:
>
>> Hello,
>>
>> I'm trying to automate a cluster install.  Ambari gets installed and the
>> server comes up fine.  But the data directories that are defaulted are
>> all prefixed with '/dev/console'.  And at times I see '/dev/tty1', etc.
>> Where does this prefix setting come from?  How can I change it?
>> I saw that there was a 'recommendations' REST API at some point,
>> but it is now gone.
>>
>> This is preventing the cluster from starting services since this is not a
>> real directory.
>>
>> My environment:
>>  CentOS 6 running within LXC.
>>  Ambari 2.2
>>
>> Thanks,
>> Martin Serrano
>>
>>
>

Re: how is the data directory prefix determined?

Posted by Alejandro Fernandez <af...@hortonworks.com>.
Hi Martin,

That code exists in Stack Advisor. If data dir is not specified, it should
default to /hadoop/hdfs/data
Make sure to check your mount points.
Which stack and version are you using?

Thanks,
Alejandro

On 5/31/16, 7:58 AM, "Martin Serrano" <ma...@attivio.com> wrote:

>Hello,
>
>I'm trying to automate a cluster install.  Ambari gets installed and the
>server comes up fine.  But the data directories that are defaulted are
>all prefixed with '/dev/console'.  And at times I see '/dev/tty1', etc.
>Where does this prefix setting come from?  How can I change it?
>I saw that there was a 'recommendations' REST API at some point,
>but it is now gone.
>
>This is preventing the cluster from starting services since this is not a
>real directory.
>
>My environment:
>  CentOS 6 running within LXC.
>  Ambari 2.2
>
>Thanks,
>Martin Serrano
>
>