should we add a start-masters.sh script in sbin?

[sequoiadb <ma...@sequoiadb.com>]

Hey,

start-slaves.sh script is able to read from slaves file and start slaves node in multiple boxes.
However in standalone mode if I want to use multiple masters, I’ll have to start masters in each individual box, and also need to provide the list of masters’ hostname+port to each worker. ( start-slaves.sh only take 1 master ip+port for now)
I wonder should we create a new script called start-masters.sh to read conf/masters file? Also start-slaves.sh script may need to change a little bit so that master list can be passed to worker nodes.

Thanks

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@spark.apache.org
For additional commands, e-mail: dev-help@spark.apache.org

Re: should we add a start-masters.sh script in sbin?

[Ted Yu <yu...@gmail.com>]

Sounds good to me.

On Tue, Mar 31, 2015 at 6:12 PM, sequoiadb <ma...@sequoiadb.com>
wrote:

> Hey,
>
> start-slaves.sh script is able to read from slaves file and start slaves
> node in multiple boxes.
> However in standalone mode if I want to use multiple masters, I’ll have to
> start masters in each individual box, and also need to provide the list of
> masters’ hostname+port to each worker. ( start-slaves.sh only take 1 master
> ip+port for now)
> I wonder should we create a new script called start-masters.sh to read
> conf/masters file? Also start-slaves.sh script may need to change a little
> bit so that master list can be passed to worker nodes.
>
> Thanks
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@spark.apache.org
> For additional commands, e-mail: dev-help@spark.apache.org
>
>