You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@mesos.apache.org by Andy Grove <an...@codefutures.com> on 2014/09/29 19:11:54 UTC

Mesos master failure question

Hi,

I'd like to clarify how master failures are handled. Do executors and tasks
continue operating while the master is down?

For example, if I have 3 master nodes and my scheduler causes a mysql
server to start on a slave and then the leader master fails, will mysql
keep running through this failure?

I read an article that said the slave stops all running tasks in the event
of a master failure, but I can't find that article now, and other articles
indicate otherwise. I couldn't find a conclusive answer in the mesos docs
either.

Thanks,

Andy.

--
Andy Grove
VP Engineering
CodeFutures Corporation

Re: Mesos master failure question

Posted by Andy Grove <an...@codefutures.com>.

Thanks, Vinod. That is good to hear.

Thanks,

Andy.

--
Andy Grove
VP Engineering
CodeFutures Corporation



On Mon, Sep 29, 2014 at 11:59 AM, Vinod Kone <vi...@gmail.com> wrote:

> Executors/tasks will keep running when the master is down. Any status
> updates sent by executors will be cached by the slaves and retried upon a
> new master being elected as leader.
>
>
> -- Vinod
>
> On Mon, Sep 29, 2014 at 10:11 AM, Andy Grove <an...@codefutures.com>
> wrote:
>
>> Hi,
>>
>> I'd like to clarify how master failures are handled. Do executors and
>> tasks continue operating while the master is down?
>>
>> For example, if I have 3 master nodes and my scheduler causes a mysql
>> server to start on a slave and then the leader master fails, will mysql
>> keep running through this failure?
>>
>> I read an article that said the slave stops all running tasks in the
>> event of a master failure, but I can't find that article now, and other
>> articles indicate otherwise. I couldn't find a conclusive answer in the
>> mesos docs either.
>>
>> Thanks,
>>
>> Andy.
>>
>> --
>> Andy Grove
>> VP Engineering
>> CodeFutures Corporation
>>
>>
>>
>

Re: Mesos master failure question

Posted by Vinod Kone <vi...@gmail.com>.

Executors/tasks will keep running when the master is down. Any status
updates sent by executors will be cached by the slaves and retried upon a
new master being elected as leader.


-- Vinod

On Mon, Sep 29, 2014 at 10:11 AM, Andy Grove <an...@codefutures.com>
wrote:

> Hi,
>
> I'd like to clarify how master failures are handled. Do executors and
> tasks continue operating while the master is down?
>
> For example, if I have 3 master nodes and my scheduler causes a mysql
> server to start on a slave and then the leader master fails, will mysql
> keep running through this failure?
>
> I read an article that said the slave stops all running tasks in the event
> of a master failure, but I can't find that article now, and other articles
> indicate otherwise. I couldn't find a conclusive answer in the mesos docs
> either.
>
> Thanks,
>
> Andy.
>
> --
> Andy Grove
> VP Engineering
> CodeFutures Corporation
>
>
>