Mailing-List: contact dev-help@aurora.apache.org; run by ezmlm
Precedence: bulk
Reply-To: dev@aurora.apache.org
MIME-Version: 1.0
In-Reply-To: <CAKTzdV02o3upJp3cN3v3otsU6mbQvKGw53Y9VV5eVWJ0WzCWaQ@mail.gmail.com>
References: <CAKTzdV02o3upJp3cN3v3otsU6mbQvKGw53Y9VV5eVWJ0WzCWaQ@mail.gmail.com>
From: David McLaughlin <dmclaughlin@apache.org>
Date: Tue, 5 Sep 2017 10:17:09 -0700
Message-ID: <CAOOJoEyO+PsSyKmtPL+ExP+bE-M7Hk0gP7KkZ=CkPAoLfTapFw@mail.gmail.com>
Subject: Re: Make drain MAX_STATUS_WAIT configurable
To: dev@aurora.apache.org
Content-Type: multipart/alternative; boundary="f403045d6c56bbb816055874654b"
archived-at: Tue, 05 Sep 2017 17:17:14 -0000

--f403045d6c56bbb816055874654b
Content-Type: text/plain; charset="UTF-8"

+1

On Tue, Sep 5, 2017 at 10:09 AM, Mauricio Garavaglia <
mauriciogaravaglia@gmail.com> wrote:

> Hi folks,
>
> The aurora-admin drain command currently has a hardcoded limit of 5 minutes
> waiting for a node to be drained, after that timeout it fails.
>
> This doesn't work very well when tasks are expected be in killing state for
> more than that, for example, if the scheduler transient_task_state_timeout
> was
> adjusted.
>
> Any objection on making this 5 minutes MAX_STATUS_WAIT in aurora-admin
> drain configurable?
>
> https://github.com/apache/aurora/blob/master/src/main/
> python/apache/aurora/admin/host_maintenance.py#L40
>
> Thanks,
>
> Mauricio
>

--f403045d6c56bbb816055874654b--