You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@mesos.apache.org by Benjamin Mahler <be...@gmail.com> on 2013/08/29 01:12:13 UTC
Reverts
Hi all,
I've just reverted two commits that were causing issues on master:
commit 5e8e4da4e8276a022fef3667f3dc1415e7236149
Author: Benjamin Mahler <bm...@twitter.com>
Date: Wed Aug 28 15:46:20 2013 -0700
Reverted "Upgraded ZooKeeper from 3.3.4 to 3.3.6.".
This reverts commit eb1cd4a7c0ad4310f090d4f0643cf4059ac5246b.
commit 400a88f9817bb102522b08b83dca400380ad8a9b
Author: Vinod Kone <vi...@twitter.com>
Date: Tue Aug 27 13:36:42 2013 -0400
Added resource checks to slave recovery tests to ensure resources
are re-offered after restarts.
From: Ian Downes <ia...@gmail.com>
Review: https://reviews.apache.org/r/13764
I will follow up with the contributors to let them know.
Thanks!
Ben
Re: Reverts
Posted by Chris Mattmann <ma...@apache.org>.
Hey Ben,
No biggie, you've done the due diligence to communicate with folks
on the team, so you are doing fine, I just didn't know since some of
the convos were off-list. But you did a great job to try and bring the
conversation
on list, prompting my email and prompting your reply which shows the
due diligence you've done.
Thanks Ben. Appreciate the discussion.
Cheers,
Chris
-----Original Message-----
From: Benjamin Mahler <bm...@twitter.com>
Reply-To: "dev@mesos.apache.org" <de...@mesos.apache.org>
Date: Wednesday, August 28, 2013 8:57 PM
To: "dev@mesos.apache.org" <de...@mesos.apache.org>
Cc: "dev@mesos.apache.org" <de...@mesos.apache.org>, "idownes@twitter.com"
<id...@twitter.com>, "vlee@twitter.com" <vl...@twitter.com>
Subject: Re: Reverts
>+Ian, Vinson
>
>I agree strongly and would generally prefer to not have reverts (they
>have been very rare in the project and we should keep it that way).
>
>Just to give a little more context on this case:
>
>With Ian's change I spoke with him personally to hear where he was with a
>fix and he was happy to revert, so I opted for that so he wouldn't be
>rushed working on his changes. In retrospect I would have preferred to
>have a fix than to revert on this.
>
>For the ZooKeeper change, it may be a library regression and so we needed
>to revert it until we can figure out whether this is indeed a library
>regression. Please see the MESOS ticket and relevant ZK ticket.
>
>I'll definitely bring up a discussion on the mailing list next time,
>thanks for checking me on this Chris!
>
>Sent from my iPhone
>
>On Aug 28, 2013, at 7:44 PM, Chris Mattmann <ma...@apache.org> wrote:
>
>> Hi Ben,
>>
>> A revert is a pretty strict/harsh thing to do. In the future, why not
>> simply fix your local copy of the code, submit JIRA tickets, attach
>> Review Board patches, etc., but please, reverts should not be used
>> in practice -- especially without discussion.
>>
>> Cheers,
>> Chris
>>
>>
>> -----Original Message-----
>> From: Benjamin Mahler <be...@gmail.com>
>> Reply-To: "dev@mesos.apache.org" <de...@mesos.apache.org>
>> Date: Wednesday, August 28, 2013 4:12 PM
>> To: dev <de...@mesos.apache.org>
>> Subject: Reverts
>>
>>> Hi all,
>>>
>>> I've just reverted two commits that were causing issues on master:
>>>
>>> commit 5e8e4da4e8276a022fef3667f3dc1415e7236149
>>> Author: Benjamin Mahler <bm...@twitter.com>
>>> Date: Wed Aug 28 15:46:20 2013 -0700
>>>
>>> Reverted "Upgraded ZooKeeper from 3.3.4 to 3.3.6.".
>>>
>>> This reverts commit eb1cd4a7c0ad4310f090d4f0643cf4059ac5246b.
>>>
>>> commit 400a88f9817bb102522b08b83dca400380ad8a9b
>>> Author: Vinod Kone <vi...@twitter.com>
>>> Date: Tue Aug 27 13:36:42 2013 -0400
>>>
>>> Added resource checks to slave recovery tests to ensure resources
>>> are re-offered after restarts.
>>>
>>> From: Ian Downes <ia...@gmail.com>
>>> Review: https://reviews.apache.org/r/13764
>>>
>>> I will follow up with the contributors to let them know.
>>>
>>> Thanks!
>>> Ben
>>
>>
Re: Reverts
Posted by Benjamin Mahler <bm...@twitter.com>.
+Ian, Vinson
I agree strongly and would generally prefer to not have reverts (they have been very rare in the project and we should keep it that way).
Just to give a little more context on this case:
With Ian's change I spoke with him personally to hear where he was with a fix and he was happy to revert, so I opted for that so he wouldn't be rushed working on his changes. In retrospect I would have preferred to have a fix than to revert on this.
For the ZooKeeper change, it may be a library regression and so we needed to revert it until we can figure out whether this is indeed a library regression. Please see the MESOS ticket and relevant ZK ticket.
I'll definitely bring up a discussion on the mailing list next time, thanks for checking me on this Chris!
Sent from my iPhone
On Aug 28, 2013, at 7:44 PM, Chris Mattmann <ma...@apache.org> wrote:
> Hi Ben,
>
> A revert is a pretty strict/harsh thing to do. In the future, why not
> simply fix your local copy of the code, submit JIRA tickets, attach
> Review Board patches, etc., but please, reverts should not be used
> in practice -- especially without discussion.
>
> Cheers,
> Chris
>
>
> -----Original Message-----
> From: Benjamin Mahler <be...@gmail.com>
> Reply-To: "dev@mesos.apache.org" <de...@mesos.apache.org>
> Date: Wednesday, August 28, 2013 4:12 PM
> To: dev <de...@mesos.apache.org>
> Subject: Reverts
>
>> Hi all,
>>
>> I've just reverted two commits that were causing issues on master:
>>
>> commit 5e8e4da4e8276a022fef3667f3dc1415e7236149
>> Author: Benjamin Mahler <bm...@twitter.com>
>> Date: Wed Aug 28 15:46:20 2013 -0700
>>
>> Reverted "Upgraded ZooKeeper from 3.3.4 to 3.3.6.".
>>
>> This reverts commit eb1cd4a7c0ad4310f090d4f0643cf4059ac5246b.
>>
>> commit 400a88f9817bb102522b08b83dca400380ad8a9b
>> Author: Vinod Kone <vi...@twitter.com>
>> Date: Tue Aug 27 13:36:42 2013 -0400
>>
>> Added resource checks to slave recovery tests to ensure resources
>> are re-offered after restarts.
>>
>> From: Ian Downes <ia...@gmail.com>
>> Review: https://reviews.apache.org/r/13764
>>
>> I will follow up with the contributors to let them know.
>>
>> Thanks!
>> Ben
>
>
Re: Reverts
Posted by Chris Mattmann <ma...@apache.org>.
Hi Ben,
A revert is a pretty strict/harsh thing to do. In the future, why not
simply fix your local copy of the code, submit JIRA tickets, attach
Review Board patches, etc., but please, reverts should not be used
in practice -- especially without discussion.
Cheers,
Chris
-----Original Message-----
From: Benjamin Mahler <be...@gmail.com>
Reply-To: "dev@mesos.apache.org" <de...@mesos.apache.org>
Date: Wednesday, August 28, 2013 4:12 PM
To: dev <de...@mesos.apache.org>
Subject: Reverts
>Hi all,
>
>I've just reverted two commits that were causing issues on master:
>
>commit 5e8e4da4e8276a022fef3667f3dc1415e7236149
>Author: Benjamin Mahler <bm...@twitter.com>
>Date: Wed Aug 28 15:46:20 2013 -0700
>
> Reverted "Upgraded ZooKeeper from 3.3.4 to 3.3.6.".
>
> This reverts commit eb1cd4a7c0ad4310f090d4f0643cf4059ac5246b.
>
>commit 400a88f9817bb102522b08b83dca400380ad8a9b
>Author: Vinod Kone <vi...@twitter.com>
>Date: Tue Aug 27 13:36:42 2013 -0400
>
> Added resource checks to slave recovery tests to ensure resources
> are re-offered after restarts.
>
> From: Ian Downes <ia...@gmail.com>
> Review: https://reviews.apache.org/r/13764
>
>I will follow up with the contributors to let them know.
>
>Thanks!
>Ben