You are viewing a plain text version of this content. The canonical link for it is here.
Posted to hdfs-dev@hadoop.apache.org by Subru Krishnan <su...@apache.org> on 2017/10/02 22:39:54 UTC

Re: [VOTE] Merge Router-Based Federation (HDFS-10467) branch into trunk/branch-3

Thanks Inigo for driving this.

+1 (binding).

I have been involved during the design and have reviewed a couple of
patches. I feel this is a good addition and is well aligned with YARN
Federation (YARN-2915).

We deployed the HDFS Router in our test cluster and mapped it against 2
clusters - one running 3.0.0-beta1 and another 2.7.4 (to demonstrate
isolation) and it works fine.

-Subru

On Fri, Sep 29, 2017 at 11:58 AM, Iñigo Goiri <el...@gmail.com> wrote:

> Hi all,
>
> Given that 3.0-beta1 is already cut, I’d like to call a vote for merging
> Router-Based Federation (HDFS-10467) to trunk and branch-3.
>
> The vote will run for 7 days as usual.
>
>
>
> We started the discussion about merging HDFS-10467 a few weeks ago [1] and
> got good feedback which we’ve incorporated already [2, 3, 4].
>
> There are a couple tasks left:
>
>    - HDFS-12273 for the UI. This should be completed in the next couple
>    days.
>    - HDFS-12284 for adding security. We can move this for v2 if not
>    completed.
>
> We have deployed this in production for 2.7 and we did a few tests with
> trunk a few months ago.
>
> This week, I’m rebasing to trunk (last one was a couple weeks ago) and
> test trunk in one of our test clusters.
>
>
> Finally, note that all the functionality is in the Router (a new
> component) so everything is isolated.
>
> In addition, no new APIs have been added and we rely fully in
> ClientProtocol.
>
>
>
> I’d like to thank the people at Microsoft (specially, Jason, Ricardo,
> Chris, Subru, Jakob, Carlo and Giovanni), Twitter (Ming and Gera), LinkedIn
> (Zhe, Erik and Konstantin), and Cloudera (Andrew and Manoj) for
> the discussion and the ideas.
>
> Special thanks to Chris Douglas for the thorough reviews!
>
>
>
> Regards,
> Inigo
>
>
>
> [1] http://mail-archives.apache.org/mod_mbox/hadoop-hdfs-dev/201708.mbox/%
> 3CCAB1dGgogTu6kHtkkYeUycmNv-H3RupfPF4Cd7rpuFi6vHGdBLg%40mail.gmail.com%3E
>
> [2] https://issues.apache.org/jira/browse/HDFS-12381
>
> [3] https://issues.apache.org/jira/browse/HDFS-12430
>
> [4] https://issues.apache.org/jira/browse/HDFS-12450
>

Re: [VOTE] Merge Router-Based Federation (HDFS-10467) branch into trunk/branch-3

Posted by Chris Douglas <cd...@apache.org>.
+1 (binding)

Inigo has done an excellent job driving this work, both hardening it
in production and merging it into Apache.

The implementation, documentation, and tooling are mature enough to
ship with the 3.0.0 GA release. -C


On Mon, Oct 2, 2017 at 3:39 PM, Subru Krishnan <su...@apache.org> wrote:
> Thanks Inigo for driving this.
>
> +1 (binding).
>
> I have been involved during the design and have reviewed a couple of
> patches. I feel this is a good addition and is well aligned with YARN
> Federation (YARN-2915).
>
> We deployed the HDFS Router in our test cluster and mapped it against 2
> clusters - one running 3.0.0-beta1 and another 2.7.4 (to demonstrate
> isolation) and it works fine.
>
> -Subru
>
> On Fri, Sep 29, 2017 at 11:58 AM, Iñigo Goiri <el...@gmail.com> wrote:
>>
>> Hi all,
>>
>> Given that 3.0-beta1 is already cut, I’d like to call a vote for merging
>> Router-Based Federation (HDFS-10467) to trunk and branch-3.
>>
>> The vote will run for 7 days as usual.
>>
>>
>>
>> We started the discussion about merging HDFS-10467 a few weeks ago [1] and
>> got good feedback which we’ve incorporated already [2, 3, 4].
>>
>> There are a couple tasks left:
>>
>> HDFS-12273 for the UI. This should be completed in the next couple days.
>> HDFS-12284 for adding security. We can move this for v2 if not completed.
>>
>> We have deployed this in production for 2.7 and we did a few tests with
>> trunk a few months ago.
>>
>> This week, I’m rebasing to trunk (last one was a couple weeks ago) and
>> test trunk in one of our test clusters.
>>
>>
>> Finally, note that all the functionality is in the Router (a new
>> component) so everything is isolated.
>>
>> In addition, no new APIs have been added and we rely fully in
>> ClientProtocol.
>>
>>
>>
>> I’d like to thank the people at Microsoft (specially, Jason, Ricardo,
>> Chris, Subru, Jakob, Carlo and Giovanni), Twitter (Ming and Gera), LinkedIn
>> (Zhe, Erik and Konstantin), and Cloudera (Andrew and Manoj) for the
>> discussion and the ideas.
>>
>> Special thanks to Chris Douglas for the thorough reviews!
>>
>>
>>
>> Regards,
>> Inigo
>>
>>
>>
>> [1]
>> http://mail-archives.apache.org/mod_mbox/hadoop-hdfs-dev/201708.mbox/%3CCAB1dGgogTu6kHtkkYeUycmNv-H3RupfPF4Cd7rpuFi6vHGdBLg%40mail.gmail.com%3E
>>
>> [2] https://issues.apache.org/jira/browse/HDFS-12381
>>
>> [3] https://issues.apache.org/jira/browse/HDFS-12430
>>
>> [4] https://issues.apache.org/jira/browse/HDFS-12450
>
>

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