You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-dev@hadoop.apache.org by Wangda Tan <wh...@gmail.com> on 2016/02/25 02:10:46 UTC
[VOTE] Merge YARN-3368 to trunk
YARN Devs,
We propose to merge YARN-3368 (YARN next generation UI) development branch
into trunk. The reason to do the merge is web UI of Timeline server v2
(YARN-2928) is depending on the framework, and YARN-2928 is planned to
merge back to trunk shortly.
Status of YARN-3368:
Now it is runnable from end-to-end, which includes:
* Scheduler page, which contains a select-able queue hierarchy view and
also related scheduler information.
* Apps table, app page, app-attempt page and also timeline view of
app-attempts/containers.
* Cluster overview page, which contains several charts to show overview of
the cluster (such as total memory, total apps, etc.)
We plan to continue develop this feature after merge back to trunk, please
refer to sub tasks of YARN-3368.
Since YARN-3368 hasn't integrated to neither Hadoop's maven build system
nor RM web service, so the merge is completely safe. Ticket for the merge
task is: YARN-4734.
The vote will run for 7 days, ending Wednesday, Mar 2. Please feel free to
comment if you have any questions/doubts. I'll start with my +1 (binding).
Thanks,
Wangda
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Li Lu <ll...@hortonworks.com>.
+1 (binding)
Thanks Wangda for proposing this. Right now YARN-2928’s proposed UI work is heavily based on YARN-3368. Merging back YARN-3368 to trunk will unblock the future progress of the next gen ATS work.
Meanwhile, the code used in YARN-3368 is completely separated with our existing YARN UIs (the new UI uses ember.js and works with YARN’s REST APIs). This means the new UI can land in trunk as an alpha and safely co-exist with the stable UI. We may potentially attract more contributors to actively working on the new UI after the merge.
Li Lu
> On Feb 24, 2016, at 17:10, Wangda Tan <wh...@gmail.com> wrote:
>
> YARN Devs,
>
> We propose to merge YARN-3368 (YARN next generation UI) development branch
> into trunk. The reason to do the merge is web UI of Timeline server v2
> (YARN-2928) is depending on the framework, and YARN-2928 is planned to
> merge back to trunk shortly.
>
> Status of YARN-3368:
> Now it is runnable from end-to-end, which includes:
> * Scheduler page, which contains a select-able queue hierarchy view and
> also related scheduler information.
> * Apps table, app page, app-attempt page and also timeline view of
> app-attempts/containers.
> * Cluster overview page, which contains several charts to show overview of
> the cluster (such as total memory, total apps, etc.)
>
> We plan to continue develop this feature after merge back to trunk, please
> refer to sub tasks of YARN-3368.
>
> Since YARN-3368 hasn't integrated to neither Hadoop's maven build system
> nor RM web service, so the merge is completely safe. Ticket for the merge
> task is: YARN-4734.
>
> The vote will run for 7 days, ending Wednesday, Mar 2. Please feel free to
> comment if you have any questions/doubts. I'll start with my +1 (binding).
>
> Thanks,
> Wangda
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Wangda Tan <wh...@gmail.com>.
I want to restart merge vote:
Updates: addressed comments from @aw/@chris, fixed:
*Licenses:*
- ASF licenses issues
- Licenses for dependencies (Added to top level LICENSE.txt). There’s
nothing new to add to NOTICE.txt
*Building:*
- Integrated yarn-ui building to Maven. Instead of build it by default, I
placed it under a separate profile (-Pyarn-ui), only people who have
interests for this in-development feature can build/try it.
- Dockerfile support for dependencies
*Documentations:*
- Modified BUILDING.txt includes how to build yarn-ui
- Moved tutorial doc to
site: hadoop-yarn-project/hadoop-yarn/hadoop-yarn-site/src/site/markdown/YarnUI2.md
*Other changes:*
- Removed Unnecessary dot files, including all .gitkeepers
- .gitignore is move to top dir
- Cleaned up all hard-coded fields for service addresses, such as
rm-address / ATS-address, etc. (YARN-4514)
Merge JIRA for Jenkins is:
<https://issues.apache.org/jira/browse/YARN-4734>
https://issues.apache.org/jira/browse/YARN-4734
Latest Jenkins gets back with +1
https://issues.apache.org/jira/browse/YARN-4734?focusedCommentId=15248772&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-15248772.
(Unit tests failure are not related).
(Again :D), the vote will run for 7 days, ending Wednesday, Apr 27. Please
feel free to comment if you have any questions/doubts. I'll start with my
+1 (binding).
+ Chris/Allen
Thanks,
Wangda
On Mon, Feb 29, 2016 at 10:50 PM, Wangda Tan <wh...@gmail.com> wrote:
> Hi Chris,
>
> Sure, will address comments at YARN-4734 and send a separate DISCUSS
> thread before call for vote.
>
> Thanks,
> Wangda
>
> On Mon, Feb 29, 2016 at 10:40 PM, Allen Wittenauer <aw...@altiscale.com>
> wrote:
>
>>
>> > On Feb 29, 2016, at 9:33 PM, Chris Douglas <cd...@apache.org> wrote:
>> >
>> > On Thu, Feb 25, 2016 at 8:10 AM, Allen Wittenauer <aw...@altiscale.com>
>> wrote:
>> >> -1 (non-binding)
>> >
>> > Your -1 is binding, per the bylaws. If you intend it only in protest,
>> > -0 is typical.
>>
>> I had it in my head that branch merges requires PMC +1s, not
>> committers. It’s definitely not a protest -1.
>>
>> Thanks.
>
>
>
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Wangda Tan <wh...@gmail.com>.
Hi Chris,
Sure, will address comments at YARN-4734 and send a separate DISCUSS thread
before call for vote.
Thanks,
Wangda
On Mon, Feb 29, 2016 at 10:40 PM, Allen Wittenauer <aw...@altiscale.com> wrote:
>
> > On Feb 29, 2016, at 9:33 PM, Chris Douglas <cd...@apache.org> wrote:
> >
> > On Thu, Feb 25, 2016 at 8:10 AM, Allen Wittenauer <aw...@altiscale.com>
> wrote:
> >> -1 (non-binding)
> >
> > Your -1 is binding, per the bylaws. If you intend it only in protest,
> > -0 is typical.
>
> I had it in my head that branch merges requires PMC +1s, not
> committers. It’s definitely not a protest -1.
>
> Thanks.
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Allen Wittenauer <aw...@altiscale.com>.
> On Feb 29, 2016, at 9:33 PM, Chris Douglas <cd...@apache.org> wrote:
>
> On Thu, Feb 25, 2016 at 8:10 AM, Allen Wittenauer <aw...@altiscale.com> wrote:
>> -1 (non-binding)
>
> Your -1 is binding, per the bylaws. If you intend it only in protest,
> -0 is typical.
I had it in my head that branch merges requires PMC +1s, not committers. It’s definitely not a protest -1.
Thanks.
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Chris Douglas <cd...@apache.org>.
On Thu, Feb 25, 2016 at 8:10 AM, Allen Wittenauer <aw...@altiscale.com> wrote:
> -1 (non-binding)
Your -1 is binding, per the bylaws. If you intend it only in protest,
-0 is typical.
Wangda, the YARN-1197 merge was pretty aggressive. It might be worth
raising a [DISCUSS] thread first, to leave time to address feedback.
-C
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Wangda Tan <wh...@gmail.com>.
Hi Allen,
Thanks for commenting,
I will update patch and keep this thread updated.
Wangda
On Thu, Feb 25, 2016 at 8:10 AM, Allen Wittenauer <aw...@altiscale.com> wrote:
>
> > On Feb 24, 2016, at 5:10 PM, Wangda Tan <wh...@gmail.com> wrote:
> >
> >
> > Since YARN-3368 hasn't integrated to neither Hadoop's maven build system
> > nor RM web service, so the merge is completely safe. Ticket for the merge
> > task is: YARN-4734.
>
>
> -1 (non-binding)
>
> I don’t even have to run the build to see that:
>
> a) dot files in subdirectories that won’t work as intended
> b) ASF license violations all over the place
>
> This branch is definitely not ready to merge (the patch doesn’t even
> apply!) and I’m *blown away* a PMC member would even suggest it is ready.
Re: [VOTE] Merge YARN-3368 to trunk
Posted by Allen Wittenauer <aw...@altiscale.com>.
> On Feb 24, 2016, at 5:10 PM, Wangda Tan <wh...@gmail.com> wrote:
>
>
> Since YARN-3368 hasn't integrated to neither Hadoop's maven build system
> nor RM web service, so the merge is completely safe. Ticket for the merge
> task is: YARN-4734.
-1 (non-binding)
I don’t even have to run the build to see that:
a) dot files in subdirectories that won’t work as intended
b) ASF license violations all over the place
This branch is definitely not ready to merge (the patch doesn’t even apply!) and I’m *blown away* a PMC member would even suggest it is ready.