You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@tez.apache.org by "Jianfeng (Jeff) Zhang" <jz...@hortonworks.com> on 2015/12/07 07:15:17 UTC

[DISCUSS] Release for 0.7.1

We have TEZ-0.7.0 release around half an year ago. During the last 6 months, we got lots of feedback from tez users, and fix many bugs. I am thinking this is a good time for the 0.7.1 release now.  And for this release, we would release binary version of tez for different hadoop versions for the first time. The candidate hadoop version in my mind is 2.2/2.4/2.6/2.7

I have go through tickets targeted for 0.7.1 and move out some tickets to 0.7.2 which I think is not critical.  And there's still 21 tickets left, you can check them on the following link, if you feel that's critical for you, please reply this thread and mark the ticket as blocker of 0.7.1
https://issues.apache.org/jira/issues/?jql=project%20%3D%20TEZ%20AND%20%22Target%20Version%2Fs%22%20%3D%200.7.1%20AND%20resolution%20%3D%20Unresolved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC



Best Regard,
Jeff Zhang


Re: [DISCUSS] Release for 0.7.1

Posted by "Jianfeng (Jeff) Zhang" <jz...@hortonworks.com>.
I have moved out all of tickets except TEZ-2594 (Fix licensing and notice
file for minimal tar-ball)
https://issues.apache.org/jira/browse/TEZ-2594

 
After this is resolved, I will start the RC.


Best Regard,
Jeff Zhang





On 12/10/15, 10:26 AM, "Jianfeng (Jeff) Zhang" <jz...@hortonworks.com>
wrote:

>
>Hi Folks,
>
>Some of tickets are Joanthan mentioned are in progress now.
>I will create branch 0.7.1 next Monday and will move these tickets to
>0.7.2 if it is not resolved at that time.
>
>
>Best Regard,
>Jeff Zhang
>
>
>
>
>
>On 12/8/15, 2:47 PM, "Siddharth Seth" <ss...@apache.org> wrote:
>
>>Jeff, Should we put a time bound on when 0.7.1 will be created. Not all
>>of
>>the 21 jiras are critical IMHO, and I don't believe all of them are
>>actively being worked on. Maybe the reporters / owners can move some of
>>these out to the next release if they're not going to be worked on soon.
>>
>>Thanks
>>- Sid
>>
>>On Mon, Dec 7, 2015 at 3:08 PM, Jonathan Eagles <je...@gmail.com>
>>wrote:
>>
>>> Did a quick scan of the jiras targeted to 0.7.1,
>>>
>>> Of the five listed below, two already have a +1, and the other three
>>>are in
>>> patch available. Of the 5, TEZ-2943 will have the largest benefit to
>>> customers.
>>>
>>> TEZ-1491 <https://issues.apache.org/jira/browse/TEZ-1491>
>>>
>>> Tez reducer-side merge's counter update is slow
>>> <https://issues.apache.org/jira/browse/TEZ-1491>
>>> TEZ-2824 <https://issues.apache.org/jira/browse/TEZ-2824>
>>>
>>> Add javadocs for Vertex.setConf and DAG.setConf
>>> <https://issues.apache.org/jira/browse/TEZ-2824>
>>> TEZ-2943 <https://issues.apache.org/jira/browse/TEZ-2943>
>>>
>>> Auto reduce parallelism estimates badly on vertices with two disparate
>>> input sources <https://issues.apache.org/jira/browse/TEZ-2943>
>>> TEZ-2966 <https://issues.apache.org/jira/browse/TEZ-2966>
>>>
>>> Tez does not honor mapreduce.task.timeout
>>> <https://issues.apache.org/jira/browse/TEZ-2966>
>>> TEZ-2972 <https://issues.apache.org/jira/browse/TEZ-2972>
>>>
>>> Ability for Tez AM to ignore node updates from YARN
>>> <https://issues.apache.org/jira/browse/TEZ-2972>
>>>
>>> On Mon, Dec 7, 2015 at 12:15 AM, Jianfeng (Jeff) Zhang <
>>> jzhang@hortonworks.com> wrote:
>>>
>>> >
>>> > We have TEZ-0.7.0 release around half an year ago. During the last 6
>>> > months, we got lots of feedback from tez users, and fix many bugs. I
>>>am
>>> > thinking this is a good time for the 0.7.1 release now.  And for this
>>> > release, we would release binary version of tez for different hadoop
>>> > versions for the first time. The candidate hadoop version in my mind
>>>is
>>> > 2.2/2.4/2.6/2.7
>>> >
>>> > I have go through tickets targeted for 0.7.1 and move out some
>>>tickets to
>>> > 0.7.2 which I think is not critical.  And there's still 21 tickets
>>>left,
>>> > you can check them on the following link, if you feel that's critical
>>>for
>>> > you, please reply this thread and mark the ticket as blocker of 0.7.1
>>> >
>>> >
>>> 
>>>https://issues.apache.org/jira/issues/?jql=project%20%3D%20TEZ%20AND%20%
>>>2
>>>2Target%20Version%2Fs%22%20%3D%200.7.1%20AND%20resolution%20%3D%20Unreso
>>>l
>>>ved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC
>>> >
>>> >
>>> >
>>> > Best Regard,
>>> > Jeff Zhang
>>> >
>>> >
>>>
>
>


Re: [DISCUSS] Release for 0.7.1

Posted by "Jianfeng (Jeff) Zhang" <jz...@hortonworks.com>.
Hi Folks,

Some of tickets are Joanthan mentioned are in progress now.
I will create branch 0.7.1 next Monday and will move these tickets to
0.7.2 if it is not resolved at that time.


Best Regard,
Jeff Zhang





On 12/8/15, 2:47 PM, "Siddharth Seth" <ss...@apache.org> wrote:

>Jeff, Should we put a time bound on when 0.7.1 will be created. Not all of
>the 21 jiras are critical IMHO, and I don't believe all of them are
>actively being worked on. Maybe the reporters / owners can move some of
>these out to the next release if they're not going to be worked on soon.
>
>Thanks
>- Sid
>
>On Mon, Dec 7, 2015 at 3:08 PM, Jonathan Eagles <je...@gmail.com> wrote:
>
>> Did a quick scan of the jiras targeted to 0.7.1,
>>
>> Of the five listed below, two already have a +1, and the other three
>>are in
>> patch available. Of the 5, TEZ-2943 will have the largest benefit to
>> customers.
>>
>> TEZ-1491 <https://issues.apache.org/jira/browse/TEZ-1491>
>>
>> Tez reducer-side merge's counter update is slow
>> <https://issues.apache.org/jira/browse/TEZ-1491>
>> TEZ-2824 <https://issues.apache.org/jira/browse/TEZ-2824>
>>
>> Add javadocs for Vertex.setConf and DAG.setConf
>> <https://issues.apache.org/jira/browse/TEZ-2824>
>> TEZ-2943 <https://issues.apache.org/jira/browse/TEZ-2943>
>>
>> Auto reduce parallelism estimates badly on vertices with two disparate
>> input sources <https://issues.apache.org/jira/browse/TEZ-2943>
>> TEZ-2966 <https://issues.apache.org/jira/browse/TEZ-2966>
>>
>> Tez does not honor mapreduce.task.timeout
>> <https://issues.apache.org/jira/browse/TEZ-2966>
>> TEZ-2972 <https://issues.apache.org/jira/browse/TEZ-2972>
>>
>> Ability for Tez AM to ignore node updates from YARN
>> <https://issues.apache.org/jira/browse/TEZ-2972>
>>
>> On Mon, Dec 7, 2015 at 12:15 AM, Jianfeng (Jeff) Zhang <
>> jzhang@hortonworks.com> wrote:
>>
>> >
>> > We have TEZ-0.7.0 release around half an year ago. During the last 6
>> > months, we got lots of feedback from tez users, and fix many bugs. I
>>am
>> > thinking this is a good time for the 0.7.1 release now.  And for this
>> > release, we would release binary version of tez for different hadoop
>> > versions for the first time. The candidate hadoop version in my mind
>>is
>> > 2.2/2.4/2.6/2.7
>> >
>> > I have go through tickets targeted for 0.7.1 and move out some
>>tickets to
>> > 0.7.2 which I think is not critical.  And there's still 21 tickets
>>left,
>> > you can check them on the following link, if you feel that's critical
>>for
>> > you, please reply this thread and mark the ticket as blocker of 0.7.1
>> >
>> >
>> 
>>https://issues.apache.org/jira/issues/?jql=project%20%3D%20TEZ%20AND%20%2
>>2Target%20Version%2Fs%22%20%3D%200.7.1%20AND%20resolution%20%3D%20Unresol
>>ved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC
>> >
>> >
>> >
>> > Best Regard,
>> > Jeff Zhang
>> >
>> >
>>


Re: [DISCUSS] Release for 0.7.1

Posted by Siddharth Seth <ss...@apache.org>.
Jeff, Should we put a time bound on when 0.7.1 will be created. Not all of
the 21 jiras are critical IMHO, and I don't believe all of them are
actively being worked on. Maybe the reporters / owners can move some of
these out to the next release if they're not going to be worked on soon.

Thanks
- Sid

On Mon, Dec 7, 2015 at 3:08 PM, Jonathan Eagles <je...@gmail.com> wrote:

> Did a quick scan of the jiras targeted to 0.7.1,
>
> Of the five listed below, two already have a +1, and the other three are in
> patch available. Of the 5, TEZ-2943 will have the largest benefit to
> customers.
>
> TEZ-1491 <https://issues.apache.org/jira/browse/TEZ-1491>
>
> Tez reducer-side merge's counter update is slow
> <https://issues.apache.org/jira/browse/TEZ-1491>
> TEZ-2824 <https://issues.apache.org/jira/browse/TEZ-2824>
>
> Add javadocs for Vertex.setConf and DAG.setConf
> <https://issues.apache.org/jira/browse/TEZ-2824>
> TEZ-2943 <https://issues.apache.org/jira/browse/TEZ-2943>
>
> Auto reduce parallelism estimates badly on vertices with two disparate
> input sources <https://issues.apache.org/jira/browse/TEZ-2943>
> TEZ-2966 <https://issues.apache.org/jira/browse/TEZ-2966>
>
> Tez does not honor mapreduce.task.timeout
> <https://issues.apache.org/jira/browse/TEZ-2966>
> TEZ-2972 <https://issues.apache.org/jira/browse/TEZ-2972>
>
> Ability for Tez AM to ignore node updates from YARN
> <https://issues.apache.org/jira/browse/TEZ-2972>
>
> On Mon, Dec 7, 2015 at 12:15 AM, Jianfeng (Jeff) Zhang <
> jzhang@hortonworks.com> wrote:
>
> >
> > We have TEZ-0.7.0 release around half an year ago. During the last 6
> > months, we got lots of feedback from tez users, and fix many bugs. I am
> > thinking this is a good time for the 0.7.1 release now.  And for this
> > release, we would release binary version of tez for different hadoop
> > versions for the first time. The candidate hadoop version in my mind is
> > 2.2/2.4/2.6/2.7
> >
> > I have go through tickets targeted for 0.7.1 and move out some tickets to
> > 0.7.2 which I think is not critical.  And there's still 21 tickets left,
> > you can check them on the following link, if you feel that's critical for
> > you, please reply this thread and mark the ticket as blocker of 0.7.1
> >
> >
> https://issues.apache.org/jira/issues/?jql=project%20%3D%20TEZ%20AND%20%22Target%20Version%2Fs%22%20%3D%200.7.1%20AND%20resolution%20%3D%20Unresolved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC
> >
> >
> >
> > Best Regard,
> > Jeff Zhang
> >
> >
>

Re: [DISCUSS] Release for 0.7.1

Posted by Jonathan Eagles <je...@gmail.com>.
Did a quick scan of the jiras targeted to 0.7.1,

Of the five listed below, two already have a +1, and the other three are in
patch available. Of the 5, TEZ-2943 will have the largest benefit to
customers.

TEZ-1491 <https://issues.apache.org/jira/browse/TEZ-1491>

Tez reducer-side merge's counter update is slow
<https://issues.apache.org/jira/browse/TEZ-1491>
TEZ-2824 <https://issues.apache.org/jira/browse/TEZ-2824>

Add javadocs for Vertex.setConf and DAG.setConf
<https://issues.apache.org/jira/browse/TEZ-2824>
TEZ-2943 <https://issues.apache.org/jira/browse/TEZ-2943>

Auto reduce parallelism estimates badly on vertices with two disparate
input sources <https://issues.apache.org/jira/browse/TEZ-2943>
TEZ-2966 <https://issues.apache.org/jira/browse/TEZ-2966>

Tez does not honor mapreduce.task.timeout
<https://issues.apache.org/jira/browse/TEZ-2966>
TEZ-2972 <https://issues.apache.org/jira/browse/TEZ-2972>

Ability for Tez AM to ignore node updates from YARN
<https://issues.apache.org/jira/browse/TEZ-2972>

On Mon, Dec 7, 2015 at 12:15 AM, Jianfeng (Jeff) Zhang <
jzhang@hortonworks.com> wrote:

>
> We have TEZ-0.7.0 release around half an year ago. During the last 6
> months, we got lots of feedback from tez users, and fix many bugs. I am
> thinking this is a good time for the 0.7.1 release now.  And for this
> release, we would release binary version of tez for different hadoop
> versions for the first time. The candidate hadoop version in my mind is
> 2.2/2.4/2.6/2.7
>
> I have go through tickets targeted for 0.7.1 and move out some tickets to
> 0.7.2 which I think is not critical.  And there's still 21 tickets left,
> you can check them on the following link, if you feel that's critical for
> you, please reply this thread and mark the ticket as blocker of 0.7.1
>
> https://issues.apache.org/jira/issues/?jql=project%20%3D%20TEZ%20AND%20%22Target%20Version%2Fs%22%20%3D%200.7.1%20AND%20resolution%20%3D%20Unresolved%20ORDER%20BY%20due%20ASC%2C%20priority%20DESC%2C%20created%20ASC
>
>
>
> Best Regard,
> Jeff Zhang
>
>