As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was  
released 3 months ago and appears to be stable.  It also seems to be  
considered a more solid release than 0.19 (at least by the Hadoop  
developers and users I've talked to).  Given this, I propose that for  
our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3  
branch will continue to use Hadoop 0.18, and any major bug fixes in  
Pig could be subsequently released as 0.3.x releases.

Thoughts?

Alan.

Will this be a compatibility release or will this also include a move to
the new APIs?

Santhosh

-----Original Message-----
From: Alan Gates [mailto:[EMAIL PROTECTED]]
Sent: Tuesday, July 07, 2009 4:18 PM
To: [EMAIL PROTECTED]
Subject: Moving to Hadoop 0.20

As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was  
released 3 months ago and appears to be stable.  It also seems to be  
considered a more solid release than 0.19 (at least by the Hadoop  
developers and users I've talked to).  Given this, I propose that for  
our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3  
branch will continue to use Hadoop 0.18, and any major bug fixes in  
Pig could be subsequently released as 0.3.x releases.

Thoughts?

Alan.

I'd like to move to Hadoop 0.20 sooner rather than later.  Given that  
a patch exists to move in compatibility mode (https://issues.apache.org/jira/browse/PIG-660
) I propose we do that first.  We can convert to the new APIs as needed.

Alan.

On Jul 7, 2009, at 4:21 PM, Santhosh Srinivasan wrote:

> Will this be a compatibility release or will this also include a  
> move to
> the new APIs?
>
> Santhosh
>
> -----Original Message-----
> From: Alan Gates [mailto:[EMAIL PROTECTED]]
> Sent: Tuesday, July 07, 2009 4:18 PM
> To: [EMAIL PROTECTED]
> Subject: Moving to Hadoop 0.20
>
> As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was
> released 3 months ago and appears to be stable.  It also seems to be
> considered a more solid release than 0.19 (at least by the Hadoop
> developers and users I've talked to).  Given this, I propose that for
> our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3
> branch will continue to use Hadoop 0.18, and any major bug fixes in
> Pig could be subsequently released as 0.3.x releases.
>
> Thoughts?
>
> Alan.

Alan,
A lot of people are still running 18 in production. Unless a
backwards-compatibility patch is provided (or, better yet, a config
parameter that flips between 18 and 20), a change to 20 would mean that
those folks can't use new versions of Pig. Given that a patch exists for
those who want to be on the bleeding edge, I think this move would be
premature for 0.4 -- we should wait for wider adoption of 0.20.

IIRC, Owen said at the Summit that even Y! production clusters run 18?

-Dmitriy

On Fri, Jul 10, 2009 at 10:17 AM, Alan Gates <[EMAIL PROTECTED]> wrote:

> I'd like to move to Hadoop 0.20 sooner rather than later.  Given that a
> patch exists to move in compatibility mode (
> https://issues.apache.org/jira/browse/PIG-660) I propose we do that first.
>  We can convert to the new APIs as needed.
>
> Alan.
>
>
> On Jul 7, 2009, at 4:21 PM, Santhosh Srinivasan wrote:
>
>  Will this be a compatibility release or will this also include a move to
>> the new APIs?
>>
>> Santhosh
>>
>> -----Original Message-----
>> From: Alan Gates [mailto:[EMAIL PROTECTED]]
>> Sent: Tuesday, July 07, 2009 4:18 PM
>> To: [EMAIL PROTECTED]
>> Subject: Moving to Hadoop 0.20
>>
>> As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was
>> released 3 months ago and appears to be stable.  It also seems to be
>> considered a more solid release than 0.19 (at least by the Hadoop
>> developers and users I've talked to).  Given this, I propose that for
>> our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3
>> branch will continue to use Hadoop 0.18, and any major bug fixes in
>> Pig could be subsequently released as 0.3.x releases.
>>
>> Thoughts?
>>
>> Alan.
>>
>
>

I say +1 to making 0.4 built against hadoop 0.20, and then making 0.3.x
releases against hadoop 0.18 as necessary.

What is the current expectation for the 0.4 release, btw?

Thanks,
Kevin

On Fri, Jul 10, 2009 at 10:42 AM, Dmitriy Ryaboy <[EMAIL PROTECTED]>wrote:

> Alan,
> A lot of people are still running 18 in production. Unless a
> backwards-compatibility patch is provided (or, better yet, a config
> parameter that flips between 18 and 20), a change to 20 would mean that
> those folks can't use new versions of Pig. Given that a patch exists for
> those who want to be on the bleeding edge, I think this move would be
> premature for 0.4 -- we should wait for wider adoption of 0.20.
>
> IIRC, Owen said at the Summit that even Y! production clusters run 18?
>
> -Dmitriy
>
> On Fri, Jul 10, 2009 at 10:17 AM, Alan Gates <[EMAIL PROTECTED]> wrote:
>
> > I'd like to move to Hadoop 0.20 sooner rather than later.  Given that a
> > patch exists to move in compatibility mode (
> > https://issues.apache.org/jira/browse/PIG-660) I propose we do that
> first.
> >  We can convert to the new APIs as needed.
> >
> > Alan.
> >
> >
> > On Jul 7, 2009, at 4:21 PM, Santhosh Srinivasan wrote:
> >
> >  Will this be a compatibility release or will this also include a move to
> >> the new APIs?
> >>
> >> Santhosh
> >>
> >> -----Original Message-----
> >> From: Alan Gates [mailto:[EMAIL PROTECTED]]
> >> Sent: Tuesday, July 07, 2009 4:18 PM
> >> To: [EMAIL PROTECTED]
> >> Subject: Moving to Hadoop 0.20
> >>
> >> As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was
> >> released 3 months ago and appears to be stable.  It also seems to be
> >> considered a more solid release than 0.19 (at least by the Hadoop
> >> developers and users I've talked to).  Given this, I propose that for
> >> our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3
> >> branch will continue to use Hadoop 0.18, and any major bug fixes in
> >> Pig could be subsequently released as 0.3.x releases.
> >>
> >> Thoughts?
> >>
> >> Alan.
> >>
> >
> >
>

+1 to Pig 0.40 on Hadoop 0.20
On 7/25/09, Kevin Weil <[EMAIL PROTECTED]> wrote:
> I say +1 to making 0.4 built against hadoop 0.20, and then making 0.3.x
> releases against hadoop 0.18 as necessary.
>
> What is the current expectation for the 0.4 release, btw?
>
> Thanks,
> Kevin
>
> On Fri, Jul 10, 2009 at 10:42 AM, Dmitriy Ryaboy
> <[EMAIL PROTECTED]>wrote:
>
>> Alan,
>> A lot of people are still running 18 in production. Unless a
>> backwards-compatibility patch is provided (or, better yet, a config
>> parameter that flips between 18 and 20), a change to 20 would mean that
>> those folks can't use new versions of Pig. Given that a patch exists for
>> those who want to be on the bleeding edge, I think this move would be
>> premature for 0.4 -- we should wait for wider adoption of 0.20.
>>
>> IIRC, Owen said at the Summit that even Y! production clusters run 18?
>>
>> -Dmitriy
>>
>> On Fri, Jul 10, 2009 at 10:17 AM, Alan Gates <[EMAIL PROTECTED]> wrote:
>>
>> > I'd like to move to Hadoop 0.20 sooner rather than later.  Given that a
>> > patch exists to move in compatibility mode (
>> > https://issues.apache.org/jira/browse/PIG-660) I propose we do that
>> first.
>> >  We can convert to the new APIs as needed.
>> >
>> > Alan.
>> >
>> >
>> > On Jul 7, 2009, at 4:21 PM, Santhosh Srinivasan wrote:
>> >
>> >  Will this be a compatibility release or will this also include a move
>> > to
>> >> the new APIs?
>> >>
>> >> Santhosh
>> >>
>> >> -----Original Message-----
>> >> From: Alan Gates [mailto:[EMAIL PROTECTED]]
>> >> Sent: Tuesday, July 07, 2009 4:18 PM
>> >> To: [EMAIL PROTECTED]
>> >> Subject: Moving to Hadoop 0.20
>> >>
>> >> As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was
>> >> released 3 months ago and appears to be stable.  It also seems to be
>> >> considered a more solid release than 0.19 (at least by the Hadoop
>> >> developers and users I've talked to).  Given this, I propose that for
>> >> our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The 0.3
>> >> branch will continue to use Hadoop 0.18, and any major bug fixes in
>> >> Pig could be subsequently released as 0.3.x releases.
>> >>
>> >> Thoughts?
>> >>
>> >> Alan.
>> >>
>> >
>> >
>>
>

We would like to get a couple of features in before the release such as
skew join and merge join. I think we should be able to call vote in the
second half of August.

Olga

-----Original Message-----
From: Amr Awadallah [mailto:[EMAIL PROTECTED]]
Sent: Saturday, July 25, 2009 10:02 AM
To: [EMAIL PROTECTED]
Subject: Re: Moving to Hadoop 0.20

+1 to Pig 0.40 on Hadoop 0.20
On 7/25/09, Kevin Weil <[EMAIL PROTECTED]> wrote:
> I say +1 to making 0.4 built against hadoop 0.20, and then making
0.3.x
> releases against hadoop 0.18 as necessary.
>
> What is the current expectation for the 0.4 release, btw?
>
> Thanks,
> Kevin
>
> On Fri, Jul 10, 2009 at 10:42 AM, Dmitriy Ryaboy
> <[EMAIL PROTECTED]>wrote:
>
>> Alan,
>> A lot of people are still running 18 in production. Unless a
>> backwards-compatibility patch is provided (or, better yet, a config
>> parameter that flips between 18 and 20), a change to 20 would mean
that
>> those folks can't use new versions of Pig. Given that a patch exists
for
>> those who want to be on the bleeding edge, I think this move would be
>> premature for 0.4 -- we should wait for wider adoption of 0.20.
>>
>> IIRC, Owen said at the Summit that even Y! production clusters run
18?
>>
>> -Dmitriy
>>
>> On Fri, Jul 10, 2009 at 10:17 AM, Alan Gates <[EMAIL PROTECTED]>
wrote:
>>
>> > I'd like to move to Hadoop 0.20 sooner rather than later.  Given
that a
>> > patch exists to move in compatibility mode (
>> > https://issues.apache.org/jira/browse/PIG-660) I propose we do that
>> first.
>> >  We can convert to the new APIs as needed.
>> >
>> > Alan.
>> >
>> >
>> > On Jul 7, 2009, at 4:21 PM, Santhosh Srinivasan wrote:
>> >
>> >  Will this be a compatibility release or will this also include a
move
>> > to
>> >> the new APIs?
>> >>
>> >> Santhosh
>> >>
>> >> -----Original Message-----
>> >> From: Alan Gates [mailto:[EMAIL PROTECTED]]
>> >> Sent: Tuesday, July 07, 2009 4:18 PM
>> >> To: [EMAIL PROTECTED]
>> >> Subject: Moving to Hadoop 0.20
>> >>
>> >> As of release 0.3.0, Pig is still on Hadoop 0.18.  Hadoop 0.20 was
>> >> released 3 months ago and appears to be stable.  It also seems to
be
>> >> considered a more solid release than 0.19 (at least by the Hadoop
>> >> developers and users I've talked to).  Given this, I propose that
for
>> >> our next release (assumably 0.4.0) Pig move to Hadoop 0.20.   The
0.3
>> >> branch will continue to use Hadoop 0.18, and any major bug fixes
in
>> >> Pig could be subsequently released as 0.3.x releases.
>> >>
>> >> Thoughts?
>> >>
>> >> Alan.
>> >>
>> >
>> >
>>
>