Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Pig >> mail # dev >> pig 0.11 candidate 2 feedback: Several problems


Copy link to this message
-
Re: pig 0.11 candidate 2 feedback: Several problems
I stand corrected. Cool, 0.11 is good!
On Wed, Feb 20, 2013 at 1:15 PM, Jarek Jarcec Cecho <[EMAIL PROTECTED]>wrote:

> Just a unrelated note: The CDH3 is more closer to Hadoop 1.x than to 0.20.
>
> Jarcec
>
> On Wed, Feb 20, 2013 at 12:04:51PM -0800, Dmitriy Ryaboy wrote:
> > I agree -- this is a good release. The bugs Kai pointed out should be
> > fixed, but as they are not critical regressions, we can fix them in
> 0.11.1
> > (if someone wants to roll 0.11.1 the minute these fixes are committed, I
> > won't mind and will dutifully vote for the release).
> >
> > I think the Hadoop 20.2 incompatibility is unfortunate but iirc this is
> > fixable by setting HADOOP_USER_CLASSPATH_FIRST=true (was that in 20.2?)
> >
> > FWIW Twitter's running CDH3 and this release works in our environment.
> >
> > At this point things that block a release are critical regressions in
> > performance or correctness.
> >
> > D
> >
> >
> > On Wed, Feb 20, 2013 at 11:52 AM, Alan Gates <[EMAIL PROTECTED]>
> wrote:
> >
> > > No.  Bugs like these are supposed to be found and fixed after we branch
> > > from trunk (which happened several months ago in the case of 0.11).
>  The
> > > point of RCs are to check that it's a good build, licenses are right,
> etc.
> > >  Any bugs found this late in the game have to be seen as failures of
> > > earlier testing.
> > >
> > > Alan.
> > >
> > > On Feb 20, 2013, at 11:33 AM, Russell Jurney wrote:
> > >
> > > > Isn't the point of an RC to find and fix bugs like these>
> > > >
> > > >
> > > > On Wed, Feb 20, 2013 at 11:31 AM, Bill Graham <[EMAIL PROTECTED]>
> > > wrote:
> > > >
> > > >> Regarding Pig 11 rc2, I propose we continue with the current vote
> as is
> > > >> (which closes today EOD). Patches for 0.20.2 issues can be rolled
> into a
> > > >> Pig 0.11.1 release whenever they're available and tested.
> > > >>
> > > >>
> > > >>
> > > >> On Wed, Feb 20, 2013 at 9:24 AM, Olga Natkovich <
> [EMAIL PROTECTED]
> > > >>> wrote:
> > > >>
> > > >>> I agree that supporting as much as we can is a good goal. The
> issue is
> > > >> who
> > > >>> is going to be testing against all these versions? We found the
> issues
> > > >>> under discussion because of a customer report, not because we
> > > >> consistently
> > > >>> test against all versions. Perhaps when we decide which versions to
> > > >> support
> > > >>> for next release we need also to agree who is going to be testing
> and
> > > >>> maintaining compatibility with a particular version.
> > > >>>
> > > >>> For instance since Hadoop 23 compatibility is important for us at
> Yahoo
> > > >> we
> > > >>> have been maintaining compatibility with this version for 0.9,
> 0.10 and
> > > >>> will do the same for 0.11 and going forward. I think we would need
> > > others
> > > >>> to step in and claim the versions of their interest.
> > > >>>
> > > >>> Olga
> > > >>>
> > > >>>
> > > >>> ________________________________
> > > >>> From: Kai Londenberg <[EMAIL PROTECTED]>
> > > >>> To: [EMAIL PROTECTED]
> > > >>> Sent: Wednesday, February 20, 2013 1:51 AM
> > > >>> Subject: Re: pig 0.11 candidate 2 feedback: Several problems
> > > >>>
> > > >>> Hi,
> > > >>>
> > > >>> I stronly agree with Jonathan here. If there are good reasons why
> you
> > > >>> can't support an older version of Hadoop any more, that's one
> thing.
> > > >>> But having to change 2 lines of code doesn't really qualify as
> such in
> > > >>> my point of view ;)
> > > >>>
> > > >>> At least for me, pig support for 0.20.2 is essential - without it,
> I
> > > >>> can't use it. If it doesn't support it, I'll have to branch pig and
> > > >>> hack it myself, or stop using it.
> > > >>>
> > > >>> I guess, there are a lot of people still running 0.20.2 Clusters.
> If
> > > >>> you really have lots of data stored on HDFS and a continuously busy
> > > >>> cluster, an upgrade is nothing you do "just because".
> > > >>>
> > > >>>
> > > >>> 2013/2/20 Jonathan Coveney <[EMAIL PROTECTED]>:
> > > >>>> I agree that we shouldn't have to support old versions forever.

Russell Jurney twitter.com/rjurney [EMAIL PROTECTED] datasyndrome.com