You are viewing a plain text version of this content. The canonical link for it is here.

Posted to user@hbase.apache.org by Nick Dimiduk <nd...@apache.org> on 2015/03/25 23:05:20 UTC

Spinning up for 1.1 Release

With about 100 issues beyond the 1.0.x line [0], I think it's time to start
talking about our 1.1.0 release. Noteworthy goodness already committed to
the branch includes:

 - Async RPC client (HBASE-12684)
 - Improved compaction controls (HBASE-8329, HBASE-12859)
 - Per-column family flush (HBASE-10201)
 - WAL on SSD (HBASE-12848)
 - Tons of region replica enhancements around META, WAL, and bulk loading
(HBASE-11574, HBASE-11568, HBASE-11571, HBASE-11567)

There's also talk of porting over the MOB goodness (HBASE-11339), simple
RPC throttling (HBASE-11598), and bits from the new procedure bus
(HBASE-12439). Remember, this is a "minor release" [1] according to our
newly adopted semantic versioning guide [2], which means if you want new
features in, you've got to speak up now or wait for 1.2.

So timeline. Given HBaseCon [3] is right around the corner, I'd love to see
the 1.1.0 bits on every attendee's laptops and fobs. To make that happen, I
figure we'll need to start voting on RC's around April 24, give me time to
work out this whole RM thing and give you time to test 'em! Working back
from there, we should have the major feature landed and the release branch
cut on or around April 15. Are you up for this schedule?

For anyone who's bored, there's a couple unresolved blockers for this
release. I'll also take your help in monitoring the jenkins builds against
branch-1.

Thanks for your help in making this next release another great one!

Cheers,
-n

[0]: http://s.apache.org/hbase-1.1-issues
[1]: http://hbase.apache.org/book.html#hbase.versioning.post10
[2]: http://semver.org/
[3]: http://hbasecon.com/

Re: Spinning up for 1.1 Release

Posted by Nick Dimiduk <nd...@gmail.com>.

On Mon, Apr 27, 2015 at 10:11 AM, Elliott Clark <ec...@apache.org> wrote:

> Any thoughts on HBASE-13517 ? It has no extra code changes so it shouldn't
> be de-stabilizing.
>

I'm inclined to include it -- someone else have a strong argument against?
Any suggestions on how we can test it properly? A new integration test
maybe?

On Mon, Apr 27, 2015 at 9:32 AM, Cody Marcel <cm...@salesforce.com> wrote:
>
> > The port for PHOENIX-1728
> > <https://issues.apache.org/jira/browse/PHOENIX-1728> from master is
> still
> > in progress. I should have the rest of the branches update today.
> >
> > On Sun, Apr 26, 2015 at 11:57 AM, Nick Dimiduk <nd...@gmail.com>
> wrote:
> >
> > > Devs --
> > >
> > > Branch-1.1 is officially closed for feature enhancements, with the
> > > following exceptions: HBASE-11339, HBASE-13260, HBASE-13431.
> > >
> > > Please limit commits to blocker and critical bug fixes. I also ask that
> > you
> > > please ping me as patches go in, so that I might review them in
> relation
> > to
> > > the testing that's been done. I do not require you wait for a +1 from
> me
> > on
> > > these issues, simply a courtesy notification. Please use your best
> > > judgement re: the above request.
> > >
> > > Thanks,
> > > Nick
> > >
> > > On Thu, Apr 23, 2015 at 11:14 AM, Enis Söztutar <en...@gmail.com>
> > > wrote:
> > >
> > > > No.
> > > >
> > > > semver[1] has an explicit policy to have identifiers in this format:
> > > >
> > > > MAJOR.MINOR.PATCH[-IDENTIFIER]
> > > >
> > > > From [1]:
> > > >
> > > > Additional labels for pre-release and build metadata are available as
> > > > extensions to the MAJOR.MINOR.PATCH format.
> > > > Bullet point 9 in [1] talks about this. If we want to do "developer
> > > > preview" releases, we should use 1.1.0-alpha, 1.1.0-beta kind of
> > version
> > > > name.
> > > >
> > > > [1] http://semver.org/
> > > >
> > > > Enis
> > > >
> > > > On Thu, Apr 23, 2015 at 5:48 AM, Jean-Marc Spaggiari <
> > > > jean-marc@spaggiari.org> wrote:
> > > >
> > > > > Sorry guys,
> > > > >
> > > > > I'm know I'm very late in the process, but should we not release
> > 1.2.0
> > > > > instead of 1.1.0?
> > > > >
> > > > > 0.97 was for dev
> > > > > 0.99 was for dev
> > > > > Should 1.1 be for dev too? Should we keep even numbers for prod
> > > releases
> > > > > and odd for dev?
> > > > >
> > > > > JM
> > > > >
> > > > > 2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:
> > > > >
> > > > > > I would love to see
> > > https://issues.apache.org/jira/browse/HBASE-13260
> > > > in
> > > > > > if
> > > > > > possible.
> > > > > >
> > > > > > Enis
> > > > > >
> > > > > > On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> > > > > > chrajeshbabu32@gmail.com> wrote:
> > > > > >
> > > > > > > I would like to have it
> > > > > > https://issues.apache.org/jira/browse/HBASE-13431
> > > > > > > in.
> > > > > > >
> > > > > > > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com>
> > > > wrote:
> > > > > > >
> > > > > > > > You have my attention all the time, Nick :-)
> > > > > > > > We need the folks to get an agreement on HBASE-13149.  Sean,
> > > Andy,
> > > > > > Enis?
> > > > > > > > Of course the last resort is we do nothing but document it
> with
> > > > > manual
> > > > > > > > workaround or manual jar replacement in 1.1.
> > > > > > > > I am available to do any verification if there is any
> proposal.
> > > > > > > >
> > > > > > > > Jerry
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Elliott Clark <ec...@apache.org>.

Any thoughts on HBASE-13517 ? It has no extra code changes so it shouldn't
be de-stabilizing.

On Mon, Apr 27, 2015 at 9:32 AM, Cody Marcel <cm...@salesforce.com> wrote:

> The port for PHOENIX-1728
> <https://issues.apache.org/jira/browse/PHOENIX-1728> from master is still
> in progress. I should have the rest of the branches update today.
>
> On Sun, Apr 26, 2015 at 11:57 AM, Nick Dimiduk <nd...@gmail.com> wrote:
>
> > Devs --
> >
> > Branch-1.1 is officially closed for feature enhancements, with the
> > following exceptions: HBASE-11339, HBASE-13260, HBASE-13431.
> >
> > Please limit commits to blocker and critical bug fixes. I also ask that
> you
> > please ping me as patches go in, so that I might review them in relation
> to
> > the testing that's been done. I do not require you wait for a +1 from me
> on
> > these issues, simply a courtesy notification. Please use your best
> > judgement re: the above request.
> >
> > Thanks,
> > Nick
> >
> > On Thu, Apr 23, 2015 at 11:14 AM, Enis Söztutar <en...@gmail.com>
> > wrote:
> >
> > > No.
> > >
> > > semver[1] has an explicit policy to have identifiers in this format:
> > >
> > > MAJOR.MINOR.PATCH[-IDENTIFIER]
> > >
> > > From [1]:
> > >
> > > Additional labels for pre-release and build metadata are available as
> > > extensions to the MAJOR.MINOR.PATCH format.
> > > Bullet point 9 in [1] talks about this. If we want to do "developer
> > > preview" releases, we should use 1.1.0-alpha, 1.1.0-beta kind of
> version
> > > name.
> > >
> > > [1] http://semver.org/
> > >
> > > Enis
> > >
> > > On Thu, Apr 23, 2015 at 5:48 AM, Jean-Marc Spaggiari <
> > > jean-marc@spaggiari.org> wrote:
> > >
> > > > Sorry guys,
> > > >
> > > > I'm know I'm very late in the process, but should we not release
> 1.2.0
> > > > instead of 1.1.0?
> > > >
> > > > 0.97 was for dev
> > > > 0.99 was for dev
> > > > Should 1.1 be for dev too? Should we keep even numbers for prod
> > releases
> > > > and odd for dev?
> > > >
> > > > JM
> > > >
> > > > 2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:
> > > >
> > > > > I would love to see
> > https://issues.apache.org/jira/browse/HBASE-13260
> > > in
> > > > > if
> > > > > possible.
> > > > >
> > > > > Enis
> > > > >
> > > > > On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> > > > > chrajeshbabu32@gmail.com> wrote:
> > > > >
> > > > > > I would like to have it
> > > > > https://issues.apache.org/jira/browse/HBASE-13431
> > > > > > in.
> > > > > >
> > > > > > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com>
> > > wrote:
> > > > > >
> > > > > > > You have my attention all the time, Nick :-)
> > > > > > > We need the folks to get an agreement on HBASE-13149.  Sean,
> > Andy,
> > > > > Enis?
> > > > > > > Of course the last resort is we do nothing but document it with
> > > > manual
> > > > > > > workaround or manual jar replacement in 1.1.
> > > > > > > I am available to do any verification if there is any proposal.
> > > > > > >
> > > > > > > Jerry
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Cody Marcel <cm...@salesforce.com>.

The port for PHOENIX-1728
<https://issues.apache.org/jira/browse/PHOENIX-1728> from master is still
in progress. I should have the rest of the branches update today.

On Sun, Apr 26, 2015 at 11:57 AM, Nick Dimiduk <nd...@gmail.com> wrote:

> Devs --
>
> Branch-1.1 is officially closed for feature enhancements, with the
> following exceptions: HBASE-11339, HBASE-13260, HBASE-13431.
>
> Please limit commits to blocker and critical bug fixes. I also ask that you
> please ping me as patches go in, so that I might review them in relation to
> the testing that's been done. I do not require you wait for a +1 from me on
> these issues, simply a courtesy notification. Please use your best
> judgement re: the above request.
>
> Thanks,
> Nick
>
> On Thu, Apr 23, 2015 at 11:14 AM, Enis Söztutar <en...@gmail.com>
> wrote:
>
> > No.
> >
> > semver[1] has an explicit policy to have identifiers in this format:
> >
> > MAJOR.MINOR.PATCH[-IDENTIFIER]
> >
> > From [1]:
> >
> > Additional labels for pre-release and build metadata are available as
> > extensions to the MAJOR.MINOR.PATCH format.
> > Bullet point 9 in [1] talks about this. If we want to do "developer
> > preview" releases, we should use 1.1.0-alpha, 1.1.0-beta kind of version
> > name.
> >
> > [1] http://semver.org/
> >
> > Enis
> >
> > On Thu, Apr 23, 2015 at 5:48 AM, Jean-Marc Spaggiari <
> > jean-marc@spaggiari.org> wrote:
> >
> > > Sorry guys,
> > >
> > > I'm know I'm very late in the process, but should we not release 1.2.0
> > > instead of 1.1.0?
> > >
> > > 0.97 was for dev
> > > 0.99 was for dev
> > > Should 1.1 be for dev too? Should we keep even numbers for prod
> releases
> > > and odd for dev?
> > >
> > > JM
> > >
> > > 2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:
> > >
> > > > I would love to see
> https://issues.apache.org/jira/browse/HBASE-13260
> > in
> > > > if
> > > > possible.
> > > >
> > > > Enis
> > > >
> > > > On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> > > > chrajeshbabu32@gmail.com> wrote:
> > > >
> > > > > I would like to have it
> > > > https://issues.apache.org/jira/browse/HBASE-13431
> > > > > in.
> > > > >
> > > > > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com>
> > wrote:
> > > > >
> > > > > > You have my attention all the time, Nick :-)
> > > > > > We need the folks to get an agreement on HBASE-13149.  Sean,
> Andy,
> > > > Enis?
> > > > > > Of course the last resort is we do nothing but document it with
> > > manual
> > > > > > workaround or manual jar replacement in 1.1.
> > > > > > I am available to do any verification if there is any proposal.
> > > > > >
> > > > > > Jerry
> > > > > >
> > > > >
> > > >
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Nick Dimiduk <nd...@gmail.com>.

Devs --

Branch-1.1 is officially closed for feature enhancements, with the
following exceptions: HBASE-11339, HBASE-13260, HBASE-13431.

Please limit commits to blocker and critical bug fixes. I also ask that you
please ping me as patches go in, so that I might review them in relation to
the testing that's been done. I do not require you wait for a +1 from me on
these issues, simply a courtesy notification. Please use your best
judgement re: the above request.

Thanks,
Nick

On Thu, Apr 23, 2015 at 11:14 AM, Enis Söztutar <en...@gmail.com> wrote:

> No.
>
> semver[1] has an explicit policy to have identifiers in this format:
>
> MAJOR.MINOR.PATCH[-IDENTIFIER]
>
> From [1]:
>
> Additional labels for pre-release and build metadata are available as
> extensions to the MAJOR.MINOR.PATCH format.
> Bullet point 9 in [1] talks about this. If we want to do "developer
> preview" releases, we should use 1.1.0-alpha, 1.1.0-beta kind of version
> name.
>
> [1] http://semver.org/
>
> Enis
>
> On Thu, Apr 23, 2015 at 5:48 AM, Jean-Marc Spaggiari <
> jean-marc@spaggiari.org> wrote:
>
> > Sorry guys,
> >
> > I'm know I'm very late in the process, but should we not release 1.2.0
> > instead of 1.1.0?
> >
> > 0.97 was for dev
> > 0.99 was for dev
> > Should 1.1 be for dev too? Should we keep even numbers for prod releases
> > and odd for dev?
> >
> > JM
> >
> > 2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:
> >
> > > I would love to see https://issues.apache.org/jira/browse/HBASE-13260
> in
> > > if
> > > possible.
> > >
> > > Enis
> > >
> > > On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> > > chrajeshbabu32@gmail.com> wrote:
> > >
> > > > I would like to have it
> > > https://issues.apache.org/jira/browse/HBASE-13431
> > > > in.
> > > >
> > > > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com>
> wrote:
> > > >
> > > > > You have my attention all the time, Nick :-)
> > > > > We need the folks to get an agreement on HBASE-13149.  Sean, Andy,
> > > Enis?
> > > > > Of course the last resort is we do nothing but document it with
> > manual
> > > > > workaround or manual jar replacement in 1.1.
> > > > > I am available to do any verification if there is any proposal.
> > > > >
> > > > > Jerry
> > > > >
> > > >
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Enis Söztutar <en...@gmail.com>.

No.

semver[1] has an explicit policy to have identifiers in this format:

MAJOR.MINOR.PATCH[-IDENTIFIER]

>From [1]:

Additional labels for pre-release and build metadata are available as
extensions to the MAJOR.MINOR.PATCH format.
Bullet point 9 in [1] talks about this. If we want to do "developer
preview" releases, we should use 1.1.0-alpha, 1.1.0-beta kind of version
name.

[1] http://semver.org/

Enis

On Thu, Apr 23, 2015 at 5:48 AM, Jean-Marc Spaggiari <
jean-marc@spaggiari.org> wrote:

> Sorry guys,
>
> I'm know I'm very late in the process, but should we not release 1.2.0
> instead of 1.1.0?
>
> 0.97 was for dev
> 0.99 was for dev
> Should 1.1 be for dev too? Should we keep even numbers for prod releases
> and odd for dev?
>
> JM
>
> 2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:
>
> > I would love to see https://issues.apache.org/jira/browse/HBASE-13260 in
> > if
> > possible.
> >
> > Enis
> >
> > On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> > chrajeshbabu32@gmail.com> wrote:
> >
> > > I would like to have it
> > https://issues.apache.org/jira/browse/HBASE-13431
> > > in.
> > >
> > > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com> wrote:
> > >
> > > > You have my attention all the time, Nick :-)
> > > > We need the folks to get an agreement on HBASE-13149.  Sean, Andy,
> > Enis?
> > > > Of course the last resort is we do nothing but document it with
> manual
> > > > workaround or manual jar replacement in 1.1.
> > > > I am available to do any verification if there is any proposal.
> > > >
> > > > Jerry
> > > >
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Jean-Marc Spaggiari <je...@spaggiari.org>.

Sorry guys,

I'm know I'm very late in the process, but should we not release 1.2.0
instead of 1.1.0?

0.97 was for dev
0.99 was for dev
Should 1.1 be for dev too? Should we keep even numbers for prod releases
and odd for dev?

JM

2015-04-22 17:32 GMT-04:00 Enis Söztutar <en...@apache.org>:

> I would love to see https://issues.apache.org/jira/browse/HBASE-13260 in
> if
> possible.
>
> Enis
>
> On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
> chrajeshbabu32@gmail.com> wrote:
>
> > I would like to have it
> https://issues.apache.org/jira/browse/HBASE-13431
> > in.
> >
> > On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com> wrote:
> >
> > > You have my attention all the time, Nick :-)
> > > We need the folks to get an agreement on HBASE-13149.  Sean, Andy,
> Enis?
> > > Of course the last resort is we do nothing but document it with manual
> > > workaround or manual jar replacement in 1.1.
> > > I am available to do any verification if there is any proposal.
> > >
> > > Jerry
> > >
> >
>

Re: Spinning up for 1.1 Release

Posted by Enis Söztutar <en...@apache.org>.

I would love to see https://issues.apache.org/jira/browse/HBASE-13260 in if
possible.

Enis

On Wed, Apr 22, 2015 at 12:49 PM, rajeshbabu@apache.org <
chrajeshbabu32@gmail.com> wrote:

> I would like to have it https://issues.apache.org/jira/browse/HBASE-13431
> in.
>
> On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com> wrote:
>
> > You have my attention all the time, Nick :-)
> > We need the folks to get an agreement on HBASE-13149.  Sean, Andy, Enis?
> > Of course the last resort is we do nothing but document it with manual
> > workaround or manual jar replacement in 1.1.
> > I am available to do any verification if there is any proposal.
> >
> > Jerry
> >
>

Re: Spinning up for 1.1 Release

Posted by "rajeshbabu@apache.org" <ch...@gmail.com>.

I would like to have it https://issues.apache.org/jira/browse/HBASE-13431
in.

On Thu, Apr 23, 2015 at 1:10 AM, Jerry He <je...@gmail.com> wrote:

> You have my attention all the time, Nick :-)
> We need the folks to get an agreement on HBASE-13149.  Sean, Andy, Enis?
> Of course the last resort is we do nothing but document it with manual
> workaround or manual jar replacement in 1.1.
> I am available to do any verification if there is any proposal.
>
> Jerry
>

Re: Spinning up for 1.1 Release

Posted by Jerry He <je...@gmail.com>.

You have my attention all the time, Nick :-)
We need the folks to get an agreement on HBASE-13149.  Sean, Andy, Enis?
Of course the last resort is we do nothing but document it with manual
workaround or manual jar replacement in 1.1.
I am available to do any verification if there is any proposal.

Jerry

Re: Spinning up for 1.1 Release

Posted by Lars Francke <la...@gmail.com>.

I'd love to have <https://issues.apache.org/jira/browse/HBASE-13466> in if
no one objects.

On Wed, Apr 22, 2015 at 7:51 PM, Nick Dimiduk <nd...@gmail.com> wrote:

> My mistake Jerry, you're correct. My pre-coffee brain was thinking that by
> upgrading Hadoop we'd promote the version of Jackson under us. This is not
> the case because we have an explicit dependency (though maybe maven is
> promoting this for us for the builds anyway?)
>
> Point is, we have issues that need attention before the release. I'm glad
> to see I have your attention ;)
>
> On Wed, Apr 22, 2015 at 10:43 AM, Jerry He <je...@gmail.com> wrote:
>
> > Hi, Nick
> >
> > I probably missed something here.  Upgrading to hadoop 2.6 would help
> > HBASE-13149?
> > >> HBASE-13149 and HBASE-13339 are making a strong argument for us to
> bump
> > >> the Hadoop version to 2.6 for this release.
> >
> > Jerry
> >
>

Re: Spinning up for 1.1 Release

Posted by Nick Dimiduk <nd...@gmail.com>.

My mistake Jerry, you're correct. My pre-coffee brain was thinking that by
upgrading Hadoop we'd promote the version of Jackson under us. This is not
the case because we have an explicit dependency (though maybe maven is
promoting this for us for the builds anyway?)

Point is, we have issues that need attention before the release. I'm glad
to see I have your attention ;)

On Wed, Apr 22, 2015 at 10:43 AM, Jerry He <je...@gmail.com> wrote:

> Hi, Nick
>
> I probably missed something here.  Upgrading to hadoop 2.6 would help
> HBASE-13149?
> >> HBASE-13149 and HBASE-13339 are making a strong argument for us to bump
> >> the Hadoop version to 2.6 for this release.
>
> Jerry
>

Re: Spinning up for 1.1 Release

Posted by Jerry He <je...@gmail.com>.

Hi, Nick

I probably missed something here.  Upgrading to hadoop 2.6 would help
HBASE-13149?
>> HBASE-13149 and HBASE-13339 are making a strong argument for us to bump
>> the Hadoop version to 2.6 for this release.

Jerry