You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@kudu.apache.org by Jean-Daniel Cryans <jd...@apache.org> on 2016/02/26 23:38:37 UTC
Thinking of doing a 0.7.1
Hello Kudu devs,
Now that 0.7.0 is out of the way, we can think about what's next.
During our testing of this last version we found some non-critical issues
that I think may be good bringing to folks earlier than our next planned
major release. This is my short list of commits:
KUDU-1325: more crash avoidance during remote bootstrap and tablet deletion
KUDU-1343. [java client] Scan sequence id is never reset between tablets
KUDU-1345. Fix case in which hybrid clock can run backwards
Also, it'd be nice to fix our small license issues:
https://issues.apache.org/jira/browse/KUDU-1347
Anything else you good folks would like to see in there? FWIW I don't want
too much bloat, just what's really useful. Following the plan from my "1.0"
email, a 0.8.0 would come out in April anyways.
As for a timeline, I'm hoping to get it all together by the end of next
week.
Thanks,
J-D
Re: Thinking of doing a 0.7.1
Posted by Jean-Daniel Cryans <jd...@apache.org>.
Think we've got everything. Tagging and building the RC.
J-D
On Wed, Mar 2, 2016 at 2:37 PM, Todd Lipcon <to...@cloudera.com> wrote:
> http://gerrit.cloudera.org:8080/#/c/2417/1
>
> Hopefully I navigated the complexity of these underspecified policies
> correctly :)
>
> -Todd
>
>
> On Wed, Mar 2, 2016 at 2:28 PM, Jean-Daniel Cryans <jd...@apache.org>
> wrote:
>
> > Right, I had just copied what we were doing for the other branches.
> Honest
> > mistake, I swear!
> >
> > Also, the branch is almost ready, just need the last patch from Todd for
> > license tweaking and my patch to remove -SNAPSHOT is already up for
> review.
> >
> > J-D
> >
> > On Wed, Mar 2, 2016 at 12:18 PM, Todd Lipcon <to...@cloudera.com> wrote:
> >
> > > On Wed, Mar 2, 2016 at 12:10 PM, Adar Dembo <ad...@cloudera.com> wrote:
> > >
> > > > Ahh, I see. Perhaps we can name our future branches after y releases
> > and
> > > > not z releases?
> > > >
> > > > The only reason to have a branch for each z release is if we need to
> > > patch
> > > > z release n after z release n+1 has shipped.
> > > >
> > >
> > > +1, I dont think it was a particulraly intentional decision :)
> > >
> > > -Todd
> > >
> > >
> > > >
> > > > On Wed, Mar 2, 2016 at 6:30 AM, Jean-Daniel Cryans <
> > jdcryans@apache.org>
> > > > wrote:
> > > >
> > > > > We named that branch branch-0.7.0, not 0.7
> > > > >
> > > > > On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com>
> > wrote:
> > > > >
> > > > > > Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release
> be
> > > > based
> > > > > > there?
> > > > > >
> > > > > > On Tuesday, March 1, 2016, Jean-Daniel Cryans <
> jdcryans@apache.org
> > >
> > > > > wrote:
> > > > > >
> > > > > > > I pushed all the backports for review on a new branch called
> > > > > > branch-0.7.x.
> > > > > > >
> > > > > > > There's still the question of what we want to fix from
> KUDU-1347.
> > > > > > >
> > > > > > > J-D
> > > > > > >
> > > > > > > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <
> todd@cloudera.com
> > > > > > > <javascript:;>> wrote:
> > > > > > >
> > > > > > > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> > > > > > jdcryans@apache.org
> > > > > > > <javascript:;>>
> > > > > > > > wrote:
> > > > > > > > > Yeah good point about known issues.
> > > > > > > >
> > > > > > > > Right, I guess I should have mentioned that we should doc the
> > > > issues
> > > > > > > > you mentioned as 'known issues' in 0.7.0, especially if any
> are
> > > > > > > > regressions (I think KUDU-1325 is, but the others aren't)
> > > > > > > >
> > > > > > > > >
> > > > > > > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <
> > > todd@cloudera.com
> > > > > > > <javascript:;>> wrote:
> > > > > > > > >
> > > > > > > > >> +1 from me.
> > > > > > > > >>
> > > > > > > > >> I'll try to wrap up KUDU-1341 as well, but given that
> it's a
> > > bit
> > > > > of
> > > > > > a
> > > > > > > > >> riskier change we'll need some good testing of it. So,
> > > > > > reaslitically,
> > > > > > > > >> it might not make the cut. I wouldn't hold up the release
> > for
> > > > it,
> > > > > > > > >> though we should add it to "known issues".
> > > > > > > > >>
> > > > > > > > >> -Todd
> > > > > > > > >>
> > > > > > > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > > > > > > jdcryans@apache.org <javascript:;>>
> > > > > > > > >> wrote:
> > > > > > > > >> > Hello Kudu devs,
> > > > > > > > >> >
> > > > > > > > >> > Now that 0.7.0 is out of the way, we can think about
> > what's
> > > > > next.
> > > > > > > > >> >
> > > > > > > > >> > During our testing of this last version we found some
> > > > > non-critical
> > > > > > > > issues
> > > > > > > > >> > that I think may be good bringing to folks earlier than
> > our
> > > > next
> > > > > > > > planned
> > > > > > > > >> > major release. This is my short list of commits:
> > > > > > > > >> >
> > > > > > > > >> > KUDU-1325: more crash avoidance during remote bootstrap
> > and
> > > > > tablet
> > > > > > > > >> deletion
> > > > > > > > >> > KUDU-1343. [java client] Scan sequence id is never reset
> > > > between
> > > > > > > > tablets
> > > > > > > > >> > KUDU-1345. Fix case in which hybrid clock can run
> > backwards
> > > > > > > > >> >
> > > > > > > > >> > Also, it'd be nice to fix our small license issues:
> > > > > > > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > > > > > > >> >
> > > > > > > > >> > Anything else you good folks would like to see in there?
> > > FWIW
> > > > I
> > > > > > > don't
> > > > > > > > >> want
> > > > > > > > >> > too much bloat, just what's really useful. Following the
> > > plan
> > > > > from
> > > > > > > my
> > > > > > > > >> "1.0"
> > > > > > > > >> > email, a 0.8.0 would come out in April anyways.
> > > > > > > > >> >
> > > > > > > > >> > As for a timeline, I'm hoping to get it all together by
> > the
> > > > end
> > > > > of
> > > > > > > > next
> > > > > > > > >> > week.
> > > > > > > > >> >
> > > > > > > > >> > Thanks,
> > > > > > > > >> >
> > > > > > > > >> > J-D
> > > > > > > > >>
> > > > > > > > >>
> > > > > > > > >>
> > > > > > > > >> --
> > > > > > > > >> Todd Lipcon
> > > > > > > > >> Software Engineer, Cloudera
> > > > > > > > >>
> > > > > > > >
> > > > > > > >
> > > > > > > >
> > > > > > > > --
> > > > > > > > Todd Lipcon
> > > > > > > > Software Engineer, Cloudera
> > > > > > > >
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> > >
> > >
> > > --
> > > Todd Lipcon
> > > Software Engineer, Cloudera
> > >
> >
>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>
Re: Thinking of doing a 0.7.1
Posted by Todd Lipcon <to...@cloudera.com>.
http://gerrit.cloudera.org:8080/#/c/2417/1
Hopefully I navigated the complexity of these underspecified policies
correctly :)
-Todd
On Wed, Mar 2, 2016 at 2:28 PM, Jean-Daniel Cryans <jd...@apache.org>
wrote:
> Right, I had just copied what we were doing for the other branches. Honest
> mistake, I swear!
>
> Also, the branch is almost ready, just need the last patch from Todd for
> license tweaking and my patch to remove -SNAPSHOT is already up for review.
>
> J-D
>
> On Wed, Mar 2, 2016 at 12:18 PM, Todd Lipcon <to...@cloudera.com> wrote:
>
> > On Wed, Mar 2, 2016 at 12:10 PM, Adar Dembo <ad...@cloudera.com> wrote:
> >
> > > Ahh, I see. Perhaps we can name our future branches after y releases
> and
> > > not z releases?
> > >
> > > The only reason to have a branch for each z release is if we need to
> > patch
> > > z release n after z release n+1 has shipped.
> > >
> >
> > +1, I dont think it was a particulraly intentional decision :)
> >
> > -Todd
> >
> >
> > >
> > > On Wed, Mar 2, 2016 at 6:30 AM, Jean-Daniel Cryans <
> jdcryans@apache.org>
> > > wrote:
> > >
> > > > We named that branch branch-0.7.0, not 0.7
> > > >
> > > > On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com>
> wrote:
> > > >
> > > > > Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be
> > > based
> > > > > there?
> > > > >
> > > > > On Tuesday, March 1, 2016, Jean-Daniel Cryans <jdcryans@apache.org
> >
> > > > wrote:
> > > > >
> > > > > > I pushed all the backports for review on a new branch called
> > > > > branch-0.7.x.
> > > > > >
> > > > > > There's still the question of what we want to fix from KUDU-1347.
> > > > > >
> > > > > > J-D
> > > > > >
> > > > > > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> > > > > > <javascript:;>> wrote:
> > > > > >
> > > > > > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> > > > > jdcryans@apache.org
> > > > > > <javascript:;>>
> > > > > > > wrote:
> > > > > > > > Yeah good point about known issues.
> > > > > > >
> > > > > > > Right, I guess I should have mentioned that we should doc the
> > > issues
> > > > > > > you mentioned as 'known issues' in 0.7.0, especially if any are
> > > > > > > regressions (I think KUDU-1325 is, but the others aren't)
> > > > > > >
> > > > > > > >
> > > > > > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <
> > todd@cloudera.com
> > > > > > <javascript:;>> wrote:
> > > > > > > >
> > > > > > > >> +1 from me.
> > > > > > > >>
> > > > > > > >> I'll try to wrap up KUDU-1341 as well, but given that it's a
> > bit
> > > > of
> > > > > a
> > > > > > > >> riskier change we'll need some good testing of it. So,
> > > > > reaslitically,
> > > > > > > >> it might not make the cut. I wouldn't hold up the release
> for
> > > it,
> > > > > > > >> though we should add it to "known issues".
> > > > > > > >>
> > > > > > > >> -Todd
> > > > > > > >>
> > > > > > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > > > > > jdcryans@apache.org <javascript:;>>
> > > > > > > >> wrote:
> > > > > > > >> > Hello Kudu devs,
> > > > > > > >> >
> > > > > > > >> > Now that 0.7.0 is out of the way, we can think about
> what's
> > > > next.
> > > > > > > >> >
> > > > > > > >> > During our testing of this last version we found some
> > > > non-critical
> > > > > > > issues
> > > > > > > >> > that I think may be good bringing to folks earlier than
> our
> > > next
> > > > > > > planned
> > > > > > > >> > major release. This is my short list of commits:
> > > > > > > >> >
> > > > > > > >> > KUDU-1325: more crash avoidance during remote bootstrap
> and
> > > > tablet
> > > > > > > >> deletion
> > > > > > > >> > KUDU-1343. [java client] Scan sequence id is never reset
> > > between
> > > > > > > tablets
> > > > > > > >> > KUDU-1345. Fix case in which hybrid clock can run
> backwards
> > > > > > > >> >
> > > > > > > >> > Also, it'd be nice to fix our small license issues:
> > > > > > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > > > > > >> >
> > > > > > > >> > Anything else you good folks would like to see in there?
> > FWIW
> > > I
> > > > > > don't
> > > > > > > >> want
> > > > > > > >> > too much bloat, just what's really useful. Following the
> > plan
> > > > from
> > > > > > my
> > > > > > > >> "1.0"
> > > > > > > >> > email, a 0.8.0 would come out in April anyways.
> > > > > > > >> >
> > > > > > > >> > As for a timeline, I'm hoping to get it all together by
> the
> > > end
> > > > of
> > > > > > > next
> > > > > > > >> > week.
> > > > > > > >> >
> > > > > > > >> > Thanks,
> > > > > > > >> >
> > > > > > > >> > J-D
> > > > > > > >>
> > > > > > > >>
> > > > > > > >>
> > > > > > > >> --
> > > > > > > >> Todd Lipcon
> > > > > > > >> Software Engineer, Cloudera
> > > > > > > >>
> > > > > > >
> > > > > > >
> > > > > > >
> > > > > > > --
> > > > > > > Todd Lipcon
> > > > > > > Software Engineer, Cloudera
> > > > > > >
> > > > > >
> > > > >
> > > >
> > >
> >
> >
> >
> > --
> > Todd Lipcon
> > Software Engineer, Cloudera
> >
>
--
Todd Lipcon
Software Engineer, Cloudera
Re: Thinking of doing a 0.7.1
Posted by Jean-Daniel Cryans <jd...@apache.org>.
Right, I had just copied what we were doing for the other branches. Honest
mistake, I swear!
Also, the branch is almost ready, just need the last patch from Todd for
license tweaking and my patch to remove -SNAPSHOT is already up for review.
J-D
On Wed, Mar 2, 2016 at 12:18 PM, Todd Lipcon <to...@cloudera.com> wrote:
> On Wed, Mar 2, 2016 at 12:10 PM, Adar Dembo <ad...@cloudera.com> wrote:
>
> > Ahh, I see. Perhaps we can name our future branches after y releases and
> > not z releases?
> >
> > The only reason to have a branch for each z release is if we need to
> patch
> > z release n after z release n+1 has shipped.
> >
>
> +1, I dont think it was a particulraly intentional decision :)
>
> -Todd
>
>
> >
> > On Wed, Mar 2, 2016 at 6:30 AM, Jean-Daniel Cryans <jd...@apache.org>
> > wrote:
> >
> > > We named that branch branch-0.7.0, not 0.7
> > >
> > > On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com> wrote:
> > >
> > > > Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be
> > based
> > > > there?
> > > >
> > > > On Tuesday, March 1, 2016, Jean-Daniel Cryans <jd...@apache.org>
> > > wrote:
> > > >
> > > > > I pushed all the backports for review on a new branch called
> > > > branch-0.7.x.
> > > > >
> > > > > There's still the question of what we want to fix from KUDU-1347.
> > > > >
> > > > > J-D
> > > > >
> > > > > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> > > > > <javascript:;>> wrote:
> > > > >
> > > > > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> > > > jdcryans@apache.org
> > > > > <javascript:;>>
> > > > > > wrote:
> > > > > > > Yeah good point about known issues.
> > > > > >
> > > > > > Right, I guess I should have mentioned that we should doc the
> > issues
> > > > > > you mentioned as 'known issues' in 0.7.0, especially if any are
> > > > > > regressions (I think KUDU-1325 is, but the others aren't)
> > > > > >
> > > > > > >
> > > > > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <
> todd@cloudera.com
> > > > > <javascript:;>> wrote:
> > > > > > >
> > > > > > >> +1 from me.
> > > > > > >>
> > > > > > >> I'll try to wrap up KUDU-1341 as well, but given that it's a
> bit
> > > of
> > > > a
> > > > > > >> riskier change we'll need some good testing of it. So,
> > > > reaslitically,
> > > > > > >> it might not make the cut. I wouldn't hold up the release for
> > it,
> > > > > > >> though we should add it to "known issues".
> > > > > > >>
> > > > > > >> -Todd
> > > > > > >>
> > > > > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > > > > jdcryans@apache.org <javascript:;>>
> > > > > > >> wrote:
> > > > > > >> > Hello Kudu devs,
> > > > > > >> >
> > > > > > >> > Now that 0.7.0 is out of the way, we can think about what's
> > > next.
> > > > > > >> >
> > > > > > >> > During our testing of this last version we found some
> > > non-critical
> > > > > > issues
> > > > > > >> > that I think may be good bringing to folks earlier than our
> > next
> > > > > > planned
> > > > > > >> > major release. This is my short list of commits:
> > > > > > >> >
> > > > > > >> > KUDU-1325: more crash avoidance during remote bootstrap and
> > > tablet
> > > > > > >> deletion
> > > > > > >> > KUDU-1343. [java client] Scan sequence id is never reset
> > between
> > > > > > tablets
> > > > > > >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> > > > > > >> >
> > > > > > >> > Also, it'd be nice to fix our small license issues:
> > > > > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > > > > >> >
> > > > > > >> > Anything else you good folks would like to see in there?
> FWIW
> > I
> > > > > don't
> > > > > > >> want
> > > > > > >> > too much bloat, just what's really useful. Following the
> plan
> > > from
> > > > > my
> > > > > > >> "1.0"
> > > > > > >> > email, a 0.8.0 would come out in April anyways.
> > > > > > >> >
> > > > > > >> > As for a timeline, I'm hoping to get it all together by the
> > end
> > > of
> > > > > > next
> > > > > > >> > week.
> > > > > > >> >
> > > > > > >> > Thanks,
> > > > > > >> >
> > > > > > >> > J-D
> > > > > > >>
> > > > > > >>
> > > > > > >>
> > > > > > >> --
> > > > > > >> Todd Lipcon
> > > > > > >> Software Engineer, Cloudera
> > > > > > >>
> > > > > >
> > > > > >
> > > > > >
> > > > > > --
> > > > > > Todd Lipcon
> > > > > > Software Engineer, Cloudera
> > > > > >
> > > > >
> > > >
> > >
> >
>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>
Re: Thinking of doing a 0.7.1
Posted by Todd Lipcon <to...@cloudera.com>.
On Wed, Mar 2, 2016 at 12:10 PM, Adar Dembo <ad...@cloudera.com> wrote:
> Ahh, I see. Perhaps we can name our future branches after y releases and
> not z releases?
>
> The only reason to have a branch for each z release is if we need to patch
> z release n after z release n+1 has shipped.
>
+1, I dont think it was a particulraly intentional decision :)
-Todd
>
> On Wed, Mar 2, 2016 at 6:30 AM, Jean-Daniel Cryans <jd...@apache.org>
> wrote:
>
> > We named that branch branch-0.7.0, not 0.7
> >
> > On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com> wrote:
> >
> > > Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be
> based
> > > there?
> > >
> > > On Tuesday, March 1, 2016, Jean-Daniel Cryans <jd...@apache.org>
> > wrote:
> > >
> > > > I pushed all the backports for review on a new branch called
> > > branch-0.7.x.
> > > >
> > > > There's still the question of what we want to fix from KUDU-1347.
> > > >
> > > > J-D
> > > >
> > > > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> > > > <javascript:;>> wrote:
> > > >
> > > > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> > > jdcryans@apache.org
> > > > <javascript:;>>
> > > > > wrote:
> > > > > > Yeah good point about known issues.
> > > > >
> > > > > Right, I guess I should have mentioned that we should doc the
> issues
> > > > > you mentioned as 'known issues' in 0.7.0, especially if any are
> > > > > regressions (I think KUDU-1325 is, but the others aren't)
> > > > >
> > > > > >
> > > > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <todd@cloudera.com
> > > > <javascript:;>> wrote:
> > > > > >
> > > > > >> +1 from me.
> > > > > >>
> > > > > >> I'll try to wrap up KUDU-1341 as well, but given that it's a bit
> > of
> > > a
> > > > > >> riskier change we'll need some good testing of it. So,
> > > reaslitically,
> > > > > >> it might not make the cut. I wouldn't hold up the release for
> it,
> > > > > >> though we should add it to "known issues".
> > > > > >>
> > > > > >> -Todd
> > > > > >>
> > > > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > > > jdcryans@apache.org <javascript:;>>
> > > > > >> wrote:
> > > > > >> > Hello Kudu devs,
> > > > > >> >
> > > > > >> > Now that 0.7.0 is out of the way, we can think about what's
> > next.
> > > > > >> >
> > > > > >> > During our testing of this last version we found some
> > non-critical
> > > > > issues
> > > > > >> > that I think may be good bringing to folks earlier than our
> next
> > > > > planned
> > > > > >> > major release. This is my short list of commits:
> > > > > >> >
> > > > > >> > KUDU-1325: more crash avoidance during remote bootstrap and
> > tablet
> > > > > >> deletion
> > > > > >> > KUDU-1343. [java client] Scan sequence id is never reset
> between
> > > > > tablets
> > > > > >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> > > > > >> >
> > > > > >> > Also, it'd be nice to fix our small license issues:
> > > > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > > > >> >
> > > > > >> > Anything else you good folks would like to see in there? FWIW
> I
> > > > don't
> > > > > >> want
> > > > > >> > too much bloat, just what's really useful. Following the plan
> > from
> > > > my
> > > > > >> "1.0"
> > > > > >> > email, a 0.8.0 would come out in April anyways.
> > > > > >> >
> > > > > >> > As for a timeline, I'm hoping to get it all together by the
> end
> > of
> > > > > next
> > > > > >> > week.
> > > > > >> >
> > > > > >> > Thanks,
> > > > > >> >
> > > > > >> > J-D
> > > > > >>
> > > > > >>
> > > > > >>
> > > > > >> --
> > > > > >> Todd Lipcon
> > > > > >> Software Engineer, Cloudera
> > > > > >>
> > > > >
> > > > >
> > > > >
> > > > > --
> > > > > Todd Lipcon
> > > > > Software Engineer, Cloudera
> > > > >
> > > >
> > >
> >
>
--
Todd Lipcon
Software Engineer, Cloudera
Re: Thinking of doing a 0.7.1
Posted by Adar Dembo <ad...@cloudera.com>.
Ahh, I see. Perhaps we can name our future branches after y releases and
not z releases?
The only reason to have a branch for each z release is if we need to patch
z release n after z release n+1 has shipped.
On Wed, Mar 2, 2016 at 6:30 AM, Jean-Daniel Cryans <jd...@apache.org>
wrote:
> We named that branch branch-0.7.0, not 0.7
>
> On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com> wrote:
>
> > Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be based
> > there?
> >
> > On Tuesday, March 1, 2016, Jean-Daniel Cryans <jd...@apache.org>
> wrote:
> >
> > > I pushed all the backports for review on a new branch called
> > branch-0.7.x.
> > >
> > > There's still the question of what we want to fix from KUDU-1347.
> > >
> > > J-D
> > >
> > > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> > > <javascript:;>> wrote:
> > >
> > > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> > jdcryans@apache.org
> > > <javascript:;>>
> > > > wrote:
> > > > > Yeah good point about known issues.
> > > >
> > > > Right, I guess I should have mentioned that we should doc the issues
> > > > you mentioned as 'known issues' in 0.7.0, especially if any are
> > > > regressions (I think KUDU-1325 is, but the others aren't)
> > > >
> > > > >
> > > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <todd@cloudera.com
> > > <javascript:;>> wrote:
> > > > >
> > > > >> +1 from me.
> > > > >>
> > > > >> I'll try to wrap up KUDU-1341 as well, but given that it's a bit
> of
> > a
> > > > >> riskier change we'll need some good testing of it. So,
> > reaslitically,
> > > > >> it might not make the cut. I wouldn't hold up the release for it,
> > > > >> though we should add it to "known issues".
> > > > >>
> > > > >> -Todd
> > > > >>
> > > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > > jdcryans@apache.org <javascript:;>>
> > > > >> wrote:
> > > > >> > Hello Kudu devs,
> > > > >> >
> > > > >> > Now that 0.7.0 is out of the way, we can think about what's
> next.
> > > > >> >
> > > > >> > During our testing of this last version we found some
> non-critical
> > > > issues
> > > > >> > that I think may be good bringing to folks earlier than our next
> > > > planned
> > > > >> > major release. This is my short list of commits:
> > > > >> >
> > > > >> > KUDU-1325: more crash avoidance during remote bootstrap and
> tablet
> > > > >> deletion
> > > > >> > KUDU-1343. [java client] Scan sequence id is never reset between
> > > > tablets
> > > > >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> > > > >> >
> > > > >> > Also, it'd be nice to fix our small license issues:
> > > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > > >> >
> > > > >> > Anything else you good folks would like to see in there? FWIW I
> > > don't
> > > > >> want
> > > > >> > too much bloat, just what's really useful. Following the plan
> from
> > > my
> > > > >> "1.0"
> > > > >> > email, a 0.8.0 would come out in April anyways.
> > > > >> >
> > > > >> > As for a timeline, I'm hoping to get it all together by the end
> of
> > > > next
> > > > >> > week.
> > > > >> >
> > > > >> > Thanks,
> > > > >> >
> > > > >> > J-D
> > > > >>
> > > > >>
> > > > >>
> > > > >> --
> > > > >> Todd Lipcon
> > > > >> Software Engineer, Cloudera
> > > > >>
> > > >
> > > >
> > > >
> > > > --
> > > > Todd Lipcon
> > > > Software Engineer, Cloudera
> > > >
> > >
> >
>
Re: Thinking of doing a 0.7.1
Posted by Jean-Daniel Cryans <jd...@apache.org>.
We named that branch branch-0.7.0, not 0.7
On Tue, Mar 1, 2016 at 9:53 PM, Adar Dembo <ad...@cloudera.com> wrote:
> Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be based
> there?
>
> On Tuesday, March 1, 2016, Jean-Daniel Cryans <jd...@apache.org> wrote:
>
> > I pushed all the backports for review on a new branch called
> branch-0.7.x.
> >
> > There's still the question of what we want to fix from KUDU-1347.
> >
> > J-D
> >
> > On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> > <javascript:;>> wrote:
> >
> > > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <
> jdcryans@apache.org
> > <javascript:;>>
> > > wrote:
> > > > Yeah good point about known issues.
> > >
> > > Right, I guess I should have mentioned that we should doc the issues
> > > you mentioned as 'known issues' in 0.7.0, especially if any are
> > > regressions (I think KUDU-1325 is, but the others aren't)
> > >
> > > >
> > > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <todd@cloudera.com
> > <javascript:;>> wrote:
> > > >
> > > >> +1 from me.
> > > >>
> > > >> I'll try to wrap up KUDU-1341 as well, but given that it's a bit of
> a
> > > >> riskier change we'll need some good testing of it. So,
> reaslitically,
> > > >> it might not make the cut. I wouldn't hold up the release for it,
> > > >> though we should add it to "known issues".
> > > >>
> > > >> -Todd
> > > >>
> > > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > > jdcryans@apache.org <javascript:;>>
> > > >> wrote:
> > > >> > Hello Kudu devs,
> > > >> >
> > > >> > Now that 0.7.0 is out of the way, we can think about what's next.
> > > >> >
> > > >> > During our testing of this last version we found some non-critical
> > > issues
> > > >> > that I think may be good bringing to folks earlier than our next
> > > planned
> > > >> > major release. This is my short list of commits:
> > > >> >
> > > >> > KUDU-1325: more crash avoidance during remote bootstrap and tablet
> > > >> deletion
> > > >> > KUDU-1343. [java client] Scan sequence id is never reset between
> > > tablets
> > > >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> > > >> >
> > > >> > Also, it'd be nice to fix our small license issues:
> > > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > > >> >
> > > >> > Anything else you good folks would like to see in there? FWIW I
> > don't
> > > >> want
> > > >> > too much bloat, just what's really useful. Following the plan from
> > my
> > > >> "1.0"
> > > >> > email, a 0.8.0 would come out in April anyways.
> > > >> >
> > > >> > As for a timeline, I'm hoping to get it all together by the end of
> > > next
> > > >> > week.
> > > >> >
> > > >> > Thanks,
> > > >> >
> > > >> > J-D
> > > >>
> > > >>
> > > >>
> > > >> --
> > > >> Todd Lipcon
> > > >> Software Engineer, Cloudera
> > > >>
> > >
> > >
> > >
> > > --
> > > Todd Lipcon
> > > Software Engineer, Cloudera
> > >
> >
>
Re: Thinking of doing a 0.7.1
Posted by Adar Dembo <ad...@cloudera.com>.
Why aren't we reusing branch-0.7? Shouldn't every 0.7.x release be based
there?
On Tuesday, March 1, 2016, Jean-Daniel Cryans <jd...@apache.org> wrote:
> I pushed all the backports for review on a new branch called branch-0.7.x.
>
> There's still the question of what we want to fix from KUDU-1347.
>
> J-D
>
> On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <todd@cloudera.com
> <javascript:;>> wrote:
>
> > On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <jdcryans@apache.org
> <javascript:;>>
> > wrote:
> > > Yeah good point about known issues.
> >
> > Right, I guess I should have mentioned that we should doc the issues
> > you mentioned as 'known issues' in 0.7.0, especially if any are
> > regressions (I think KUDU-1325 is, but the others aren't)
> >
> > >
> > > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <todd@cloudera.com
> <javascript:;>> wrote:
> > >
> > >> +1 from me.
> > >>
> > >> I'll try to wrap up KUDU-1341 as well, but given that it's a bit of a
> > >> riskier change we'll need some good testing of it. So, reaslitically,
> > >> it might not make the cut. I wouldn't hold up the release for it,
> > >> though we should add it to "known issues".
> > >>
> > >> -Todd
> > >>
> > >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> > jdcryans@apache.org <javascript:;>>
> > >> wrote:
> > >> > Hello Kudu devs,
> > >> >
> > >> > Now that 0.7.0 is out of the way, we can think about what's next.
> > >> >
> > >> > During our testing of this last version we found some non-critical
> > issues
> > >> > that I think may be good bringing to folks earlier than our next
> > planned
> > >> > major release. This is my short list of commits:
> > >> >
> > >> > KUDU-1325: more crash avoidance during remote bootstrap and tablet
> > >> deletion
> > >> > KUDU-1343. [java client] Scan sequence id is never reset between
> > tablets
> > >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> > >> >
> > >> > Also, it'd be nice to fix our small license issues:
> > >> > https://issues.apache.org/jira/browse/KUDU-1347
> > >> >
> > >> > Anything else you good folks would like to see in there? FWIW I
> don't
> > >> want
> > >> > too much bloat, just what's really useful. Following the plan from
> my
> > >> "1.0"
> > >> > email, a 0.8.0 would come out in April anyways.
> > >> >
> > >> > As for a timeline, I'm hoping to get it all together by the end of
> > next
> > >> > week.
> > >> >
> > >> > Thanks,
> > >> >
> > >> > J-D
> > >>
> > >>
> > >>
> > >> --
> > >> Todd Lipcon
> > >> Software Engineer, Cloudera
> > >>
> >
> >
> >
> > --
> > Todd Lipcon
> > Software Engineer, Cloudera
> >
>
Re: Thinking of doing a 0.7.1
Posted by Jean-Daniel Cryans <jd...@apache.org>.
I pushed all the backports for review on a new branch called branch-0.7.x.
There's still the question of what we want to fix from KUDU-1347.
J-D
On Fri, Feb 26, 2016 at 3:15 PM, Todd Lipcon <to...@cloudera.com> wrote:
> On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <jd...@apache.org>
> wrote:
> > Yeah good point about known issues.
>
> Right, I guess I should have mentioned that we should doc the issues
> you mentioned as 'known issues' in 0.7.0, especially if any are
> regressions (I think KUDU-1325 is, but the others aren't)
>
> >
> > On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <to...@cloudera.com> wrote:
> >
> >> +1 from me.
> >>
> >> I'll try to wrap up KUDU-1341 as well, but given that it's a bit of a
> >> riskier change we'll need some good testing of it. So, reaslitically,
> >> it might not make the cut. I wouldn't hold up the release for it,
> >> though we should add it to "known issues".
> >>
> >> -Todd
> >>
> >> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <
> jdcryans@apache.org>
> >> wrote:
> >> > Hello Kudu devs,
> >> >
> >> > Now that 0.7.0 is out of the way, we can think about what's next.
> >> >
> >> > During our testing of this last version we found some non-critical
> issues
> >> > that I think may be good bringing to folks earlier than our next
> planned
> >> > major release. This is my short list of commits:
> >> >
> >> > KUDU-1325: more crash avoidance during remote bootstrap and tablet
> >> deletion
> >> > KUDU-1343. [java client] Scan sequence id is never reset between
> tablets
> >> > KUDU-1345. Fix case in which hybrid clock can run backwards
> >> >
> >> > Also, it'd be nice to fix our small license issues:
> >> > https://issues.apache.org/jira/browse/KUDU-1347
> >> >
> >> > Anything else you good folks would like to see in there? FWIW I don't
> >> want
> >> > too much bloat, just what's really useful. Following the plan from my
> >> "1.0"
> >> > email, a 0.8.0 would come out in April anyways.
> >> >
> >> > As for a timeline, I'm hoping to get it all together by the end of
> next
> >> > week.
> >> >
> >> > Thanks,
> >> >
> >> > J-D
> >>
> >>
> >>
> >> --
> >> Todd Lipcon
> >> Software Engineer, Cloudera
> >>
>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>
Re: Thinking of doing a 0.7.1
Posted by Todd Lipcon <to...@cloudera.com>.
On Fri, Feb 26, 2016 at 3:14 PM, Jean-Daniel Cryans <jd...@apache.org> wrote:
> Yeah good point about known issues.
Right, I guess I should have mentioned that we should doc the issues
you mentioned as 'known issues' in 0.7.0, especially if any are
regressions (I think KUDU-1325 is, but the others aren't)
>
> On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <to...@cloudera.com> wrote:
>
>> +1 from me.
>>
>> I'll try to wrap up KUDU-1341 as well, but given that it's a bit of a
>> riskier change we'll need some good testing of it. So, reaslitically,
>> it might not make the cut. I wouldn't hold up the release for it,
>> though we should add it to "known issues".
>>
>> -Todd
>>
>> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <jd...@apache.org>
>> wrote:
>> > Hello Kudu devs,
>> >
>> > Now that 0.7.0 is out of the way, we can think about what's next.
>> >
>> > During our testing of this last version we found some non-critical issues
>> > that I think may be good bringing to folks earlier than our next planned
>> > major release. This is my short list of commits:
>> >
>> > KUDU-1325: more crash avoidance during remote bootstrap and tablet
>> deletion
>> > KUDU-1343. [java client] Scan sequence id is never reset between tablets
>> > KUDU-1345. Fix case in which hybrid clock can run backwards
>> >
>> > Also, it'd be nice to fix our small license issues:
>> > https://issues.apache.org/jira/browse/KUDU-1347
>> >
>> > Anything else you good folks would like to see in there? FWIW I don't
>> want
>> > too much bloat, just what's really useful. Following the plan from my
>> "1.0"
>> > email, a 0.8.0 would come out in April anyways.
>> >
>> > As for a timeline, I'm hoping to get it all together by the end of next
>> > week.
>> >
>> > Thanks,
>> >
>> > J-D
>>
>>
>>
>> --
>> Todd Lipcon
>> Software Engineer, Cloudera
>>
--
Todd Lipcon
Software Engineer, Cloudera
Re: Thinking of doing a 0.7.1
Posted by Jean-Daniel Cryans <jd...@apache.org>.
Yeah good point about known issues.
On Fri, Feb 26, 2016 at 3:12 PM, Todd Lipcon <to...@cloudera.com> wrote:
> +1 from me.
>
> I'll try to wrap up KUDU-1341 as well, but given that it's a bit of a
> riskier change we'll need some good testing of it. So, reaslitically,
> it might not make the cut. I wouldn't hold up the release for it,
> though we should add it to "known issues".
>
> -Todd
>
> On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <jd...@apache.org>
> wrote:
> > Hello Kudu devs,
> >
> > Now that 0.7.0 is out of the way, we can think about what's next.
> >
> > During our testing of this last version we found some non-critical issues
> > that I think may be good bringing to folks earlier than our next planned
> > major release. This is my short list of commits:
> >
> > KUDU-1325: more crash avoidance during remote bootstrap and tablet
> deletion
> > KUDU-1343. [java client] Scan sequence id is never reset between tablets
> > KUDU-1345. Fix case in which hybrid clock can run backwards
> >
> > Also, it'd be nice to fix our small license issues:
> > https://issues.apache.org/jira/browse/KUDU-1347
> >
> > Anything else you good folks would like to see in there? FWIW I don't
> want
> > too much bloat, just what's really useful. Following the plan from my
> "1.0"
> > email, a 0.8.0 would come out in April anyways.
> >
> > As for a timeline, I'm hoping to get it all together by the end of next
> > week.
> >
> > Thanks,
> >
> > J-D
>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>
Re: Thinking of doing a 0.7.1
Posted by Todd Lipcon <to...@cloudera.com>.
+1 from me.
I'll try to wrap up KUDU-1341 as well, but given that it's a bit of a
riskier change we'll need some good testing of it. So, reaslitically,
it might not make the cut. I wouldn't hold up the release for it,
though we should add it to "known issues".
-Todd
On Fri, Feb 26, 2016 at 2:38 PM, Jean-Daniel Cryans <jd...@apache.org> wrote:
> Hello Kudu devs,
>
> Now that 0.7.0 is out of the way, we can think about what's next.
>
> During our testing of this last version we found some non-critical issues
> that I think may be good bringing to folks earlier than our next planned
> major release. This is my short list of commits:
>
> KUDU-1325: more crash avoidance during remote bootstrap and tablet deletion
> KUDU-1343. [java client] Scan sequence id is never reset between tablets
> KUDU-1345. Fix case in which hybrid clock can run backwards
>
> Also, it'd be nice to fix our small license issues:
> https://issues.apache.org/jira/browse/KUDU-1347
>
> Anything else you good folks would like to see in there? FWIW I don't want
> too much bloat, just what's really useful. Following the plan from my "1.0"
> email, a 0.8.0 would come out in April anyways.
>
> As for a timeline, I'm hoping to get it all together by the end of next
> week.
>
> Thanks,
>
> J-D
--
Todd Lipcon
Software Engineer, Cloudera