hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Ted Yu <yuzhih...@gmail.com>
Subject time based release Was: code review: HBASE-5231 Backport HBASE-3373 (per-table load balancing) to 0.92
Date Sat, 21 Jan 2012 04:55:46 GMT
I suggest first completing HBASE-2600 and HBASE-4218 before branching 0.94
We have 5 Jenkins builds so far. After branching, we need to maintain 5
Jenkins builds.

Cheers

On Fri, Jan 20, 2012 at 8:33 PM, lars hofhansl <lhofhansl@yahoo.com> wrote:

> I agree. As soon as 0.92.0 is released, let's talk about 0.94.
>
> +1 on time based releases!
>
>
> I think 0.92 shows that not doing frequent releases leads to a lot cruft
> accumulating in trunk, which then takes a lot of time to stabalize.
>
>
>
> ________________________________
>  From: Todd Lipcon <todd@cloudera.com>
> To: dev@hbase.apache.org
> Cc: stack <stack@duboce.net>; lars hofhansl <lhofhansl@yahoo.com>;
> Jean-Daniel Cryans <jdcryans@gmail.com>
> Sent: Friday, January 20, 2012 10:07 AM
> Subject: Re: code review: HBASE-5231 Backport HBASE-3373 (per-table load
> balancing) to 0.92
>
> On Fri, Jan 20, 2012 at 10:04 AM, Ted Yu <yuzhihong@gmail.com> wrote:
> > I made Fix Version of HBASE-5231 0.92.1
> >
> > I can wait till 0.92.0 is released.
>
> IMO we should push it to 0.94.0. Point releases should be for minor
> improvements and bug fixes. The fact that we divert so much energy
> into past releases is part of why we're so slow to release the new
> ones -- and the long period between the new releases I think
> contributes to how unstable trunk gets in the meantime. I won't -1 it
> for 0.92.1, but it just seems like a Bad Idea.
>
> /me goes back to beating the 'we should do time-based releases' drum
>
> -Todd
>
> >
> > On Fri, Jan 20, 2012 at 10:00 AM, Todd Lipcon <todd@cloudera.com> wrote:
> >
> >> Can we please wait until 0.92 is released before we talk about putting
> >> more stuff in there? This release process has dragged on for months
> >> because we can't seem to say "no".
> >>
> >> Todd
> >>
> >> On Thu, Jan 19, 2012 at 9:53 PM, Ted Yu <yuzhihong@gmail.com> wrote:
> >> > Hi,
> >> > The person sitting next to me at the HBase meetup tonight told me that
> >> they
> >> > once thought of allocating one cluster per table - just to keep
> regions
> >> in
> >> > each table balanced.
> >> >
> >> > With HBASE-5231, you don't have to do that anymore.
> >> >
> >> > Please review the patch.
> >> >
> >> > Thanks
> >>
> >>
> >>
> >> --
> >> Todd Lipcon
> >> Software Engineer, Cloudera
> >>
>
>
>
> --
> Todd Lipcon
> Software Engineer, Cloudera
>

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message