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

Switch to Threaded View
HBase >> mail # dev >> HEADSUP: Working on new 0.96.0RC


Copy link to this message
-
Re: HEADSUP: Working on new 0.96.0RC
> We're not releasing a new rc tonight, it seems really weird to hold up a
bug fix to try and hit some unknown, and un-agreed upon, deadline.
There is no deadline, but we cannot hold the RC for non-blocker patches
especially once the release candidate process has begun. We are not going
to solve every bug in existence (look at HBASE-9721 for example) in HBase.
In the usual case for a release, once an RC is cut, you do not want to
destabilize by adding risky patches and continue on this kind cat and mouse
game. Note that 9696 did not hold up the cut for the previous RC, why
should it hold it now?

> To me it feels like HBASE-9724 should go into 0.96
I am fine with committing HBASE-9724. I did not commit that yesterday since
I though we should be doing an RC and I did not want last minute fixes to
AM.

Let's get 9724 and a solution for 9563 in. Would that work?

Enis
On Wed, Oct 9, 2013 at 9:22 PM, Elliott Clark <[EMAIL PROTECTED]> wrote:

> To me it feels like HBASE-9724 should go into 0.96.  We're not releasing a
> new rc tonight, it seems really weird to hold up a bug fix to try and hit
> some unknown, and un-agreed upon, deadline.
>
>
> On Wed, Oct 9, 2013 at 8:08 PM, Stack <[EMAIL PROTECTED]> wrote:
>
>> On Wed, Oct 9, 2013 at 7:14 PM, Enis Söztutar <[EMAIL PROTECTED]> wrote:
>>
>>> > Anyways, if you fellas can't wait anymore, just say and we'll figure
>>> out
>>> something.
>>> As I see it, HBASE-9563 is committed,
>>
>>
>> It is still open and committed with qualification "Stack:...Was going to
>> try this first but likely needs more..."  and "Elliott: +1 I think it's
>> an improvement even if it doesn't 100% fix the master issue."
>>
>>
>>
>>> and HBASE-9696 is not a blocker
>>> against 0.96. But if you argue that 9696 is indeed a blocker, let's raise
>>> it as such.
>>
>>
>>
>> Agree.
>>
>>
>>
>>> There is no point in creating an RC, an immediately sinking it
>>> if we cannot verify the RC for a +1. We don't run into data loss issues
>>> anymore which is why I still think we can release 0.96 even without 9696
>>> and 9724. Nothing is preventing us to release 0.96.1, with this and more
>>> fixes in let's say a couple of weeks or months.
>>>
>>> I guess let's wait for tomorrow to see whether there is any progress on
>>> 9563 and 9696.
>>>
>>
>> Yes.  Lets take this up tomorrow.  Elliott and I are on the master issue,
>> HBASE-9563, this evening.
>>
>> Thanks Enis,
>> St.Ack
>>
>
>