You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@hive.apache.org by Carl Steinbach <ca...@cloudera.com> on 2011/01/24 22:11:35 UTC
Scheduling the Hive 0.7.0 Release
Hi,
There's been a lot of talk recently about Hive 0.7.0, so in the
interest of getting the process moving I'd like to propose that we
create the release branch this Friday (1/28), and aim to
have the release out the door by the following Friday (2/4).
Since the Hive project is on a time-based release model, one of the
major release tasks involves preparing an assessment of the various
in-progress features. The purpose of this exercise is to provide users
with information about the status of features that are experimental or
partially implemented. I would really appreciate help from everyone in
putting this list together. Off the top of my head I think this
section needs to include information about the new features related to
security, indexing and statistics. If you have worked on any these
features please spend some time this week updating the relevant JIRA
tickets with notes explaining whether or not the feature is ready for
users and what problems users should expect to encounter.
Indexing:
HIVE-417: Implement Indexing in Hive
Security:
HIVE-1264: Make Hive Work with Hadoop Security
HIVE-842: Authentication Infrastructure for Hive
HIVE-78: Authorization Infrastructure for Hive
Statistics:
HIVE-1361: Table/Partition Level Statistics
Thanks.
Carl
Re: Scheduling the Hive 0.7.0 Release
Posted by Namit Jain <nj...@fb.com>.
Before finalizing on the date, let us go over the 68 issues.
Based on that, let us come up with a release date.
There is no point in having a release with lots of bugs or mostly
working features.
Don¹t get me wrong - I am very interested in 0.7 release, but don¹t
want to hurry up something - a delay of 2 weeks wont cause any major
harm.
Thanks,
-namit
On 1/24/11 2:17 PM, "Edward Capriolo" <ed...@gmail.com> wrote:
>On Mon, Jan 24, 2011 at 5:03 PM, Carl Steinbach <ca...@cloudera.com> wrote:
>> Hi Ed,
>>
>> This seems aggressive. We have 4 open blockers and 68 other issues
>>> slated for 0.7.0.
>>>
>>
>> Three of the four open blockers involve writing metastore upgrade
>>scripts.
>> I'm willing to assume responsibility for this and can get it ready for
>> review in the next couple of days. The same thing holds for the
>>commons-cli
>> downgrade ticket (HIVE-1817), I just need to people to agree on whether
>>or
>> not we can desupport Hadoop 0.20.0 in Hive 0.7.0.
>>
>> As for the 68 other issues that are marked 0.7.0, I'd appreciate your
>>help
>> in determining which of these are actually blockers versus the majority
>> which I think fall into the nice-to-have category. I'm going to make a
>>first
>> pass later today. Going forward I think we should make an effort to not
>>set
>> the "Fixed Version" field until a ticket is marked as Resolved/Fixed.
>>Doing
>> otherwise seems inconsistent with a a time-based release strategy.
>>
>> Thanks.
>>
>> Carl
>>
>
>Besides the "fixed version" we have to stop marking everything 'bug'.
>No one uses 'task' 'feature' 'wish' or the priority.
Re: Scheduling the Hive 0.7.0 Release
Posted by Edward Capriolo <ed...@gmail.com>.
On Mon, Jan 24, 2011 at 5:03 PM, Carl Steinbach <ca...@cloudera.com> wrote:
> Hi Ed,
>
> This seems aggressive. We have 4 open blockers and 68 other issues
>> slated for 0.7.0.
>>
>
> Three of the four open blockers involve writing metastore upgrade scripts.
> I'm willing to assume responsibility for this and can get it ready for
> review in the next couple of days. The same thing holds for the commons-cli
> downgrade ticket (HIVE-1817), I just need to people to agree on whether or
> not we can desupport Hadoop 0.20.0 in Hive 0.7.0.
>
> As for the 68 other issues that are marked 0.7.0, I'd appreciate your help
> in determining which of these are actually blockers versus the majority
> which I think fall into the nice-to-have category. I'm going to make a first
> pass later today. Going forward I think we should make an effort to not set
> the "Fixed Version" field until a ticket is marked as Resolved/Fixed. Doing
> otherwise seems inconsistent with a a time-based release strategy.
>
> Thanks.
>
> Carl
>
Besides the "fixed version" we have to stop marking everything 'bug'.
No one uses 'task' 'feature' 'wish' or the priority.
Re: Scheduling the Hive 0.7.0 Release
Posted by Carl Steinbach <ca...@cloudera.com>.
Hi Ed,
This seems aggressive. We have 4 open blockers and 68 other issues
> slated for 0.7.0.
>
Three of the four open blockers involve writing metastore upgrade scripts.
I'm willing to assume responsibility for this and can get it ready for
review in the next couple of days. The same thing holds for the commons-cli
downgrade ticket (HIVE-1817), I just need to people to agree on whether or
not we can desupport Hadoop 0.20.0 in Hive 0.7.0.
As for the 68 other issues that are marked 0.7.0, I'd appreciate your help
in determining which of these are actually blockers versus the majority
which I think fall into the nice-to-have category. I'm going to make a first
pass later today. Going forward I think we should make an effort to not set
the "Fixed Version" field until a ticket is marked as Resolved/Fixed. Doing
otherwise seems inconsistent with a a time-based release strategy.
Thanks.
Carl
Re: Scheduling the Hive 0.7.0 Release
Posted by Edward Capriolo <ed...@gmail.com>.
On Mon, Jan 24, 2011 at 4:11 PM, Carl Steinbach <ca...@cloudera.com> wrote:
> Hi,
>
> There's been a lot of talk recently about Hive 0.7.0, so in the
> interest of getting the process moving I'd like to propose that we
> create the release branch this Friday (1/28), and aim to
> have the release out the door by the following Friday (2/4).
>
> Since the Hive project is on a time-based release model, one of the
> major release tasks involves preparing an assessment of the various
> in-progress features. The purpose of this exercise is to provide users
> with information about the status of features that are experimental or
> partially implemented. I would really appreciate help from everyone in
> putting this list together. Off the top of my head I think this
> section needs to include information about the new features related to
> security, indexing and statistics. If you have worked on any these
> features please spend some time this week updating the relevant JIRA
> tickets with notes explaining whether or not the feature is ready for
> users and what problems users should expect to encounter.
>
> Indexing:
> HIVE-417: Implement Indexing in Hive
>
> Security:
> HIVE-1264: Make Hive Work with Hadoop Security
> HIVE-842: Authentication Infrastructure for Hive
> HIVE-78: Authorization Infrastructure for Hive
>
> Statistics:
> HIVE-1361: Table/Partition Level Statistics
>
> Thanks.
>
> Carl
>
This seems aggressive. We have 4 open blockers and 68 other issues
slated for 0.7.0.
https://issues.apache.org/jira/secure/IssueNavigator.jspa?reset=true&mode=hide&sorter/order=DESC&sorter/field=priority&resolution=-1&pid=12310843&fixfor=12315150
The indexing, security, and statistics issues effect the metastore.
This type of changes should be heavily vetted.
Edward