You are viewing a plain text version of this content. The canonical link for it is here.
Posted to users@kafka.apache.org by Neha Narkhede <ne...@gmail.com> on 2011/10/14 03:54:25 UTC

[VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Hi,

This is the third candidate for the first incubator release for Apache
Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
adds a standard .rat-excludes
files along with rat scripts to Kafka.

It fixes the following issues:
http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html

*** Please download, test and vote by noon Oct, 18th, 6 pm

Note that we are voting upon the source (tag), binaries are provided
for convenience.

Source and binary files:
http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/

The tag to be voted upon:
http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3

Kafka's KEYS file containing PGP keys we use to sign the release:
http://svn.apache.org/repos/asf/incubator/kafka/KEYS

Note that the Incubator PMC needs to vote upon the release after a
successful PPMC vote before
any release can be made official.

Thanks,
Neha

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Chris Douglas <cd...@apache.org>.
Any combination of repository tweaks are valid as long as it's clear
what people are voting on. The signatures/etc. are helpful mechanisms
to ensure nobody is confused, but it's possible to get carried away.
You don't need to follow a specific incantation to generate a valid
release, as long as the artifact approved by the voters is
unambiguously identified. Identifying it in the repository is just
good release management.

One release process that's worked in the past (completely
optional/pedantic steps in square brackets):

Let ROOT be http://svn.apache.org/repos/asf/incubator/kafka
Let VERSION be a version string, e.g. kafka-0.5.0-incubating
[1.] If releasing a new minor version (x.y), create a branch at
$ROOT/branches/$VERSION; patch releases x.y.z don't need branches)
2. Modify CHANGES.txt, follow whatever packaging is appropriate for the release
3. Create an tarball from that branch, contents matching the final release
4. Create signature/MD5 and upload to people.apache.org/~username/$VERSION-$RC
[5.] Create a tag in the repository at $ROOT/tags/$VERSION-$RC
6. Run a vote pointing to the artifact/tag
7. If the vote fails or problems are found, bump $RC, fix on
branch/trunk, and goto 3
[8.] Rename the tag of the approved $RC to $VERSION, cleaning up other RC tags

The above is pretty much exactly what you described, but adding a
release branch and RC tags (again, just a convenience for devs). If
you're only releasing from trunk, you probably don't need a release
branch, but they're cheap and often useful. -C

On Fri, Oct 14, 2011 at 4:42 PM, Neha Narkhede <ne...@gmail.com> wrote:
> 1. Create tgz named kafka-0.7.0-incubating.tar.gz
> 2. Create the PGP signature for it - kafka-0.7.0-incubating.tar.gz.asc
> 3. Create the MD5 signature for it - kafka-0.7.0-incubating.tar.gz.md5
> 4. Specify the svn revision to be voted upon. For example,
> http://svn.apache.org/repos/asf/incubator/kafka/trunk@r1181176
> 5. Upload the tgz to a URL that looks like
> http://*/kafka-0.7.0-incubating-candidate-1/
>
> Now, say if RC1 vote doesn't pass -
>
> 1. Fix the bugs in RC1
> 2. Create tgz named kafka-0.7.0-incubating.tar.gz
> 3. Create the PGP signature for it - kafka-0.7.0-incubating.tar.gz.asc
> 4. Create the MD5 signature for it - kafka-0.7.0-incubating.tar.gz.md5
> 5. Specify the NEW svn revision to be voted upon. For example,
> http://svn.apache.org/repos/asf/incubator/kafka/trunk@r1181186
> 6. Upload the tgz to a URL that looks like
> http://*/kafka-0.7.0-incubating-candidate-2/
>
> If RC1 vote passes, then -
>
> 1. Create svn tag named kafka-0.7.0-incubating, pointing to the SVN
> revision that was voted upon
> 2. Upload the tgz to a URL that looks like http://*/kafka-0.7.0-incubating/
>
> It will be good to iron this out once and follow it in the forthcoming releases.
> We plan to use this process to start a new vote on Monday morning, Oct 17th.
> Please let us know if this process looks good to the mentors.
>
> Thanks,
> Neha
>
> On Fri, Oct 14, 2011 at 2:06 PM, Chris Douglas <cd...@apache.org> wrote:
>>> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.
>>
>> As long as the signature and MD5 don't change, renaming the artifact
>> is fine (though it's more convenient for the release manager to put
>> the final artifacts in a <version>-RC<n> directory). In this case, the
>> tarball unpacks to "kafka-0.7.0-incubating-candidate-3", which is kind
>> of unfortunate. Might as well clean that up.
>>
>> It's a software release, not a shuttle launch. Is this the only
>> outstanding issue (I saw that RAT passed in KAFKA-143)? JIRA looks
>> clean.
>>
>> We should find someone to sign the release key. -C
>>
>> On Fri, Oct 14, 2011 at 1:18 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>>
>>> On Oct 14, 2011, at 1:08 PM, Neha Narkhede wrote:
>>>
>>>> Alan,
>>>>
>>>>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>>>
>>>> We intended to let people vote on the RC3, let the vote pass, and then
>>>> just rename the tag and tgz to remove the "-candidate-3" from the
>>>> name. The contents will stay the same.
>>>
>>> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.
>>>
>>>> Our understanding is that if
>>>> the contents stay the same, we don't need another vote. Is that right
>>>> ?
>>>
>>> Restarting the vote is very common even for well established projects.  It's to be expected for podlings that are working out the kinks.  I wouldn't worry about it.
>>>
>>>
>>> Regards,
>>> Alan
>>>
>>>>
>>>> Thanks,
>>>> Neha
>>>>
>>>> On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>>>>
>>>>> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>>>>>
>>>>>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>>>>>
>>>>>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>>>>>> specific work, we can pass on that for this release. We can start
>>>>>> publishing to Maven after the next release.
>>>>>
>>>>> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>>>>>
>>>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>>>>
>>>>> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>>>>>
>>>>>
>>>>>
>>>>> Regards,
>>>>> Alan
>>>>>
>>>>>
>>>
>>>
>>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
Chris,

>> In this case, the tarball unpacks to "kafka-0.7.0-incubating-candidate-3", which is kind of unfortunate. Might as well clean that up.

I agree. In this case, we'd need to rename the directory as well,
which would change the MD5, so I guess that needs another vote.

Going forward, we want to confirm the exact process to follow for
publishing a release. Let me state our understanding here. Please
correct
us now, if you don't agree with the following

For example, we are release RC1 for Kafka 0.7.0-incubating -

1. Create tgz named kafka-0.7.0-incubating.tar.gz
2. Create the PGP signature for it - kafka-0.7.0-incubating.tar.gz.asc
3. Create the MD5 signature for it - kafka-0.7.0-incubating.tar.gz.md5
4. Specify the svn revision to be voted upon. For example,
http://svn.apache.org/repos/asf/incubator/kafka/trunk@r1181176
5. Upload the tgz to a URL that looks like
http://*/kafka-0.7.0-incubating-candidate-1/

Now, say if RC1 vote doesn't pass -

1. Fix the bugs in RC1
2. Create tgz named kafka-0.7.0-incubating.tar.gz
3. Create the PGP signature for it - kafka-0.7.0-incubating.tar.gz.asc
4. Create the MD5 signature for it - kafka-0.7.0-incubating.tar.gz.md5
5. Specify the NEW svn revision to be voted upon. For example,
http://svn.apache.org/repos/asf/incubator/kafka/trunk@r1181186
6. Upload the tgz to a URL that looks like
http://*/kafka-0.7.0-incubating-candidate-2/

If RC1 vote passes, then -

1. Create svn tag named kafka-0.7.0-incubating, pointing to the SVN
revision that was voted upon
2. Upload the tgz to a URL that looks like http://*/kafka-0.7.0-incubating/

It will be good to iron this out once and follow it in the forthcoming releases.
We plan to use this process to start a new vote on Monday morning, Oct 17th.
Please let us know if this process looks good to the mentors.

Thanks,
Neha

On Fri, Oct 14, 2011 at 2:06 PM, Chris Douglas <cd...@apache.org> wrote:
>> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.
>
> As long as the signature and MD5 don't change, renaming the artifact
> is fine (though it's more convenient for the release manager to put
> the final artifacts in a <version>-RC<n> directory). In this case, the
> tarball unpacks to "kafka-0.7.0-incubating-candidate-3", which is kind
> of unfortunate. Might as well clean that up.
>
> It's a software release, not a shuttle launch. Is this the only
> outstanding issue (I saw that RAT passed in KAFKA-143)? JIRA looks
> clean.
>
> We should find someone to sign the release key. -C
>
> On Fri, Oct 14, 2011 at 1:18 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>
>> On Oct 14, 2011, at 1:08 PM, Neha Narkhede wrote:
>>
>>> Alan,
>>>
>>>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>>
>>> We intended to let people vote on the RC3, let the vote pass, and then
>>> just rename the tag and tgz to remove the "-candidate-3" from the
>>> name. The contents will stay the same.
>>
>> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.
>>
>>> Our understanding is that if
>>> the contents stay the same, we don't need another vote. Is that right
>>> ?
>>
>> Restarting the vote is very common even for well established projects.  It's to be expected for podlings that are working out the kinks.  I wouldn't worry about it.
>>
>>
>> Regards,
>> Alan
>>
>>>
>>> Thanks,
>>> Neha
>>>
>>> On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>>>
>>>> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>>>>
>>>>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>>>>
>>>>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>>>>> specific work, we can pass on that for this release. We can start
>>>>> publishing to Maven after the next release.
>>>>
>>>> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>>>>
>>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>>>
>>>> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>>>>
>>>>
>>>>
>>>> Regards,
>>>> Alan
>>>>
>>>>
>>
>>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Chris Douglas <cd...@apache.org>.
> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.

As long as the signature and MD5 don't change, renaming the artifact
is fine (though it's more convenient for the release manager to put
the final artifacts in a <version>-RC<n> directory). In this case, the
tarball unpacks to "kafka-0.7.0-incubating-candidate-3", which is kind
of unfortunate. Might as well clean that up.

It's a software release, not a shuttle launch. Is this the only
outstanding issue (I saw that RAT passed in KAFKA-143)? JIRA looks
clean.

We should find someone to sign the release key. -C

On Fri, Oct 14, 2011 at 1:18 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>
> On Oct 14, 2011, at 1:08 PM, Neha Narkhede wrote:
>
>> Alan,
>>
>>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>
>> We intended to let people vote on the RC3, let the vote pass, and then
>> just rename the tag and tgz to remove the "-candidate-3" from the
>> name. The contents will stay the same.
>
> The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.
>
>> Our understanding is that if
>> the contents stay the same, we don't need another vote. Is that right
>> ?
>
> Restarting the vote is very common even for well established projects.  It's to be expected for podlings that are working out the kinks.  I wouldn't worry about it.
>
>
> Regards,
> Alan
>
>>
>> Thanks,
>> Neha
>>
>> On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>>
>>> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>>>
>>>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>>>
>>>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>>>> specific work, we can pass on that for this release. We can start
>>>> publishing to Maven after the next release.
>>>
>>> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>>>
>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>>>
>>> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>>>
>>>
>>>
>>> Regards,
>>> Alan
>>>
>>>
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Oct 14, 2011, at 1:08 PM, Neha Narkhede wrote:

> Alan,
> 
>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
> 
> We intended to let people vote on the RC3, let the vote pass, and then
> just rename the tag and tgz to remove the "-candidate-3" from the
> name. The contents will stay the same.

The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.   

> Our understanding is that if
> the contents stay the same, we don't need another vote. Is that right
> ?

Restarting the vote is very common even for well established projects.  It's to be expected for podlings that are working out the kinks.  I wouldn't worry about it.


Regards,
Alan

> 
> Thanks,
> Neha
> 
> On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>> 
>> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>> 
>>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>> 
>>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>>> specific work, we can pass on that for this release. We can start
>>> publishing to Maven after the next release.
>> 
>> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>> 
>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>> 
>> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>> 
>> 
>> 
>> Regards,
>> Alan
>> 
>> 


Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Oct 14, 2011, at 1:08 PM, Neha Narkhede wrote:

> Alan,
> 
>>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
> 
> We intended to let people vote on the RC3, let the vote pass, and then
> just rename the tag and tgz to remove the "-candidate-3" from the
> name. The contents will stay the same.

The *contents* will contain the "-candidate-3".  Not sure why the project would want that especially since you intend to rename the tag.  Also, when you rename the tag there's no way to tell from the proposal that's being voted on that this is what happened.   

> Our understanding is that if
> the contents stay the same, we don't need another vote. Is that right
> ?

Restarting the vote is very common even for well established projects.  It's to be expected for podlings that are working out the kinks.  I wouldn't worry about it.


Regards,
Alan

> 
> Thanks,
> Neha
> 
> On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>> 
>> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>> 
>>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>> 
>>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>>> specific work, we can pass on that for this release. We can start
>>> publishing to Maven after the next release.
>> 
>> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>> 
>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>> 
>> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>> 
>> 
>> 
>> Regards,
>> Alan
>> 
>> 


Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
Alan,

>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.

We intended to let people vote on the RC3, let the vote pass, and then
just rename the tag and tgz to remove the "-candidate-3" from the
name. The contents will stay the same. Our understanding is that if
the contents stay the same, we don't need another vote. Is that right
?

Thanks,
Neha

On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>
> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>
>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>
>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>> specific work, we can pass on that for this release. We can start
>> publishing to Maven after the next release.
>
> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>
> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>
> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>
>
>
> Regards,
> Alan
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Chris Burroughs <ch...@gmail.com>.
On 10/14/2011 03:44 PM, Alan D. Cabrera wrote:
> That's fine.  What do others in the community think about this?  Were
> they counting on a release of Maven artifacts as well?  If that's
> what is everyone's understanding then the release should be good on
> this front.

My expectation was that maven artifacts were a "after the first release"
thing.  I don't recall anyone expressing expectations to the contrary.

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Chris Burroughs <ch...@gmail.com>.
On 10/14/2011 03:44 PM, Alan D. Cabrera wrote:
> That's fine.  What do others in the community think about this?  Were
> they counting on a release of Maven artifacts as well?  If that's
> what is everyone's understanding then the release should be good on
> this front.

My expectation was that maven artifacts were a "after the first release"
thing.  I don't recall anyone expressing expectations to the contrary.

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
Alan,

>> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.

We intended to let people vote on the RC3, let the vote pass, and then
just rename the tag and tgz to remove the "-candidate-3" from the
name. The contents will stay the same. Our understanding is that if
the contents stay the same, we don't need another vote. Is that right
?

Thanks,
Neha

On Fri, Oct 14, 2011 at 12:44 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>
> On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:
>
>>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>>
>> So since KAFKA-133 is not yet resolved and will take quite some sbt
>> specific work, we can pass on that for this release. We can start
>> publishing to Maven after the next release.
>
> That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.
>
> Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.
>
> Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.
>
>
>
> Regards,
> Alan
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:

>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
> 
> So since KAFKA-133 is not yet resolved and will take quite some sbt
> specific work, we can pass on that for this release. We can start
> publishing to Maven after the next release.

That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.

Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.   

Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.



Regards,
Alan


Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Oct 14, 2011, at 11:51 AM, Neha Narkhede wrote:

>>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
> 
> So since KAFKA-133 is not yet resolved and will take quite some sbt
> specific work, we can pass on that for this release. We can start
> publishing to Maven after the next release.

That's fine.  What do others in the community think about this?  Were they counting on a release of Maven artifacts as well?  If that's what is everyone's understanding then the release should be good on this front.

Were you going to rebuild the tgz so that it did not have the candidate-3 bit in its path?  What about the rename of the tag?  If you change any of these you need to restart the vote since this email is the record of what's been voted on.   

Thanks for being so patent.  Getting the first release out for an incubating project always has its fits and starts.



Regards,
Alan


Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.

So since KAFKA-133 is not yet resolved and will take quite some sbt
specific work, we can pass on that for this release. We can start
publishing to Maven after the next release.

Seems like we are unblocked on voting for this RC then. Lets start the
vote, the deadline is still Oct 18th, 6pm

Thanks,
Neha


On Fri, Oct 14, 2011 at 11:37 AM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>
> On Oct 14, 2011, at 11:21 AM, Neha Narkhede wrote:
>
>> Alan,
>>
>> Thanks for looking into this.
>>
>>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>>
>> Yes. We intend to publish the tgz too. That helps users run the
>> quickstart and unit tests on a release quickly. I checked some other
>> incubator projects and top-level Apache projects. They seemed to do
>> the same.
>
> Yes, that's fine.  Your statement "binaries are provided for convenience" is what confused me.  When a vote occurs all the artifacts and their signatures that will be published should be provided for review.
>
>>
>>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>>
>> We actually haven't yet resolved
>> https://issues.apache.org/jira/browse/KAFKA-133. From what you had
>> mentioned, our understanding was that you have to wait for at least
>> one release to be able to publish to a Maven repo. Do you think this
>> blocks us ?
>
> If it's the intention of the project to publish Maven artifacts as part of this release then they need to be provided following the guidelines outlined in http://www.apache.org/dev/publishing-maven-artifacts.html.  This means that KAFKA-133 needs to be resolved before the vote.  Maven artifacts cannot be published after the fact even though they may be built from a tag.
>
> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>
>>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>>
>> I'm thinking that since RC3 and the release tag is the same, we can
>> add the kafka-0.7.0-incubating tag and delete the
>> kafka-0.7.0-incubating-candidate-3 tag.
>
> Simply moving the tag should be fine unless you want to pick up more stuff from trunk.
>
>
> Regards,
> Alan
>
>>
>> Let us know if this sounds good.
>>
>> Thanks,
>> Neha
>>
>> On Thu, Oct 13, 2011 at 9:26 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>>>
>>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>>>
>>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>>>
>>>
>>> Regards,
>>> Alan
>>>
>>> On Oct 13, 2011, at 6:54 PM, Neha Narkhede wrote:
>>>
>>>> Hi,
>>>>
>>>> This is the third candidate for the first incubator release for Apache
>>>> Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
>>>> adds a standard .rat-excludes
>>>> files along with rat scripts to Kafka.
>>>>
>>>> It fixes the following issues:
>>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html
>>>>
>>>> *** Please download, test and vote by noon Oct, 18th, 6 pm
>>>>
>>>> Note that we are voting upon the source (tag), binaries are provided
>>>> for convenience.
>>>>
>>>> Source and binary files:
>>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/
>>>>
>>>> The tag to be voted upon:
>>>> http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3
>>>>
>>>> Kafka's KEYS file containing PGP keys we use to sign the release:
>>>> http://svn.apache.org/repos/asf/incubator/kafka/KEYS
>>>>
>>>> Note that the Incubator PMC needs to vote upon the release after a
>>>> successful PPMC vote before
>>>> any release can be made official.
>>>>
>>>> Thanks,
>>>> Neha
>>>
>>>
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
>> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.

So since KAFKA-133 is not yet resolved and will take quite some sbt
specific work, we can pass on that for this release. We can start
publishing to Maven after the next release.

Seems like we are unblocked on voting for this RC then. Lets start the
vote, the deadline is still Oct 18th, 6pm

Thanks,
Neha


On Fri, Oct 14, 2011 at 11:37 AM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>
> On Oct 14, 2011, at 11:21 AM, Neha Narkhede wrote:
>
>> Alan,
>>
>> Thanks for looking into this.
>>
>>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>>
>> Yes. We intend to publish the tgz too. That helps users run the
>> quickstart and unit tests on a release quickly. I checked some other
>> incubator projects and top-level Apache projects. They seemed to do
>> the same.
>
> Yes, that's fine.  Your statement "binaries are provided for convenience" is what confused me.  When a vote occurs all the artifacts and their signatures that will be published should be provided for review.
>
>>
>>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>>
>> We actually haven't yet resolved
>> https://issues.apache.org/jira/browse/KAFKA-133. From what you had
>> mentioned, our understanding was that you have to wait for at least
>> one release to be able to publish to a Maven repo. Do you think this
>> blocks us ?
>
> If it's the intention of the project to publish Maven artifacts as part of this release then they need to be provided following the guidelines outlined in http://www.apache.org/dev/publishing-maven-artifacts.html.  This means that KAFKA-133 needs to be resolved before the vote.  Maven artifacts cannot be published after the fact even though they may be built from a tag.
>
> To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.
>
>>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>>
>> I'm thinking that since RC3 and the release tag is the same, we can
>> add the kafka-0.7.0-incubating tag and delete the
>> kafka-0.7.0-incubating-candidate-3 tag.
>
> Simply moving the tag should be fine unless you want to pick up more stuff from trunk.
>
>
> Regards,
> Alan
>
>>
>> Let us know if this sounds good.
>>
>> Thanks,
>> Neha
>>
>> On Thu, Oct 13, 2011 at 9:26 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>>>
>>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>>>
>>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>>>
>>>
>>> Regards,
>>> Alan
>>>
>>> On Oct 13, 2011, at 6:54 PM, Neha Narkhede wrote:
>>>
>>>> Hi,
>>>>
>>>> This is the third candidate for the first incubator release for Apache
>>>> Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
>>>> adds a standard .rat-excludes
>>>> files along with rat scripts to Kafka.
>>>>
>>>> It fixes the following issues:
>>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html
>>>>
>>>> *** Please download, test and vote by noon Oct, 18th, 6 pm
>>>>
>>>> Note that we are voting upon the source (tag), binaries are provided
>>>> for convenience.
>>>>
>>>> Source and binary files:
>>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/
>>>>
>>>> The tag to be voted upon:
>>>> http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3
>>>>
>>>> Kafka's KEYS file containing PGP keys we use to sign the release:
>>>> http://svn.apache.org/repos/asf/incubator/kafka/KEYS
>>>>
>>>> Note that the Incubator PMC needs to vote upon the release after a
>>>> successful PPMC vote before
>>>> any release can be made official.
>>>>
>>>> Thanks,
>>>> Neha
>>>
>>>
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
On Oct 14, 2011, at 11:21 AM, Neha Narkhede wrote:

> Alan,
> 
> Thanks for looking into this.
> 
>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
> 
> Yes. We intend to publish the tgz too. That helps users run the
> quickstart and unit tests on a release quickly. I checked some other
> incubator projects and top-level Apache projects. They seemed to do
> the same.

Yes, that's fine.  Your statement "binaries are provided for convenience" is what confused me.  When a vote occurs all the artifacts and their signatures that will be published should be provided for review.

> 
>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
> 
> We actually haven't yet resolved
> https://issues.apache.org/jira/browse/KAFKA-133. From what you had
> mentioned, our understanding was that you have to wait for at least
> one release to be able to publish to a Maven repo. Do you think this
> blocks us ?

If it's the intention of the project to publish Maven artifacts as part of this release then they need to be provided following the guidelines outlined in http://www.apache.org/dev/publishing-maven-artifacts.html.  This means that KAFKA-133 needs to be resolved before the vote.  Maven artifacts cannot be published after the fact even though they may be built from a tag.

To recap, all the artifacts that are going to be published as part of the release have to be provided for review.  Artifacts cannot be published after the fact even though they may be built from a tag.

>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
> 
> I'm thinking that since RC3 and the release tag is the same, we can
> add the kafka-0.7.0-incubating tag and delete the
> kafka-0.7.0-incubating-candidate-3 tag.

Simply moving the tag should be fine unless you want to pick up more stuff from trunk.


Regards,
Alan

> 
> Let us know if this sounds good.
> 
> Thanks,
> Neha
> 
> On Thu, Oct 13, 2011 at 9:26 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
>> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>> 
>> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>> 
>> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>> 
>> 
>> Regards,
>> Alan
>> 
>> On Oct 13, 2011, at 6:54 PM, Neha Narkhede wrote:
>> 
>>> Hi,
>>> 
>>> This is the third candidate for the first incubator release for Apache
>>> Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
>>> adds a standard .rat-excludes
>>> files along with rat scripts to Kafka.
>>> 
>>> It fixes the following issues:
>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html
>>> 
>>> *** Please download, test and vote by noon Oct, 18th, 6 pm
>>> 
>>> Note that we are voting upon the source (tag), binaries are provided
>>> for convenience.
>>> 
>>> Source and binary files:
>>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/
>>> 
>>> The tag to be voted upon:
>>> http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3
>>> 
>>> Kafka's KEYS file containing PGP keys we use to sign the release:
>>> http://svn.apache.org/repos/asf/incubator/kafka/KEYS
>>> 
>>> Note that the Incubator PMC needs to vote upon the release after a
>>> successful PPMC vote before
>>> any release can be made official.
>>> 
>>> Thanks,
>>> Neha
>> 
>> 


Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by Neha Narkhede <ne...@gmail.com>.
Alan,

Thanks for looking into this.

> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.

Yes. We intend to publish the tgz too. That helps users run the
quickstart and unit tests on a release quickly. I checked some other
incubator projects and top-level Apache projects. They seemed to do
the same.

> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html

We actually haven't yet resolved
https://issues.apache.org/jira/browse/KAFKA-133. From what you had
mentioned, our understanding was that you have to wait for at least
one release to be able to publish to a Maven repo. Do you think this
blocks us ?

> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.

I'm thinking that since RC3 and the release tag is the same, we can
add the kafka-0.7.0-incubating tag and delete the
kafka-0.7.0-incubating-candidate-3 tag.

Let us know if this sounds good.

Thanks,
Neha

On Thu, Oct 13, 2011 at 9:26 PM, Alan D. Cabrera <li...@toolazydogs.com> wrote:
> The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.
>
> IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html
>
> Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.
>
>
> Regards,
> Alan
>
> On Oct 13, 2011, at 6:54 PM, Neha Narkhede wrote:
>
>> Hi,
>>
>> This is the third candidate for the first incubator release for Apache
>> Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
>> adds a standard .rat-excludes
>> files along with rat scripts to Kafka.
>>
>> It fixes the following issues:
>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html
>>
>> *** Please download, test and vote by noon Oct, 18th, 6 pm
>>
>> Note that we are voting upon the source (tag), binaries are provided
>> for convenience.
>>
>> Source and binary files:
>> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/
>>
>> The tag to be voted upon:
>> http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3
>>
>> Kafka's KEYS file containing PGP keys we use to sign the release:
>> http://svn.apache.org/repos/asf/incubator/kafka/KEYS
>>
>> Note that the Incubator PMC needs to vote upon the release after a
>> successful PPMC vote before
>> any release can be made official.
>>
>> Thanks,
>> Neha
>
>

Re: [VOTE] Release Kafka 0.7.0-incubating (candidate 3)

Posted by "Alan D. Cabrera" <li...@toolazydogs.com>.
The binary tgz should not be offered to assist in the review unless you're going to publish it as well, i.e. provide it as a download.  Signing it will only make people think you're going to publish it.  Not a showstopper.  Something to consider next time.

IIRC, the actual artifacts that get published to maven should also be available to review.  http://www.apache.org/dev/publishing-maven-artifacts.html

Is kafka-0.7.0-incubating-candidate-3 the intended final tag?  If not maybe it should just be kafka-0.7.0-incubating.  Once this goes out you can never change it.


Regards,
Alan

On Oct 13, 2011, at 6:54 PM, Neha Narkhede wrote:

> Hi,
> 
> This is the third candidate for the first incubator release for Apache
> Kafka, version 0.7.0-incubating. This fixes ALL rat failures and also
> adds a standard .rat-excludes
> files along with rat scripts to Kafka.
> 
> It fixes the following issues:
> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/RELEASE-NOTES.html
> 
> *** Please download, test and vote by noon Oct, 18th, 6 pm
> 
> Note that we are voting upon the source (tag), binaries are provided
> for convenience.
> 
> Source and binary files:
> http://people.apache.org/~nehanarkhede/kafka-0.7.0-incubating-candidate-3/
> 
> The tag to be voted upon:
> http://svn.apache.org/repos/asf/incubator/kafka/tags/kafka-0.7.0-incubating-candidate-3
> 
> Kafka's KEYS file containing PGP keys we use to sign the release:
> http://svn.apache.org/repos/asf/incubator/kafka/KEYS
> 
> Note that the Incubator PMC needs to vote upon the release after a
> successful PPMC vote before
> any release can be made official.
> 
> Thanks,
> Neha