incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Marvin Humphrey <>
Subject Re: Sentry 1.5.0 release (was Re: [RESULT] [VOTE] Apache TinkerPop 3.0.0.M8-incubating Release (Part Deux))
Date Fri, 29 May 2015 16:48:09 GMT
On Fri, May 29, 2015 at 5:26 AM, Joe Brockmeier <> wrote:
> On Sun, May 10, 2015, at 03:20 PM, Marvin Humphrey wrote:
>> > Apache Sentry
>> This one looks like a problem.  A vote was taken for Sentry 1.5.0 on the
>> podling dev list, but there has been no vote on general@incubator.
>> It would be perfectly understandable for an individual release manager to
>> misunderstand the incubating release process, but it's troubling that not
>> a single podling community member spoke up.  What's going on?
> Apologies for just noticing this email, and for missing that. I'm
> looking into this, I think there may have been a misunderstanding about
> the need for IPMC votes vs. actually conducting a vote on general@.
> (IIRC all three binding +1s are IPMC members.)

Thanks, Joe.

To clarify and elaborate:

The Apache Incubator is a "top-level project" at the Apache Software
Foundation -- like Apache Tomcat, like Apache Lucene, like Apache HTTPD. The
Incubator makes official Apache releases following a specific process, just
like those other top-level projects do.

Release candidates prepared by incubating podlings become official Apache
releases when the Incubator releases them. This requires a vote by the
Incubator PMC.

It is not valid for a subset of any Apache PMC to approve a release candidate
without giving the entire PMC sufficient opportunity for review.  This
prevents, for example, a handful of employees from ExampleCo making a release
of Apache Foo that other members of the Foo PMC find objectionable.

While the Sentry release glitch appears to have arisen from a
misunderstanding, there have been multiple occasions in the ASF's history
where failure to understand or uphold release protocols has resulted in
distorted project governance and disenfranchised PMC members.

>From time to time, you will hear people argue that the incubation process
should be reorganized and that the release votes on general@incubator should
be streamlined away.  Such arguments would be more persuasive, IMO, if it were
not routine to catch problems during the general@incubator vote phase.  In any
case, that is not the situation that exists today.

So we now have Sentry artifacts being distributed which have not been approved
by the full Incubator PMC.  A similar situation with another project was once
resolved by having the full PMC vote on the already-distributed artifacts.  In
my view, the Sentry podling should initiate such a vote, and a passing vote
would settle the issue.

Marvin Humphrey

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message