incubator-general mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Kevan Miller <>
Subject Re: [VOTE] Release Kafka 0.7.1-incubating (Candidate 2)
Date Tue, 19 Jun 2012 14:31:57 GMT

On Jun 18, 2012, at 9:34 PM, Joe Stein wrote:

> I uploaded a md5 checksum
> to
> go along with the detached signature that was generated already


> starting the kafka server is in the quick start

I did track that doc down.

I find it easier to read that info in the download. Rather than bring up a README and then
have to visit a web site (which may not be documenting the kafka version that I'm trying to
start). But that's just my opinion and it's a minor point… I'm happy for the Kafka project
to decide how they document their project...

Here's what I got trying to start, yesterday:

[2012-06-18 16:52:09,788] INFO Opening socket connection to server localhost/0:0:0:0:0:0:0:1:2181
[2012-06-18 16:52:09,798] WARN Session 0x0 for server null, unexpected error, closing socket
connection and attempting reconnect (org.apache.zookeeper.ClientCnxn) Connection refused
	at Method)
	at org.apache.zookeeper.ClientCnxn$

Seems to be working today. Don't know what's changed… Again not a big deal. You guys are
in control of these issues/non-issues. When reviewing a release, I try to build (and start)
a project.

> looking at Cassandra and Hadoop source distributions both have jars in them
> too, so not sure the issue you are raising others might have a better
> understanding or I am missing something you are trying to point out

We (the ASF) are not always terribly consistent. And there could be some debate on this. And
I'm not sure there's official documented policy around this… Finally, I confess that in
the past, I've been part of a project that has released a jar file(s) as part of a source
release. However, IMO, this was a mistake. And since corrected…  

Apache releases are *source* releases. See --
"All releases are in the form of the source materials needed to make changes to the software
being released." 

IMO, we should not be including jar's, dll's, and other binary artifacts in our source releases.
You may find some alternate opinions. I'm not aware of documentation that makes this explicit…

Build utilities (e.g. ant, maven, sbt, ivy) are usually a requirement to build a release.
Not included in the release. I don't know why you are including the other jar/dll files in
your source. But would encourage you to stop...

> I would like to keep the RC2 voting open for another few days to see if
> this release passes (as folks have not had time to review) or we can get to
> what issues we need to have resolved that can be done for RC3 for a new
> vote either with objective to getting 0.7.1 released.

You are the release manager. You're in control of the release timing/process. It's certainly
frustrating to have a release vote, a problem found, and then have people hold off reviewing
the release until there's a new RC. So, encouraging additional reviews/votes is quite understandable.

Also, there aren't VETO's of release votes. So my vote does not necessarily mean your release
vote fails.

FYI, out-of-date copyright statements are not typically considered a blocking issue for a

> current votes:
> +1 [1] +0 [0] -1 [0] from the PPMC
> and
> +1 [0] +0 [0] -1 [0] from the IPMC
> please let me know if there is something different we want or should have
> been doing here or that I can be or should be doing differently.

I see that I didn't register my vote in my original email. Here's mine -- -1 (binding). 

Binding just means that I'm a member of the Incubator PMC. Beyond the requirement for 3 +1
IPMC votes, it doesn't count more or less than any other votes… There aren't veto's of release
votes. As a release manager, you can decide if problems found merit a new release candidate.

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

View raw message