+1 on both matters.
As far as what goes in, I feel strongly that FileChannel needs to be
stabilized(see FLUME-1232 we have had multiple people reporting the
same issue in different scenarios). We should also probably remove the
hadoop.io dependencies on FileChannel and ResumableMemoryChannel. The
JDBC channel is just not efficient enough, and I think that having
a(more or less) reliable channel for high throughput applications is
important.
Other than that, I think we're in a pretty good place right now. I'm
working on the guava service issue which is rather far-reaching and
would definitely be better suited for the next version.
On 06/13/2012 05:03 PM, Jarek Jarcec Cecho wrote:
+1 on making next stable release (1.2.)
+1 for Mike to be the release manager
Jarcec
On Jun 13, 2012, at 9:57 AM, Mike Percy wrote:
Developers,
It has been nearly 3 months since we released Flume 1.1.0. Since the 1.1.0 release, we
have had 117 commits to the trunk, and 114 JIRAs have been marked "Resolved
Fixed" with a target of v1.2.0. This flurry of activity represents a whole slew of
new features and a bunch of bug fixes.
Because of the above, I think we should release a Flume 1.2.0 to the world
within 2-3 weeks. I propose that I be the release manager for this release. To
that end, I would like to cut an SVN branch pretty soon to stabilize and test
the code in advance of the release.
In addition, I'd like to gather input on what other devs would like to see
included in a 1.2.0 release.
Aside from a few JIRAs already marked as Patch Available, the one thing I'd
really like to make happen is a once-over on the user guide to cover the latest
features. I'm willing to help.
Regards,
Mike
