Excellent :)
Makes me wonder if we could actually determine when DNS is hosed when
launching MiniAccumuloCluster and fail in an obvious manner..
Aaron D. Mihalik wrote:
Builds on my machine were failing and I was concerned that builds.a.o was
uncovering additional tests that were going to fail.
Turns out, this is not the case. All tests pass on my local machine I'm
continuing the process to cut a RC now.
I was having issues running the Mini Accumulo Cluster tests on my local
machine because my DNS set up was screwy. That's fixed now and the tests
pass.
--Aaron
On Wed, Sep 7, 2016 at 12:41 PM Josh Elser<[email protected]> wrote:
IMO, a flakey test should just be disabled until someone makes time to
fix it. But, why is builds.a.o blocking you from making a release?
You'll be building the release locally anyways [1].
[1] http://www.apache.org/dev/release.html#owned-controlled-hardware
Aaron D. Mihalik wrote:
Caleb and I went to cut a RC yesterday, and it seems like the build is in
an unstable state. It seems like a strange, non-deterministic test
failure
that might be related to the builder server. I've documented it here:
RYA-171 - Mongo Tests are not Cleaning Up
https://issues.apache.org/jira/browse/RYA-171
Should we continue with the 3.2.10-RC or wait until this is fixed?
Also, the Apache Build server is slammed. Yesterday it took 1.5 hours to
get a machine to start our build :/
--Aaron
