I tried reverting it and it still failed. Furthermore, I tried a Maven
clean/test/verify from 1.7.0 with a binDir set to 1.7.1-RC1’s release build and
it still failed. The same Maven build with 1.7.0’s release build failed too now
though, so it seems this is either a flaky test, or more likely m
Hm, that's interesting. KUDU-2379 is the only one I can think of that might
have caused this issue. If you revert that change, does it fix the test?
On Wed, May 9, 2018 at 11:46 AM, Attila Bukor wrote:
> Hi Todd,
>
> Yes, Kudu Client tests all pass on 1.7.0 (although I tried it only once)
> and
Hi Todd,
Yes, Kudu Client tests all pass on 1.7.0 (although I tried it only once) and it
fails consistently on 1.7.1-RC1. Also I’m not convinced this would be a
test-only issue as none of these changes seem to have touched the tests between
the releases:
$ git log --oneline 1.7.0..1.7.1-RC1 --
Hey Attila,
Is this a regression from 1.7.0 for you? If not, I don't think it should
block the maintenance release. We can always do more maintenance releases
to address further bugs.
Additionally my guess is this is a test-only issue since ITClientStress
does some strange stuff reaching into Net
I made a mistake and ran the Java tests agains another binDir. With 1.7.1-RC1
release build I actually have a Kudu Java Client error:
Tests run: 1, Failures: 0, Errors: 0, Skipped: 0, Time elapsed: 61.261 sec - in
org.apache.kudu.client.ITClient
Results :
Failed tests:
ITClientStress.testMan
+1 (non-binding)
Built on CentOS 7.3 in release mode as well. I had one C++ test failure when I
built this with Ninja due to it running "which ninja" and on this machine it's
called ninja-build. I was able to work around it by adding a symlink called
ninja to the $PATH, that way the actual test
+1
I built on CentOS 7.3 and ran all C++ tests in release mode and with TSAN.
All passed.
On Wed, May 9, 2018 at 12:08 AM, Mike Percy wrote:
> I hit that issue too, but running
>
> unset GLOG_colorlogtostderr
>
> Made those tests pass for me as well.
>
> Mike
>
> Sent from my iPhone
>
> > On
Thanks, Todd.
I did not notice this comment. Sorry for that:(
何李夫
2017-04-10 16:06:24
-邮件原件-
发件人: dev-return-5994-hzhelifu=corp.netease@kudu.apache.org
代表 Todd Lipcon
发送时间: 2018年5月8日 23:03
收件人: dev
主题: Re: A question about the culling row-sets in
’Tablet::CaptureConsistentIterator
+1
* built in debug/release mode on debian8.7, all tests passed.
何李夫
2017-04-10 16:06:24
-邮件原件-
发件人: dev-return-5996-hzhelifu=corp.netease@kudu.apache.org
代表 Hao Hao
发送时间: 2018年5月9日 5:07
收件人: dev@kudu.apache.org
主题: Re: [VOTE] Apache Kudu 1.7.1 RC1
+1
* Release note looks good
*
I hit that issue too, but running
unset GLOG_colorlogtostderr
Made those tests pass for me as well.
Mike
Sent from my iPhone
> On May 8, 2018, at 11:40 PM, Alexey Serbin wrote:
>
> +1
>
> I built 1.7.1 RC1 (release) on CentOS 6.6.
>
> All C++ tests passed except for the following ones:
>
10 matches
Mail list logo