There are 3 view filtering test failures in the 8 4.1 failures which are
known offenders as being parametrized they might take long in an
overloaded systems. Once we're back to normal capacity on jenkins they
should be ok, finger in the air estimate, so we might be even better
than 8. #collaborating
On 14/6/22 17:22, Josh McKenzie wrote:
How has it been two weeks already?
4.1 is high priority, so let's start there:
https://butler.cassandra.apache.org/#/
Last run had 8 failed tests! We've had a pretty painful run in terms
of CI infrastructure over the past week on 4.1 (see here:
https://ci-cassandra.apache.org/job/Cassandra-4.1/) - the usual
culprit of running out of disk space on a variety of agents (15, 14,
9, 9), but looks like we're back in a healthy place with a couple
successful builds now.
In terms of "0 test flakes" as beta blocker, here's our status:
https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2454
<https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2454>
28 issues in Beta, 6 of them in need of reviewers. If you're a
committer and have some spare minutes, please grab one of those
tickets for review.
We're down from 40 tickets blocking beta to 28 as of the past two
weeks - good trajectory. Won't quite put us at ready to GA in early
July but we shouldn't be too far off.
[New contributor Getting Started]
10 of those beta tickets are unassigned and are test fixes which are a
great way to get started on the project! Link:
https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2454&quickFilter=2160
<https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2454&quickFilter=2160>
Reference the "Getting started" guide on our website for help with
setting up your environment:
https://cassandra.apache.org/_/development/index.html, and the dev
community hangs out on https://the-asf.slack.com in #cassandra-dev.
There's 13 of us that have volunteered to help mentor you can reach
with the @cassandra_mentors alias in the chat room, or just float any
questions you have to the channel at large as we're a 24/7 type of
community around the world.
[Dev list Digest]
https://lists.apache.org/list?dev@cassandra.apache.org:lte=2w:
The highlight this past couple weeks has been discussion around the
right way to shape an API around multi-key transactions in Cassandra
for CEP-15:
https://lists.apache.org/thread/5sds3968mnnk42c24pvgwphg4qvo2xk0.
There's been a lot of great back and forth here, and the more
perspectives we can get on this thread and about this API the better,
so if you have some opinions about the shape a transactional / batch
API in Cassandra should look, please engage!
I'm going to jump off the email list to JIRA for this one - we've had
a discussion ongoing about when we cut a Major vs. a Minor, what
qualifies as an API, etc on CASSANDRA-16844
(https://issues.apache.org/jira/browse/CASSANDRA-16844). Expect
something to formally hit the dev mailing list about this soon, but
until then we can keep going on the JIRA ticket.
A question floated to the dev mailing list about slow unit tests with
Cassandra 4.x on macOS:
https://lists.apache.org/thread/8lr5pcrqqxm1fltgzok2b2j8jo04zskh
And last but not least, we've been without a formal Build Lead for a
couple of weeks now. This is a great way to get involved with the
project and get more exposure to the CI/CD system, as well as help us
drive to a 4.1 GA. The Wiki covering the build lead role can be found
here:
https://cwiki.apache.org/confluence/display/CASSANDRA/Build+Lead. You
can expect it to take a minimum of a few hours of your entire week to
keep up with new test failure ticket creation and butler linkage, and
it's been a huge help in keeping us on track to hit near our 12 month
commitment between releases.
[CI Trends]
https://butler.cassandra.apache.org/#/
All our branches excepting trunk are slowly trending down; this makes
sense as we burn down test failures for 4.1 as they're often
multi-branch impacting.
3.0: 13
3.11: 23
4.0: 2
4.1: 8
trunk: 20
66 in total, so one of the lowest marks we've seen across all our
branches. Curious to see how low the entire body of tests gets when we
hit 0 on 4.1, and to have some discussions about ways we can keep it
there.
[Release progress]
https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2175
<https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2175>
4.1 alpha:
3 test issues closed out
4.1 beta:
10+ test failure fixes. An issue that could lead to disk read errors
during repair session failure was resolved in CASSANDRA-17663
4.1 rc / ga:
Test fix, an NPE in SimpleBuilders was addressed in CASSANDRA-17596,
and a slight tightening of our consideration of escaped unicode
control characters in CASSANDRA-17617
It's great to see this momentum on burning down these test failures.
Let's keep it going!
~Josh
