Cassandra project biweekly status update 2022-05-17

[Josh McKenzie <jm...@apache.org>]

The ApacheCon CFP for 2022 closes in 6 days! If you have something you've been thinking about presenting (about Cassandra or another Apache project), an interesting use-case, some fascinating work you did or production lessons you learned in the past year, join us at ApacheCon 2022 in New Orleans from October 3 to October 6 and tell us about it! You can submit CFP's at the following URL: https://apachecon.com/acna2022/cfp.html

The ApacheCon site can be found here: https://www.apachecon.com/acna2022/

We're making steady progress during the freeze on cassandra-4.1. Here's a couple high level links and insights:

* butler CI dashboard for 4.1 (16 failures): https://butler.cassandra.apache.org/#/
* 4.1 release blockers (44): https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2455&quickFilter=2454
* 4.1 tickets closed in the past 2 weeks (13): https://issues.apache.org/jira/issues/?jql=project%20%3D%20cassandra%20and%20fixversion%20%3D%204.1%20and%20resolution%20!%3D%20unresolved%20and%20resolved%20%3E%20-2w

[New contributor Getting Started]
:wave: Welcome! Glad to have you join us. :D

We use JIRA for our work tracking, and as mentioned above we're in a pre-GA freeze period where we're trying to stabilize our release; your help would be greatly appreciated and we have interesting work that needs to be done that's well suited to a newcomer.

We have 27 unassigned tickets that are currently blocking the 4.1 release, many of them failing or flaky tests which are a great place to get started with the project. If you're interested, here's a convenient link to unassigned 4.1 blockers: https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2455&quickFilter=2454&quickFilter=2160

Join us on slack at the apache server at https://the-asf.slack.com, channel #cassandra-dev, and feel free to ping @cassandra_mentors alias if you need help deciding where to get started.


[Dev list Digest]
https://lists.apache.org/list?dev@cassandra.apache.org:lte=2w:

I'll repeat the call to action: if you have tickets that you think should be blockers for cassandra-4.1, please take a few minutes to update the FixVersion on the tickets to 4.1-alpha, 4.1-beta, or 4.1-rc as appropriate. Email thread: https://lists.apache.org/thread/pzbrnmhk3749g831g3w8bmstq7z97783, cwiki article on which release work fits into: https://cwiki.apache.org/confluence/display/CASSANDRA/Release+Lifecycle

Mick's been a busy bee with all the releases over the past couple of weeks; we have 4.0.4, 3.11.13, and 3.0.27 that all went through the vote and release process.

Benedict revisited the code contribution and style guide and there was some discussion back and forth about annotation usage: https://lists.apache.org/thread/fh1ylkp7wwsr3b3sbnym9gl3fl9l70tt

We had an early draft review of a blog post I authored about the path to green CI you can find here: https://lists.apache.org/thread/796wcpnlpm63zov5s5f7c8xqogmv24c8 as well as a couple other blog post emails that have gone live.

Ekaterina reached out about CASSANDRA-15234, CASSANDRA-17562, and the future of JMX and virtual tables: https://lists.apache.org/thread/lnm5ly4tzyzlf4zcmsg9okx6rgxmf0rr


[CI Trends]
Butler dashboard: https://butler.cassandra.apache.org/#/

We continue to have some pain here with our CI infrastructure largely centered around agents running out of disk space. Brandon has a ticket open with ASF Infra to try and get us physical access to a representative sample of our agents to try and root cause why disk space issues are persisting.

For those that don't know, we have 5 different sets of machines donated by different parties that make up our CI infrastructure (list available here: https://github.com/apache/cassandra-builds/blob/trunk/ASF-jenkins-agents.md#current-agents). While these agents are uniform in their memory availability, their processing environments and disk space can differ which has been leading to some repeated headaches managing this cluster of machines. To make matters more complex, we don't have physical console access to the boxes and have to go through some fairly obtuse mechanisms to run simple commands on them, hence working with infra as mentioned above.

Trunk remains at around 20 failures, 4.0 in the single digits, and 4.1 has been bouncing from low teens up to a high of 40 when we had some infrastructure troubles.

As I mention in the blog post that should be posted within the next week, we're adding more and more tests over time and holding steady with a relatively low number of test failures; given the diversity in our commit gatekeepers (circle mid, high, or ASF Jenkins), our merge strategy (multi-branch, not blocking merge on CI), and the complexity of the space we're working in, we _still_ effectively have fewer and fewer failures relative to our total test load as time goes by; this is something we should be proud of!

This last mile of getting to green and staying there is definitely tough but we're making progress.


[Release progress]
https://issues.apache.org/jira/secure/RapidBoard.jspa?rapidView=484&quickFilter=2175

4.0.x:
7 issues closed out, website, release cutting, and some test fixes. Nothing disruptive, as expected.

4.1:
19 issues closed out. This work is dominated by test fixes as we expect at this point in our release cycle. I renamed a guardrail I'd added so we didn't end up married to a restrictive API / config param name (thanks again Andres!), David fixed a regression introduced into the jvm-dtest framework causing it to drop unhandled exceptions rather than failing tests on them in CASSANDRA-17549, and he also took care of a regression in our handling of expected exceptions during repair in CASSANDRA-17620

Thanks everyone for your continued focus on stabilizing 4.1; it's great to see this chunk of work shrink as we drive towards GA!

~Josh