Thanks for raising the question, Benjamin! Notes on a few tickets inline below.
Non-Blocking: – CASSANDRA-15537 Local Read/Write Path: Upgrade and Diff Test I think it’s reasonable to consider this ticket complete. Yifan and others have worked to execute several dozen diff tests and while I’m sure others will continue, it’s reasonable to say cassandra-diff has been used to compare 3.0 vs. 4.0 clusters with a wide variety of data models. I’ll check with Yifan on Tuesday re: updating the status of the ticket. It would be wonderful to hear of diff runs and experience from additional contributors if others can share. – CASSANDRA-15584 Tooling - External Ecosystem Great collaboration on this one (including issues filed arising from this coverage, such as a recent ticket related to Medusa). Blocking GA: – CASSANDRA-15579 Distributed Read/Write Path The coordination and replication subtasks (16180, 16181) are making good progress. I’ll check with Caleb and David on 16262 (the fuzz testing subtask on Tuesday). – CASSANDRA-15581 Compaction Most of these are perf tests rather than development tasks, though the ones complete are listed as Patch Available. I’ll check with Yifan if it’d make sense to move those for which no planned work remains to Resolved. I don’t think there’s a lot left here. – CASSANDRA-15538 Local Read/Write Path - Other Areas Will see if anything specific is planned, as scope is relatively undefined. With the exception of 15538, most of these look to be moving along or nearly complete. I don’t think I’d shift others aside from it into the non-blocking category - but the silver lining is that it shouldn’t be long before the others wrap up. We do have several flaky test tickets that could use attention, though — these may be quick to push through if anyone is able to pick them up: – CASSANDRA-16236: Fix flaky testTrackMaxDeletionTime – CASSANDRA-16238: Fix flaky test test_insert_data_during_replace_same_address - replace_address_test.TestReplaceAddress – CASSANDRA-16239: Fix flaky test org.apache.cassandra.distributed.test.NetstatsRepairStreamingTest testWithCompressionDisabled – CASSANDRA-16317: Fix flaky test incompleteCommit - org.apache.cassandra.distributed.test.CASTest – CASSANDRA-16355: Fix flaky test incompletePropose - org.apache.cassandra.distributed.test.CASTest – CASSANDRA-16382: Fix flaky LongSharedExecutorPoolTest.testPromptnessOfExecution – CASSANDRA-16358: Minor Flakiness in ProxyHandlerConnectionsTest#testExpireSomeFromBatch – CASSANDRA-16229: Flaky jvm-dtest: org.apache.cassandra.distributed.test.ring.NodeNotInRingTest.nodeNotInRingTest – CASSANDRA-16061: transient_replication_ring_test.py::TestTransientReplicationRing::test_move_forwards_and_cleanup Cheers, – Scott > On Jan 14, 2021, at 9:05 AM, Benjamin Lerer <benjamin.le...@datastax.com> > wrote: > > Hi everybody, > > As discussed before the holidays, it might make sense to revisit the scope > of the quality testing tickets for 4.0 GA to ensure that the 4.0 release is > not held for longer than necessary. > > The current status of the quality testing tasks are the following: > > *DONE:* > > * CASSANDRA-15583 <https://issues.apache.org/jira/browse/CASSANDRA-15583> > Tooling, Bundled and First Party* > CASSANDRA-15586 <https://issues.apache.org/jira/browse/CASSANDRA-15586> > Cluster Setup and Maintenance > CASSANDRA-15587 <https://issues.apache.org/jira/browse/CASSANDRA-15587> > Platforms and Runtimes > > > *NON BLOCKING:* > > The goals of the following ticket have been reached. Once GA is closed they > will be marked as done. > > CASSANDRA-15537 <https://issues.apache.org/jira/browse/CASSANDRA-15537> > Local Read/Write Path: Upgrade and Diff Test > CASSANDRA-15584 <https://issues.apache.org/jira/browse/CASSANDRA-15584> > Tooling - External Ecosystem > > If I understood Jordan comment correctly on the following ticket, its > should also not be a blocker for 4.0 > CASSANDRA-15585 <https://issues.apache.org/jira/browse/CASSANDRA-15585> > Test Frameworks, Tooling, Infra / Automation > > *BLOCKING GA:* > > CASSANDRA-15579 <https://issues.apache.org/jira/browse/CASSANDRA-15579> > Distributed Read/Write Path > 4 sub-tasks: 1 resolved, 2 in progress, 1 open > > CASSANDRA-15580 <https://issues.apache.org/jira/browse/CASSANDRA-15580> > Repair > Test scenarios are ready, working on integrating them to circle-ci > > CASSANDRA-15581 <https://issues.apache.org/jira/browse/CASSANDRA-15581> > Compaction > 9 sub-tasks: 5 patch available, 1 review in progress, 3 triage needed > > CASSANDRA-15582 <https://issues.apache.org/jira/browse/CASSANDRA-15582> > Metrics > 16 sub-tasks: 9 resolved, 5 patch available, 5 open > > CASSANDRA-15588 <https://issues.apache.org/jira/browse/CASSANDRA-15588> > Cluster Upgrade > 6 sub-tasks: 4 resolved, 1 in progress, 1 open > CASSANDRA-15538 <https://issues.apache.org/jira/browse/CASSANDRA-15538> > Local Read/Write Path No progress has been made on that ticket. The > conclusion so far is that Harry is our best choice to uncover issues in > that area but there is no clear plan on how to move forward. > We have made some progress across the quality testing tickets. Nevertheless > there is still a significant amount of tickets to fix. As our time and > resources are limited it might make sense to focus on what we believe are > the most critical for 4.0 and relax our constraints on others. For example > it seems to me that the metrics tickets will mainly help to discover non > critical old issues that are not blockers for 4.0. It is clear to me that > they should be fixed but that could probably be done for the 4.0.x/4.1 > release (I fully volunteer for that :-)). The same could be true for some > other areas of the code. > > In my opinion the important questions we would need to answer are: > > 1. Are there some tickets that we should make non-blocking for 4.0 ? > 2. What do we do about CASSANDRA-15538 > <https://issues.apache.org/jira/browse/CASSANDRA-15538> Local Read/Write > Path? > > Thanks in advance for your feedback :-)
