Based on the discussions I had with Alex, it is clear that Harry has been a really valuable tool to find some issues that we would have otherwise missed. Part of the problem with CASSANDRA-16262 was in my opinion that we lacked experience with the tool itself and did not know where to start. Alex nicely proposed to take over the ownership of the ticket. Harry apparently has nearly all the functionality required to accomplish this. That sounds to me like a great proposal. :-) Thanks for that Alex
On Wed, Feb 17, 2021 at 10:39 AM Oleksandr Petrov < oleksandr.pet...@gmail.com> wrote: > In the absence of a fuzz testing tool I would probably support excluding > this ticket from GA, but speaking from recent experience it feels to me > that fixing bugs is blocking us from further fuzz testing more than fuzz > testing is blocking us from releases. In other words, you run a fuzz test > for a relatively short amount of time, and then spend several days to fix > and merge the issue that was triggered to unblock further testing. > > > In a relatively short amount of time, we've been able to hit at least these > four issues: > > - Group By in-jvm paging issue: > https://issues.apache.org/jira/browse/CASSANDRA-16427 > - Group By breaks range tombstone closer: > https://issues.apache.org/jira/browse/CASSANDRA-16431 > - Reverse iteration + paging: > https://issues.apache.org/jira/browse/CASSANDRA-16435 > - NPE in Slice#make on RT + partition deletion reconciliation: > https://issues.apache.org/jira/browse/CASSANDRA-16453 > > I think we would've hit the first one without a fuzz testing tool, since it > was a relatively obvious one, but looking at the output from Harry it was > immediately clear what's going on, so I still consider its output useful. > There was one more issue that was discovered independently, but was also > caught with a fuzz test tool later. Its prior discovery with a test tool > was blocked by the fix of an issue that was occurring more frequently. > > Maybe it's ok to release a release candidate, but we probably should still > wait a bit to cut a final release. WDYT? > > > > > On Mon, Feb 1, 2021 at 1:33 PM Benjamin Lerer <benjamin.le...@datastax.com > > > wrote: > > > It seems to me that CASSANDRA-16180 > > <https://issues.apache.org/jira/browse/CASSANDRA-16180>, CASSANDRA-16262 > > <https://issues.apache.org/jira/browse/CASSANDRA-16262> and > > CASSANDRA-16181 > > <https://issues.apache.org/jira/browse/CASSANDRA-16181> already highly > > improve the test coverage on distributed read and write paths. > > CASSANDRA-16262 sounds like a nice to have but should probably not block > > 4.0 GA > > I am +1 to push it out of 4.0-rc. > > > > Thanks to you and Andres for all the work you contributed on > > CASSANDRA-15579 > > <https://issues.apache.org/jira/browse/CASSANDRA-15579>. It is some > great > > work :-) > > > > On Thu, Jan 28, 2021 at 6:45 PM Caleb Rackliffe < > calebrackli...@gmail.com> > > wrote: > > > > > Hey everyone, > > > > > > I wanted to have a quick conversation about CASSANDRA-16262 > > > <https://issues.apache.org/jira/browse/CASSANDRA-16262>. As I > mentioned > > in > > > the Jira > > > < > > > > > > https://issues.apache.org/jira/browse/CASSANDRA-16262?focusedCommentId=17273891&page=com.atlassian.jira.plugin.system.issuetabpanels%3Acomment-tabpanel#comment-17273891 > > > > > > > ... > > > > > > *"I was chatting w/ Andres de la Peña > > > <https://issues.apache.org/jira/secure/ViewProfile.jspa?name=adelapena > > > > > the > > > other day, and it feels like there's an argument for allowing 4.0 to > > > release without this work being complete. We've certainly come a long > way > > > w/ CASSANDRA-15579 > > > <https://issues.apache.org/jira/browse/CASSANDRA-15579> already, > filling > > > in > > > a number of gaps that existed."* > > > > > > Any strong opinions out there? > > > > > > > > -- > alex p >
