[jira] [Updated] (CASSANDRA-20544) AST Harry's multi node tests are flakey when multi cell list happens cross instances

["David Capwell (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/CASSANDRA-20544?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

David Capwell updated CASSANDRA-20544:
--------------------------------------
     Bug Category: Parent values: Correctness(12982)Level 1 values: Test Failure(12990)
       Complexity: Low Hanging Fruit
    Discovered By: Fuzz Test
    Fix Version/s: NA
         Severity: Low
           Status: Open  (was: Triage Needed)

> AST Harry's multi node tests are flakey when multi cell list happens cross instances
> ------------------------------------------------------------------------------------
>
>                 Key: CASSANDRA-20544
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-20544
>             Project: Apache Cassandra
>          Issue Type: Bug
>          Components: Test/fuzz
>            Reporter: David Capwell
>            Assignee: David Capwell
>            Priority: Normal
>             Fix For: NA
>
>
> List append relies on TimeUUID.Generator.nextUnixMicro, which relies on the local clock’s millisecond clock.  In jvm-dtest all instances share the same physical clock, so most of the time time moves forward uniformly, but there is an edge case when the test “runs too fast” and creates a new TimeUUID within the same millisecond, when this happens a logically later timeuuid can happens-before a logically earlier timeuuid!



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@cassandra.apache.org
For additional commands, e-mail: commits-help@cassandra.apache.org