[ https://issues.apache.org/jira/browse/CASSANDRA-10762?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Ryan McGuire reopened CASSANDRA-10762: -------------------------------------- > select_distinct_with_deletions_test failing in mixed version cluster > -------------------------------------------------------------------- > > Key: CASSANDRA-10762 > URL: https://issues.apache.org/jira/browse/CASSANDRA-10762 > Project: Cassandra > Issue Type: Sub-task > Components: CQL > Reporter: Philip Thompson > Assignee: Sylvain Lebresne > Fix For: 3.0.3, 3.3 > > > The dtests > {{upgrade_tests.cql_tests.TestCQLNodes3RF3.select_distinct_with_deletions_test}} > and > {{upgrade_tests.cql_tests.TestCQLNodes2RF1.select_distinct_with_deletions_test}} > are failing. > While in a mixed node state, they do {{'SELECT DISTINCT k FROM t1'}} on this > table > {{CREATE TABLE t1 (k int PRIMARY KEY, c int, v int)}} > More rows than expected are returned. Here is the sample output of one of > those queries. As you can see, there are duplicate values of k returned in a > DISTINCT k query. Seems like a bug when communicating between the 2.2 and 3.0 > node. > {code} > [Row(k=5), Row(k=1), Row(k=1), Row(k=8), Row(k=8), Row(k=2), Row(k=2), > Row(k=4), Row(k=4), Row(k=7), Row(k=7), Row(k=6), Row(k=6), Row(k=9), > Row(k=9), Row(k=3), Row(k=3)] > {code} -- This message was sent by Atlassian JIRA (v6.3.4#6332)