[ https://issues.apache.org/jira/browse/CASSANDRA-15879?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17138670#comment-17138670 ]
Caleb Rackliffe edited comment on CASSANDRA-15879 at 6/17/20, 5:45 PM: ----------------------------------------------------------------------- So {{CorruptedSSTablesCompactionsTest}} was inserting rows with invalid {{LongType}} partition keys (i.e. keys with length 1, rather than 0 or 8). That wasn't really a problem by itself, but the seed {{9823169134884L}} somehow generated some duplicate clusterings, and when {{DuplicateRowChecker}} is notified of the partition closing in {{onPartitionClose()}}, it [tries to print|https://github.com/apache/cassandra/blob/cassandra-3.0/src/java/org/apache/cassandra/db/transform/DuplicateRowChecker.java#L93] a human-readable partition key, which of course fails validation. Simply fixing {{CorruptedSSTablesCompactionsTest}} to generate valid partition keys seems to solve the problem. ||3.0||3.11||trunk|| |patch|https://github.com/apache/cassandra/pull/638/commits/17a7842dd6f71935f8e5c5b8f7c0a1dd7e85c566]|[patch|https://github.com/apache/cassandra/pull/639]|[patch|https://github.com/apache/cassandra/pull/640]| |[utest|https://app.circleci.com/pipelines/github/maedhroz/cassandra/8/workflows/c4704991-07c8-473b-b99e-97d93b4c5f90]|[utest|https://app.circleci.com/pipelines/github/maedhroz/cassandra/9/workflows/14c42f46-bb04-488f-a2d8-2718d600b848]|[utest|https://app.circleci.com/pipelines/github/maedhroz/cassandra?branch=CASSANDRA-15879-trunk]| was (Author: maedhroz): So {{CorruptedSSTablesCompactionsTest}} was inserting rows with invalid {{LongType}} partition keys (i.e. keys with length 1, rather than 0 or 8). That wasn't really a problem by itself, but the seed {{9823169134884L}} somehow generated some duplicate clusterings, and when {{DuplicateRowChecker}} is notified of the partition closing in {{onPartitionClose()}}, it [tries to print|https://github.com/apache/cassandra/blob/cassandra-3.0/src/java/org/apache/cassandra/db/transform/DuplicateRowChecker.java#L93] a human-readable partition key, which of course fails validation. Simply fixing {{CorruptedSSTablesCompactionsTest}} to generate valid partition keys seems to solve the problem. ||3.0||3.11||trunk|| |[patch|https://github.com/apache/cassandra/pull/638/commits/17a7842dd6f71935f8e5c5b8f7c0a1dd7e85c566]|[patch|https://github.com/apache/cassandra/pull/639]|[patch|https://github.com/apache/cassandra/pull/640]| > Flaky unit test: > BlacklistingCompactionsTest.testBlacklistingWithSizeTieredCompactionStrategy > --------------------------------------------------------------------------------------------- > > Key: CASSANDRA-15879 > URL: https://issues.apache.org/jira/browse/CASSANDRA-15879 > Project: Cassandra > Issue Type: Bug > Components: Test/unit > Reporter: Caleb Rackliffe > Assignee: Caleb Rackliffe > Priority: Normal > Fix For: 3.0.x, 3.11.x, 4.0-beta > > Time Spent: 0.5h > Remaining Estimate: 0h > > CASSANDRA-14238 addressed the failure in > {{BlacklistingCompactionsTest.testBlacklistingWithSizeTieredCompactionStrategy}}, > but only on 2.2. While working on CASSANDRA-14888, we’ve reproduced [the > failure|https://app.circleci.com/pipelines/github/dineshjoshi/cassandra/47/workflows/de5f7cdb-06b6-4869-9d19-81a145e79f3f/jobs/2516/tests] > on trunk. > It looks like this should be a clean merge forward. -- This message was sent by Atlassian Jira (v8.3.4#803005) --------------------------------------------------------------------- To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org