[jira] [Commented] (CASSANDRA-13764) SelectTest.testMixedTTLOnColumnsWide is flaky
[ https://issues.apache.org/jira/browse/CASSANDRA-13764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16146196#comment-16146196 ] Joel Knighton commented on CASSANDRA-13764: --- The idea looks sound to me - I think you need to remove the TTL column from the select [here|https://github.com/jeffjirsa/cassandra/commit/a1e49db69622de11a996d09105e5ebf3b54c58c3#diff-7f5981228f9d9428fb164aa91316aa85R2976], as you did in {{testMixedTTLOnColumnsWide}}. If you agree, I'm comfortable with you doing that on commit and don't need to rereview if CI looks good. > SelectTest.testMixedTTLOnColumnsWide is flaky > - > > Key: CASSANDRA-13764 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13764 > Project: Cassandra > Issue Type: Bug > Components: Testing >Reporter: Joel Knighton >Assignee: Jeff Jirsa >Priority: Trivial > > {{org.apache.cassandra.cql3.validation.operations.SelectTest.testMixedTTLOnColumnsWide}} > is flaky. This is because it inserts rows and then asserts their contents > using {{ttl()}} in the select, but if the test is sufficiently slow, the > remaining ttl may change by the time the select is run. Anecdotally, > {{testSelectWithAlias}} in the same class uses a fudge factor of 1 second > that would fix all the failures I've seen, but it might make more sense to > measure the elapsed time in the test and calculate the acceptable variation > from that time. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13764) SelectTest.testMixedTTLOnColumnsWide is flaky
[ https://issues.apache.org/jira/browse/CASSANDRA-13764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16145997#comment-16145997 ] Jeff Jirsa commented on CASSANDRA-13764: [~slebresne] says he's ok with ninja'ing this in, but in case you ( [~jkni] ) want to review before I do it (I need to push for Circle anyway just to be safe, so this'll sit here until Circle gives me a green run): 3.0: https://github.com/jeffjirsa/cassandra/tree/cassandra-3.0-13764 (Circle: https://circleci.com/gh/jeffjirsa/cassandra/tree/cassandra-3.0-13764 ) 3.11: https://github.com/jeffjirsa/cassandra/tree/cassandra-3.11-13764 (Circle: https://circleci.com/gh/jeffjirsa/cassandra/tree/cassandra-3.11-13764 ) trunk: https://github.com/jeffjirsa/cassandra/tree/cassandra-13764 (Circle: https://circleci.com/gh/jeffjirsa/cassandra/tree/cassandra-13764 ) > SelectTest.testMixedTTLOnColumnsWide is flaky > - > > Key: CASSANDRA-13764 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13764 > Project: Cassandra > Issue Type: Bug > Components: Testing >Reporter: Joel Knighton >Assignee: Jeff Jirsa >Priority: Trivial > > {{org.apache.cassandra.cql3.validation.operations.SelectTest.testMixedTTLOnColumnsWide}} > is flaky. This is because it inserts rows and then asserts their contents > using {{ttl()}} in the select, but if the test is sufficiently slow, the > remaining ttl may change by the time the select is run. Anecdotally, > {{testSelectWithAlias}} in the same class uses a fudge factor of 1 second > that would fix all the failures I've seen, but it might make more sense to > measure the elapsed time in the test and calculate the acceptable variation > from that time. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org
[jira] [Commented] (CASSANDRA-13764) SelectTest.testMixedTTLOnColumnsWide is flaky
[ https://issues.apache.org/jira/browse/CASSANDRA-13764?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16127438#comment-16127438 ] Joel Knighton commented on CASSANDRA-13764: --- This also affects {{SelectTest.testMixedTTLOnColumns}}. > SelectTest.testMixedTTLOnColumnsWide is flaky > - > > Key: CASSANDRA-13764 > URL: https://issues.apache.org/jira/browse/CASSANDRA-13764 > Project: Cassandra > Issue Type: Bug > Components: Testing >Reporter: Joel Knighton >Priority: Trivial > > {{org.apache.cassandra.cql3.validation.operations.SelectTest.testMixedTTLOnColumnsWide}} > is flaky. This is because it inserts rows and then asserts their contents > using {{ttl()}} in the select, but if the test is sufficiently slow, the > remaining ttl may change by the time the select is run. Anecdotally, > {{testSelectWithAlias}} in the same class uses a fudge factor of 1 second > that would fix all the failures I've seen, but it might make more sense to > measure the elapsed time in the test and calculate the acceptable variation > from that time. -- This message was sent by Atlassian JIRA (v6.4.14#64029) - To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org For additional commands, e-mail: commits-h...@cassandra.apache.org