[ 
https://issues.apache.org/jira/browse/CASSANDRA-14956?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16757661#comment-16757661
 ] 

Joseph Lynch commented on CASSANDRA-14956:
------------------------------------------

Turns out it was just that in trunk we don't include the port in the 
[cassandra-murmur3.yaml|https://github.com/apache/cassandra/blob/7f634feb7cf1fdb135133946ffd75efa681b8cb7/test/conf/cassandra-murmur.yaml#L27]
 so the check 
[here|https://github.com/apache/cassandra/blob/7f634feb7cf1fdb135133946ffd75efa681b8cb7/src/java/org/apache/cassandra/service/StorageService.java#L556]
 fails and we actually try to gossip with peers. A quick patch to just include 
the port appears to fix the issue. Other tests are broken but PagingTest is now 
fixed:

 
||trunk||
|[bb4b0626|https://github.com/jolynch/cassandra/commit/bb4b06265cfc7a9cf915e8022feb9ccd83390d5d]|
|[!https://circleci.com/gh/jolynch/cassandra/tree/CASSANDRA-14956-fix-unit.png?circle-token=
 
1102a59698d04899ec971dd36e925928f7b521f5!|https://circleci.com/gh/jolynch/cassandra/tree/CASSANDRA-14956-fix-unit]|

Test failures:
{{org.apache.cassandra.transport.frame.checksum.ChecksummingTransformerTest#corruptionCausesFailure}}:
 Pretty sure this is CASSANDRA-15002, debugging that separately
{{org.apache.cassandra.distributed.DistributedReadWritePathTest#readRepairTest}}:
 I believe this is CASSANDRA-14922, debugging that separately

> Paged Range Slice queries with DISTINCT can drop rows from results
> ------------------------------------------------------------------
>
>                 Key: CASSANDRA-14956
>                 URL: https://issues.apache.org/jira/browse/CASSANDRA-14956
>             Project: Cassandra
>          Issue Type: Bug
>          Components: CQL/Interpreter
>            Reporter: Sam Tunnicliffe
>            Assignee: Sam Tunnicliffe
>            Priority: Major
>             Fix For: 2.1.21, 2.2.14
>
>
> If we have a partition where the first CQL row is fully deleted (possibly via 
> TTLs), and that partition happens to fall on the page boundary of a paged 
> range query which is using SELECT DISTINCT, the next live partition *after* 
> it is omitted from the result set. This is due to over fetching of the pages 
> and a bug in trimming those pages where overlap occurs.
> This does not affect 3.0+.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to