[
https://issues.apache.org/jira/browse/CASSANDRA-14873?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Aleksey Yeschenko updated CASSANDRA-14873:
------------------------------------------
Description: If a partition has a static row and is large enough to be
indexed, then {{firstName}} of the first index block will be set to a static
clustering. When deserializing the column index we then incorrectly deserialize
the {{firstName}} as a regular, non-{{STATIC}} {{Clustering}} - a singleton
array with an empty {{ByteBuffer}} to be exact. Depending on the clustering
comparator, this can trip up binary search over {{IndexInfo}} list and cause an
incorrect resultset to be returned. (was: TBD)
> Fix missing rows when reading 2.1 SSTables with static columns in 3.0
> ---------------------------------------------------------------------
>
> Key: CASSANDRA-14873
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14873
> Project: Cassandra
> Issue Type: Bug
> Reporter: Aleksey Yeschenko
> Assignee: Aleksey Yeschenko
> Priority: Major
> Fix For: 3.0.x, 3.11.x
>
>
> If a partition has a static row and is large enough to be indexed, then
> {{firstName}} of the first index block will be set to a static clustering.
> When deserializing the column index we then incorrectly deserialize the
> {{firstName}} as a regular, non-{{STATIC}} {{Clustering}} - a singleton array
> with an empty {{ByteBuffer}} to be exact. Depending on the clustering
> comparator, this can trip up binary search over {{IndexInfo}} list and cause
> an incorrect resultset to be returned.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@cassandra.apache.org
For additional commands, e-mail: commits-h...@cassandra.apache.org
