[ https://issues.apache.org/jira/browse/CASSANDRA-5492?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13639071#comment-13639071 ]
Sylvain Lebresne commented on CASSANDRA-5492: --------------------------------------------- bq. but I think the code removal is fairly straightforward I wouldn't go all the way to 'straighforward', it did took a few minutes to convince myself that the changes around skipBloomFilter were breaking any backward compatibility :). That being said, I did convinced myself, so +1. Nits: * In SSTableWriter, when creating the RIE, maybe we could add a comment saying we discard the range tombstones because they have already been persisted during the column indexing. * In RIE.serialize, we could move the write of the "promoted size" outside of the 'if (rie.isIndexed())' since promotedSize() returns 0 for non-indexed entries. > Backport row-level bloom filter removal to 1.2 > ---------------------------------------------- > > Key: CASSANDRA-5492 > URL: https://issues.apache.org/jira/browse/CASSANDRA-5492 > Project: Cassandra > Issue Type: Improvement > Components: Core > Affects Versions: 1.2.0 > Reporter: Jonathan Ellis > Assignee: Jonathan Ellis > Priority: Minor > Fix For: 1.2.5 > > > With the possible presence of range tombstones, it is erroneous to skip > checking for a given column in SSTableNamesIterator because the bloom filter > says it is not there. -- This message is automatically generated by JIRA. If you think it was sent incorrectly, please contact your JIRA administrators For more information on JIRA, see: http://www.atlassian.com/software/jira