[
https://issues.apache.org/jira/browse/CASSANDRA-9830?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15064573#comment-15064573
]
Paulo Motta commented on CASSANDRA-9830:
----------------------------------------
Rebased and submitted cstar_perf tests comparing LCS with default settings vs
enabled {{skip_top_level_bloom_filter}} option on
[ssd|http://cstar.datastax.com/tests/id/35023b30-a5bc-11e5-aa2d-0256e416528f]
and
[hdd|http://cstar.datastax.com/tests/id/3d14a8ac-a5bd-11e5-aa2d-0256e416528f].
Will report back when results are ready and maybe play around with parameters
(sstable size, bloom filter fp chance, etc) after initial results.
> Option to disable bloom filter in highest level of LCS sstables
> ---------------------------------------------------------------
>
> Key: CASSANDRA-9830
> URL: https://issues.apache.org/jira/browse/CASSANDRA-9830
> Project: Cassandra
> Issue Type: New Feature
> Components: Compaction
> Reporter: Jonathan Ellis
> Assignee: Paulo Motta
> Priority: Minor
> Labels: performance
> Fix For: 3.2
>
>
> We expect about 90% of data to be in the highest level of LCS in a fully
> populated series. (See also CASSANDRA-9829.)
> Thus if the user is primarily asking for data (partitions) that has actually
> been inserted, the bloom filter on the highest level only helps reject
> sstables about 10% of the time.
> We should add an option that suppresses bloom filter creation on top-level
> sstables. This will dramatically reduce memory usage for LCS and may even
> improve performance as we no longer check a low-value filter.
> (This is also an idea from RocksDB.)
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
