[
https://issues.apache.org/jira/browse/CASSANDRA-7409?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14198665#comment-14198665
]
Carl Yeksigian commented on CASSANDRA-7409:
-------------------------------------------
I've finished running a performance test; it's available
[here|http://cstar.datastax.com/tests/id/162b4540-6386-11e4-a260-bc764e04482c].
The branches tested are:
- Overlapping LCS
- New LCS, which does not use STCS in L0, but selects the compactions in a
different order
- Old LCS with STCS
- Old LCS without STCS
The time to complete the test for each:
|| || Overlapping LCS || New LCS || Old LCS/STCS || Old LCS ||
| Operations | 1:05:47 | 1:00:18 | 1:24:23 | 1:15:57 |
| Compactions to 0 | 4:59:21 | 9:56:11 | 1:12:59 | 1:42:33 |
Unfortunately, the improvements to LCS cause the compaction time (to get to 0
pending compactions) to be significantly worse. Going through the logs, it
seems that there are many fewer timeouts generated by the new configurations,
so it is providing better latency.
> Allow multiple overlapping sstables in L1
> -----------------------------------------
>
> Key: CASSANDRA-7409
> URL: https://issues.apache.org/jira/browse/CASSANDRA-7409
> Project: Cassandra
> Issue Type: Improvement
> Reporter: Carl Yeksigian
> Assignee: Carl Yeksigian
> Labels: compaction
> Fix For: 3.0
>
>
> Currently, when a normal L0 compaction takes place (not STCS), we take up to
> MAX_COMPACTING_L0 L0 sstables and all of the overlapping L1 sstables and
> compact them together. If we didn't have to deal with the overlapping L1
> tables, we could compact a higher number of L0 sstables together into a set
> of non-overlapping L1 sstables.
> This could be done by delaying the invariant that L1 has no overlapping
> sstables. Going from L1 to L2, we would be compacting fewer sstables together
> which overlap.
> When reading, we will not have the same one sstable per level (except L0)
> guarantee, but this can be bounded (once we have too many sets of sstables,
> either compact them back into the same level, or compact them up to the next
> level).
> This could be generalized to allow any level to be the maximum for this
> overlapping strategy.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
