[
https://issues.apache.org/jira/browse/CASSANDRA-10540?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14968973#comment-14968973
]
Marcus Eriksson commented on CASSANDRA-10540:
---------------------------------------------
Pushed a new branch
[here|https://github.com/krummas/cassandra/commits/marcuse/rangeawarecompaction]
for early feedback, still needs cleanup and tests
Enable like this:
{code}ALTER TABLE x.y WITH compaction={'class':'LeveledCompactionStrategy',
'range_aware_compaction':'true', 'min_range_sstable_size_in_mb':'15'}{code}
* Run a compaction strategy instance per owned range (with num_tokens=256 and
rf=3, we will have 768 * 2 instances (repaired/unrepaired data)).
* To avoid getting very many tiny sstables in the per-range strategies, we keep
them outside the strategy until the estimated size of a range-sstable is larger
than {{'min_range_sstable_size_in_mb'}}.
([estimation|https://github.com/krummas/cassandra/blob/09c58eb4689230d471ef4319733fb0e85399bd3a/src/java/org/apache/cassandra/db/compaction/writers/RangeAwareCompactionWriter.java#L115]
usually gets within a few % of the actual value).
* We do STCS among the many-range-sstables (called "L0" which might not be
optimal due to LCS)
* We currently prioritize compaction in L0 to get sstables out of there as
quickly as possible
* If an sstable fits within a range, it is added to that corresponding
range-compaction strategy - this should avoid getting a lot of L0 sstables
after streaming for example
* Adds a {{describecompactionstrategy}} nodetool command which displays
information about the configured compaction strategy (like sstables per range
etc). Example with only unrepaired data and 2 data directories - we first split
the owned ranges over those 2 directories, and then we split on a per range
basis, so the first RangeAwareCompactionStrategy is responsible for half the
data and the second one is responsible for the rest: {code} $ bin/nodetool
describecompactionstrategy keyspace1 standard1
-------------------------------------------------- keyspace1.standard1
--------------------------------------------------
Strategy=class org.apache.cassandra.db.compaction.RangeAwareCompactionStrategy,
for 167 unrepaired sstables, boundary tokens=min(-9223372036854775808) ->
max(-4095785201827646), location=/home/marcuse/c/d1
Inner strategy: class
org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy (257 instances,
162 total sstables)
sstable counts:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29
------------------------------------------------------------------------------------------
0.. 29 | 1 3 0 0 2 3 0 3 3 0 3 0 2 1 0 1 0 1 0 3 3 4 1
0 3 1 0 0 0 0
30.. 59 | 0 0 0 3 0 2 2 0 3 0 3 3 0 1 3 3 3 0 2 0 1 2 0
0 0 1 0 3 0 0
60.. 89 | 1 0 0 1 1 1 1 0 1 0 2 3 1 0 3 1 2 3 2 0 0 3 2
1 1 0 0 2 3 1
90..119 | 0 1 2 0 0 3 0 3 3 1 0 0 3 0 2 0 2 0 2 1 3 0 2
1 1 3 1 0 3 0
120..149 | 2 0 3 1 3 0 0 3 3 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
150..179 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
180..209 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
210..239 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
240..257 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Strategy=class org.apache.cassandra.db.compaction.RangeAwareCompactionStrategy,
for 221 unrepaired sstables, boundary tokens=max(-4095785201827646) ->
max(9223372036854775807), location=/var/lib/c1
Inner strategy: class
org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy (257 instances,
215 total sstables)
sstable counts:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29
------------------------------------------------------------------------------------------
0.. 29 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
30.. 59 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
60.. 89 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
90..119 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
120..149 | 0 0 0 0 0 0 0 0 0 0 1 6 0 0 3 0 3 0 3 3 3 3 1
0 1 0 2 0 3 2
150..179 | 3 3 3 0 0 3 3 0 3 2 3 1 3 3 3 3 0 0 0 3 0 1 1
0 6 3 3 0 3 3
180..209 | 0 1 1 3 1 3 1 3 3 2 3 3 0 3 0 3 1 0 0 1 2 3 0
0 1 1 0 0 3 3
210..239 | 3 3 3 2 0 6 1 3 0 0 3 3 3 1 3 4 3 3 3 0 3 0 3
1 2 2 0 2 0 0
240..257 | 1 0 3 1 0 3 3 0 0 0 0 0 0 3 3 0 0
Strategy=class org.apache.cassandra.db.compaction.RangeAwareCompactionStrategy,
for 0 repaired sstables, boundary tokens=min(-9223372036854775808) ->
max(-4095785201827646), location=/home/marcuse/c/d1
Inner strategy: class
org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy (257 instances,
0 total sstables)
sstable counts:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29
------------------------------------------------------------------------------------------
0.. 29 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
30.. 59 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
60.. 89 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
90..119 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
120..149 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
150..179 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
180..209 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
210..239 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
240..257 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Strategy=class org.apache.cassandra.db.compaction.RangeAwareCompactionStrategy,
for 0 repaired sstables, boundary tokens=max(-4095785201827646) ->
max(9223372036854775807), location=/var/lib/c1
Inner strategy: class
org.apache.cassandra.db.compaction.SizeTieredCompactionStrategy (257 instances,
0 total sstables)
sstable counts:
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22
23 24 25 26 27 28 29
------------------------------------------------------------------------------------------
0.. 29 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
30.. 59 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
60.. 89 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
90..119 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
120..149 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
150..179 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
180..209 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
210..239 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0
240..257 | 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
{code}
Comments/ideas/worries? [~yukim], [~kohlisankalp], [~iamaleksey], [~jbellis],
anyone?
> RangeAwareCompaction
> --------------------
>
> Key: CASSANDRA-10540
> URL: https://issues.apache.org/jira/browse/CASSANDRA-10540
> Project: Cassandra
> Issue Type: New Feature
> Reporter: Marcus Eriksson
> Assignee: Marcus Eriksson
> Fix For: 3.2
>
>
> Broken out from CASSANDRA-6696, we should split sstables based on ranges
> during compaction.
> Requirements;
> * dont create tiny sstables - keep them bunched together until a single vnode
> is big enough (configurable how big that is)
> * make it possible to run existing compaction strategies on the per-range
> sstables
> We should probably add a global compaction strategy parameter that states
> whether this should be enabled or not.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
