[
https://issues.apache.org/jira/browse/CASSANDRA-14556?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16558442#comment-16558442
]
ASF GitHub Bot commented on CASSANDRA-14556:
--------------------------------------------
Github user iamaleksey commented on a diff in the pull request:
https://github.com/apache/cassandra/pull/239#discussion_r205504297
--- Diff:
src/java/org/apache/cassandra/db/streaming/CassandraStreamHeader.java ---
@@ -65,18 +85,43 @@ private CassandraStreamHeader(Version version,
SSTableFormat.Type format, long e
this.compressionInfo = compressionInfo;
this.sstableLevel = sstableLevel;
this.header = header;
-
+ this.fullStream = fullStream;
+ this.componentManifest = componentManifest;
+ this.firstKey = firstKey;
+ this.tableId = tableId;
this.size = calculateSize();
}
- public CassandraStreamHeader(Version version, SSTableFormat.Type
format, long estimatedKeys, List<SSTableReader.PartitionPositionBounds>
sections, CompressionMetadata compressionMetadata, int sstableLevel,
SerializationHeader.Component header)
+ private CassandraStreamHeader(Version version, SSTableFormat.Type
format, long estimatedKeys,
--- End diff --
The introduction of the new fields and constructors got us to 5
constructors total with up to 10 arguments, which is no longer manageable, and
calls for a builder. It's boring and tedious work, so I did it myself and
pushed here -
https://github.com/iamaleksey/cassandra/commit/321d21747faa46afcf34518ebdeb811f2a805de8
- please feel free to cherry-pick.
In addition to introducing the builder, the commit renames `fullStream` to
something a bit more meaningful (`isEntireSSTable`) that clearly reflects
what's actually happening, fixes a bug in `serializedSize()` where compression
info isn't initialized, and removes some fields without `toString()`
implementations from header's own `toString()`.
> Optimize streaming path in Cassandra
> ------------------------------------
>
> Key: CASSANDRA-14556
> URL: https://issues.apache.org/jira/browse/CASSANDRA-14556
> Project: Cassandra
> Issue Type: Improvement
> Components: Streaming and Messaging
> Reporter: Dinesh Joshi
> Assignee: Dinesh Joshi
> Priority: Major
> Labels: Performance
> Fix For: 4.x
>
>
> During streaming, Cassandra reifies the sstables into objects. This creates
> unnecessary garbage and slows down the whole streaming process as some
> sstables can be transferred as a whole file rather than individual
> partitions. The objective of the ticket is to detect when a whole sstable can
> be transferred and skip the object reification. We can also use a zero-copy
> path to avoid bringing data into user-space on both sending and receiving
> side.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
