[
https://issues.apache.org/jira/browse/CASSANDRA-21224?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Marcus Eriksson updated CASSANDRA-21224:
----------------------------------------
Test and Documentation Plan: ci run
Status: Patch Available (was: Open)
https://github.com/apache/cassandra/pull/4679
> Cost of deserializing cluster metadata may become prohibitive with a large
> number of peers
> ------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-21224
> URL: https://issues.apache.org/jira/browse/CASSANDRA-21224
> Project: Apache Cassandra
> Issue Type: Improvement
> Components: Transactional Cluster Metadata
> Reporter: Marcus Eriksson
> Assignee: Marcus Eriksson
> Priority: Normal
> Fix For: 6.x
>
>
> When building the Directory component of ClusterMetadata, the entire set of
> nodes is traversed to calculate the min, max and common serialization
> versions. Deserializing a directory from a set of bytes (e.g. when reading a
> snapshot from the system table or off the wire) is inefficient and creates a
> new temporary Directory instance each time a node is added. In large clusters
> this may have a noticeable effect on loading time.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
