[
https://issues.apache.org/jira/browse/IGNITE-22938?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Pavel Tupitsyn updated IGNITE-22938:
------------------------------------
Epic Link: IGNITE-19479
> Thin 3.0: Improve getAll and putAll partition awareness
> -------------------------------------------------------
>
> Key: IGNITE-22938
> URL: https://issues.apache.org/jira/browse/IGNITE-22938
> Project: Ignite
> Issue Type: Improvement
> Components: thin client
> Reporter: Pavel Tupitsyn
> Assignee: Pavel Tupitsyn
> Priority: Major
> Labels: ignite-3
> Fix For: 3.0.0-beta2
>
>
> Currently, *getAll* and *putAll* [use the first key for partition
> awareness|https://github.com/apache/ignite-3/blob/9b271581a7ed3d8d6d620af465404b387bfdd84d/modules/client/src/main/java/org/apache/ignite/internal/client/table/ClientRecordBinaryView.java#L103].
> This causes extra network hops when other keys in the list belong to
> different nodes.
> * Client should split the data set into per-node sets and send multiple
> requests in parallel. This should improve performance significantly for large
> sets.
> * To ensure atomicity, use client observable timestamp. Talk to
> [~ascherbakov] for more details.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
