[
https://issues.apache.org/jira/browse/IGNITE-19682?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17845191#comment-17845191
]
Pavel Tupitsyn commented on IGNITE-19682:
-----------------------------------------
Merged to main:
[1c0ee275624f07b1dab1a2c0a0307ae74ab0a579|https://github.com/apache/ignite-3/commit/1c0ee275624f07b1dab1a2c0a0307ae74ab0a579]
> .NET: Thin 3.0: Tx partition awareness
> --------------------------------------
>
> Key: IGNITE-19682
> URL: https://issues.apache.org/jira/browse/IGNITE-19682
> Project: Ignite
> Issue Type: Improvement
> Components: thin client
> Affects Versions: 3.0.0-beta1
> Reporter: Pavel Tupitsyn
> Assignee: Pavel Tupitsyn
> Priority: Major
> Labels: .NET, ignite-3, performance
> Fix For: 3.0.0-beta2
>
> Time Spent: 20m
> Remaining Estimate: 0h
>
> Currently, client sends a separate *TX_BEGIN* request when the user invokes
> *ITransactions.BeginAsync* API:
> * Extra network request.
> * Chosen tx coordinator (server node that handles TX_BEGIN request) is random
> and in most cases won't be the primary node for enlisted keys.
> Solution:
> * On the client, do not send *TX_BEGIN* request when the user invokes
> *ITransactions.BeginAsync*. Instead, start the tx "on demand" when it is
> first used in some API.
> * Send two requests at once to the same node where the first enlisted
> operation goes (according to partition awareness, if applicable).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
