[
https://issues.apache.org/jira/browse/IGNITE-11029?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16753975#comment-16753975
]
Andrew Mashenkov commented on IGNITE-11029:
-------------------------------------------
[~Pavlukhin],
I've just one note: Seems, there is no need to mark
DeadlockDetectionManager.detectionStartDelay as volatile until you are going to
change its state concurrently. DeadlockDetectionManager is fully initialized
(start() method called) before node join a grid.
PR looks good.
> Public API for edge-chasing deadlock detection configuration
> ------------------------------------------------------------
>
> Key: IGNITE-11029
> URL: https://issues.apache.org/jira/browse/IGNITE-11029
> Project: Ignite
> Issue Type: Task
> Components: mvcc
> Reporter: Ivan Pavlukhin
> Assignee: Ivan Pavlukhin
> Priority: Major
> Fix For: 2.8
>
> Time Spent: 50m
> Remaining Estimate: 0h
>
> As deadlock detection introduce overhead during transaction processing some
> configuration means should be introduced. Following aspects should be
> addressed:
> 1. Turning detection on/off.
> 2. Adjusting a timeout before starting actual detection.
> It is proposed to introduce single property in {{TransactionConfiguration}}
> to address both concerns.
> Also it is desired that configuration is consistent across all cluster nodes
> (detection turned off on even one node can lead to undetected deadlocks).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
