Some Eclipse Foundation services are deprecated, or will be soon. Please ensure you've read this important communication.

Bug 317034

Summary:	Block Read And Write Operations Until Discovery Had Finished If Cache Coordination Is Configured
Product:	z_Archived	Reporter:	Reinhard Girstenbrei <reinhard.girstenbrei>
Component:	Eclipselink	Assignee:	Project Inbox <eclipselink.foundation-inbox>
Status:	NEW ---	QA Contact:
Severity:	enhancement
Priority:	P3
Version:	unspecified
Target Milestone:	---
Hardware:	All
OS:	All
Whiteboard:

Description Reinhard Girstenbrei

2010-06-16 08:00:30 EDT

Build Identifier: 2.0.2.v20100323-r6872

EclipseLink allows to execute read and write operations immediately after EclipseLink session login.
This can lead to problems with stale objects if cache coordination is configured.
A example scenario that would run into problems with stale objects is described in "Steps to Reproduce".

Reproducible: Always

Steps to Reproduce:
Problem Scenario:
1. Cluster Node C1 starts EclipseLink session S1.
2. S1 sends announcement but does not get a response because there is no other session taking part in the coordinated cache yet.
3. Cluster node C2 starts EclipseLink session S2.
4. C2 reads entity objects using S2.
5. C1 updates the same entity objects using S1, but does not propagate changes as S2 had not send the announcement yet.
6. S2 sends announcement and joins the coordinated cache.
6.1. Now S2 has stale instances of the entities updated by S1 in the shared cache.
6.2. Updating this objects will fail with a lock exception if locking had been configured.

Comment 1 Eclipse Webmaster

2022-06-09 10:07:28 EDT

The Eclipselink project has moved to Github: https://github.com/eclipse-ee4j/eclipselink