0052455: Import Entry processed twice when processing node is changed

Description

On a clustered environment with 2 application nodes.

If the node handling import entries is changed from node A to node B while node A is still processing Import Entries.

When the node B start processing it searches import entries and queue some that are also queued and to be processed in node A.

These Import Entries are processed in both nodes. When the Import Entry is 1st processed by node B and later in node A.

The Import Entry processed in node B ends with success.

The Import Entry processed in node A can fail generating a false error or duplicating a processing.

In case of POS requests like Order Loader the process generates a Error While Importing with an OutdatedDateException.
In case of a Integration request the process might send again the same data to the external system generating a duplicate.

Steps To Reproduce

Rough steps:
- Build a cluster environment with node A and B.
- Generate some Import Entries.
- When all Import Entries are queued in node A switch to node B.
- node B queue all the Import Entries again.
- Make the processing in node A slower so a Import Entry is processed first in node B
- Check that this import entry is processed again in node A with unpredictable results.

Proposed Solution

- Before starting to process an Import Entry check the status of the Import Entry to avoid processing again a Import Entry that it is not in Initial status anymore.

- Additionally consider to add a delay time when the handling node is switch before starting to search for new Import Entries to allow the node A to process queues entries.

Tags

No tags attached.

Attached Files

Relationships [ Relation Graph ] [ Dependency Graph ]

duplicate of	defect	0048454	closed	caristu	When a node is demoted as cluster node IE can be processed twice under some circumstances
related to	defect	0052457	closed	Triage Platform Conn	On cluster environments processing node can be switch even if current node is alive
related to	defect	0052445	closed	AugustoMauch	The margin when checking if we need to replace a cluster instance is too small
related to	defect	0052456	closed	caristu	Import Entry post process hook not executed when processing node is changed

Notes
There are no notes attached to this issue.

Issue History
Date Modified	Username	Field	Change
2023-05-16 09:45	gorkaion	New Issue
2023-05-16 09:45	gorkaion	Assigned To	=> Triage Platform Conn
2023-05-16 09:45	gorkaion	Modules	=> Core
2023-05-16 09:45	gorkaion	Triggers an Emergency Pack	=> No
2023-05-16 10:17	gorkaion	Issue cloned	0052456
2023-05-16 10:17	gorkaion	Relationship added	related to 0052456
2023-05-16 10:30	gorkaion	Relationship added	related to 0052457
2023-05-16 10:31	gorkaion	Relationship added	related to 0052445
2023-05-18 14:33	ngarcia	Issue Monitored: ngarcia
2023-05-18 17:09	adrianromero	Assigned To	Triage Platform Conn => caristu
2023-05-19 11:36	caristu	Relationship added	related to 0048454
2023-05-22 11:45	caristu	Relationship deleted	related to 0048454
2023-05-22 11:46	caristu	Relationship added	duplicate of 0048454
2023-05-22 11:46	caristu	Review Assigned To	=> caristu
2023-05-22 11:46	caristu	Status	new => closed
2023-05-22 11:46	caristu	Resolution	open => duplicate