Details
-
Task
-
Resolution: Done
-
L3 - Default
-
None
-
None
Description
Context:
In Elasticsearch between the get and indexing phases of the update, it is possible that another process might have already updated the same document. To make sure that there is a retry, the retry_on_conflict parameter was used. However, when changing to the high level rest client, we had to remove this parameter, since it is not support the way the high level rest client adds this parameter. In Elasticsearch 6.0.0 it is expected that the parameter is called _retry_on_conflict and from 6.1.0 it is called retry_on_conflict. Once we only allow Elasticsearch version 6.1+, we should add the parameter again. Unfortunately, there is no way to change that request manually.
AT:
- when there are conflict while updating a document, it is retried to update the document
- the value hard coded
- in particular the retry mechanism should be applied for the following resources:
- process definitions
- process definition xmls
- variables
- (running/completed) activity instances
- (running/completed) process instances
- decision definitions
- decision definition xmls
- the number of retries constant is removed from the configuration, which also includes removing it from the documentation
mgm-controller-panel
This is the controller panel for Smart Panels app
Attachments
Issue Links
- mentioned in
-
Page Loading...