[
https://issues.apache.org/jira/browse/IGNITE-20766?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17789952#comment-17789952
]
Vladislav Pyatkov commented on IGNITE-20766:
--------------------------------------------
Merged da2a4f374af4503e158a91828f0a800d5af93b73
> Deal with the primary replica provider in integration tests
> -----------------------------------------------------------
>
> Key: IGNITE-20766
> URL: https://issues.apache.org/jira/browse/IGNITE-20766
> Project: Ignite
> Issue Type: Bug
> Reporter: Kirill Tkalenko
> Assignee: Vladislav Pyatkov
> Priority: Major
> Labels: ignite-3
> Time Spent: 50m
> Remaining Estimate: 0h
>
> In the process of implementing IGNITE-20678, it turned out that using
> *TestPlacementDriver* in tests *ItRebalanceDistributedTest* and
> *ItIgniteNodeRestartTest* does not look correct, since for each node the
> primary replica will be defined as the node (primary replica) itself, which
> can lead to unpredictable behavior, we need to figure this out.
> Definition of done
> Both thests have an integration nature; the cluster contains more than one
> node, and the nodes use the network. In these circumstances, the primary
> replica can be changed during an integration test and is not predetermined.
> # I believe the test implementation of the placement drive is not applicable
> here; this is a better way to instance the production implementation. Look at
> the PR attached.
> # Or it would be better to rewrite the test to use the original Ignite node
> instead of the particular instantiated nodes.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
