[
https://issues.apache.org/jira/browse/IGNITE-23483?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Vladimir Steshin resolved IGNITE-23483.
---------------------------------------
Resolution: Resolved
> Restore snapshot check from lesser topology for snapshot check distributed
> process.
> -----------------------------------------------------------------------------------
>
> Key: IGNITE-23483
> URL: https://issues.apache.org/jira/browse/IGNITE-23483
> Project: Ignite
> Issue Type: Bug
> Reporter: Vladimir Steshin
> Assignee: Vladimir Steshin
> Priority: Major
> Labels: ise
> Time Spent: 1h 20m
> Remaining Estimate: 0h
>
> Before the refactoring of snapshot validation, a single node had the ability
> to check several snaphot parts if they were placed on the same node. As a
> result, a snapshot can be checked by just one node, can be checked by
> insufficient topology and can be restored from other topologies.
> This was achieved by mapping of available snapshot metadatas to actual online
> topology. See _AbstractSnapshotVerificationTask#map(...)_. Now, we check only
> the snapshot data related to the current node despite we read more than one
> meta and there aren't enough online nodes.
> _testMetaWarningRestoredByOnlyOneNode()_ indicates the problem. A broken
> snapshot is sucessfully validated because node doesn't receive all the
> available snapshot parts.
> We should restore the feature of checking several snapshot parts on one node.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
