[
https://issues.apache.org/jira/browse/IGNITE-23553?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17893801#comment-17893801
]
Kirill Tkalenko commented on IGNITE-23553:
------------------------------------------
Looks good.
> Only notify FailureHandler about first watch notification failure
> -----------------------------------------------------------------
>
> Key: IGNITE-23553
> URL: https://issues.apache.org/jira/browse/IGNITE-23553
> Project: Ignite
> Issue Type: Improvement
> Reporter: Roman Puchkovskiy
> Assignee: Roman Puchkovskiy
> Priority: Major
> Labels: ignite-3
> Time Spent: 40m
> Remaining Estimate: 0h
>
> When a failure happens in any watch, any subsequent handling of any revision
> or idle safe time propagation event will fail with the same original
> exception, and each time the FailureHandler is notified. This creates a storm
> of logging if the FH does not halt the node immediately.
> We should only notify FH about first such a failure (and also write to the
> logs that no more watch notifications will happen).
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
