[
https://issues.apache.org/jira/browse/IGNITE-10003?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16728369#comment-16728369
]
Alexey Goncharuk commented on IGNITE-10003:
-------------------------------------------
[~agura], [~andrey-kuznetsov], I agree that we need another failure type, but I
do not like {{SYSTEM_CRITICAL_OPERATION_FAILED}} - semantically it is
incorrect, checkpoint read lock timeout does not necessarily mean the operation
failed - our recent load tests show this clearly. I agree with
[~andrey-kuznetsov] here that {{SYSTEM_CRITICAL_OPERATION_TIMEOUT}} is a better
fit here. For example, in future we may use the same failure type for PME
timeouts.
> Raise SYSTEM_WORKER_BLOCKED instead of CRITICAL_ERROR when checkpoint read
> lock timeout detected
> ------------------------------------------------------------------------------------------------
>
> Key: IGNITE-10003
> URL: https://issues.apache.org/jira/browse/IGNITE-10003
> Project: Ignite
> Issue Type: Task
> Affects Versions: 2.7
> Reporter: Andrey Kuznetsov
> Assignee: Andrey Kuznetsov
> Priority: Trivial
> Fix For: 2.8
>
>
> {{GridCacheDatabaseSharedManager#failCheckpointReadLock}} should report
> {{SYSTEM_WORKER_BLOCKED}} to failure handler: it is closer to the truth and
> default consequenses are not so severe as opposed to {{CRITICAL_ERROR}}.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
