[
https://issues.apache.org/jira/browse/IGNITE-11536?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16800177#comment-16800177
]
Ivan Rakov commented on IGNITE-11536:
-------------------------------------
[~antonovsergey93], your PR has conflicts with master. Also a few comments:
1. GridCacheDatabaseSharedManager: new field
{code:java}
private final LongJVMPauseDetector pauseDetector;
{code}
2. Do we really need to explicitly pass instance of LongJVMPauseDetector to
every needed grid component? There's only one instance per node anyway - we can
obtain it via getter from kernal context instead.
> Add information about possible long GC pause to checkpoint started message.
> ---------------------------------------------------------------------------
>
> Key: IGNITE-11536
> URL: https://issues.apache.org/jira/browse/IGNITE-11536
> Project: Ignite
> Issue Type: Improvement
> Reporter: Sergey Antonov
> Assignee: Sergey Antonov
> Priority: Major
> Fix For: 2.8
>
> Time Spent: 1h 40m
> Remaining Estimate: 0h
>
> One of possible reasons long hold of write lock on checkpoint is GC pause.
> Now you must check logs around on {{Possible too long JVM pause}} message. We
> should print possible long GC pause time in {{Checkpoint started}} message,
> if threshold was passed.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
