[
https://issues.apache.org/jira/browse/IGNITE-6748?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16218252#comment-16218252
]
Alexey Goncharuk commented on IGNITE-6748:
------------------------------------------
In the scope of IGNITE-6030 it looks like there is no much sense to have
checkpoint page buffer size in DataStorageConfiguration. The checkpoint page
buffer size should be moved to DataRegionConfiguration, then the size
calculation will be straightforward.
> Invalid calculation of required node memory
> -------------------------------------------
>
> Key: IGNITE-6748
> URL: https://issues.apache.org/jira/browse/IGNITE-6748
> Project: Ignite
> Issue Type: Bug
> Security Level: Public(Viewable by anyone)
> Components: persistence
> Affects Versions: 2.3
> Reporter: Vasiliy Sisko
> Assignee: Alexey Goncharuk
>
> #
> org.apache.ignite.internal.managers.discovery.GridDiscoveryManager#requiredOffheap
> calculate size of system and configured data regions 2 times.
> #
> org.apache.ignite.internal.processors.cache.persistence.GridCacheDatabaseSharedManager#getDataStorageConfiguration
> initialize assign 2 variable by getDataStorageConfiguration result
> # When data regions are not configured
> org.apache.ignite.internal.processors.cache.persistence.GridCacheDatabaseSharedManager#checkpointBufferSize
> get DataStorageConfiguration.DFLT_DATA_REGION_MAX_SIZE constant instead of
> maximum size of default data region.
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
