[
https://issues.apache.org/jira/browse/GEODE-10067?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17508632#comment-17508632
]
ASF subversion and git services commented on GEODE-10067:
---------------------------------------------------------
Commit ffde8c788e97a09203dd1eb9385f2355f9c408e9 in geode's branch
refs/heads/develop from Alberto Gomez
[ https://gitbox.apache.org/repos/asf?p=geode.git;h=ffde8c7 ]
GEODE-10067: Optimize wan-copy region command (#7420)
When the region to copy is partitioned and
the gateway sender is parallel, each gateway
sender sends the primary copies of the
buckets in that server.
Deserialization of the entries is also avoided.
> WANCopyRegionFunctionDelegate needs to be optimized to handle large regions
> ---------------------------------------------------------------------------
>
> Key: GEODE-10067
> URL: https://issues.apache.org/jira/browse/GEODE-10067
> Project: Geode
> Issue Type: Improvement
> Components: wan
> Affects Versions: 1.15.0
> Reporter: Udo Kohlmeyer
> Assignee: Alberto Gomez
> Priority: Major
> Labels: pull-request-available
> Fix For: 1.15.0
>
>
> The current WanCopyRegionFunctionDelegate may cause significant memory issues
> in with really large regions.
> The
> [getEntries|https://github.com/apache/geode/blob/develop/geode-wan/src/main/java/org/apache/geode/management/internal/cli/functions/WanCopyRegionFunctionDelegate.java#L102]
> method returns both primary and redundant Region Entries for local server.
> The invocation of getEntries from the PartitionedRegionDataStore will cause
> all entries to be deserialized
> https://github.com/apache/geode/blob/develop/geode-core/src/main/java/org/apache/geode/internal/cache/PartitionedRegionDataStore.java#L2501-L2515
> The
> [createBatch|https://github.com/apache/geode/blob/develop/geode-wan/src/main/java/org/apache/geode/management/internal/cli/functions/WanCopyRegionFunctionDelegate.java#L107-L108]
> method creates a List<GatewayQueueEvent> equally that of the local Region
> size.
> Essentially ... In a system with VERY large regions this might cause
> significant memory issues
>
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
