[
https://issues.apache.org/jira/browse/IGNITE-27999?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18068215#comment-18068215
]
Nikita Amelchev commented on IGNITE-27999:
------------------------------------------
Should be fixed after IGNITE-27631. It is recommended to check the issue.
> FullMessage consume huge heap allocation during snapshot creation
> -----------------------------------------------------------------
>
> Key: IGNITE-27999
> URL: https://issues.apache.org/jira/browse/IGNITE-27999
> Project: Ignite
> Issue Type: Improvement
> Components: general
> Affects Versions: 2.17
> Reporter: Evgeny Stanilovsky
> Priority: Major
> Labels: ignite-2, ise
>
> During snapshot creation
> {noformat}
> org.apache.ignite.internal.util.distributed.FullMessage
> {noformat}
> will be send through discovery. It will conatain
> {noformat}
> SnapshotPartitionsQuickVerifyHandler
> {noformat}
> with Map<PartitionKey, PartitionHashRecord> through all snapshot involved
> caches - which can consume huge (10 - 100 megabytes) if sufficient (>100)
> caches are involved. Seems it need to be reduced.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
