[
https://issues.apache.org/jira/browse/HDDS-12502?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Wei-Chiu Chuang resolved HDDS-12502.
------------------------------------
Resolution: Implemented
Resolving as implemented/superseded by the Snapshot Defragmentation project.
SnapshotDefragService performs full defragmentation and compacts the processed
snapshot tables with bottommost compaction forced to remove tombstones. The
repository also includes SnapshotCache.compactSnapshotDB(), which compacts
tables not tracked in the snapshot-diff DAG when configured. The original
request to reclaim space in non-diff snapshot tables is therefore covered by
the implemented snapshot maintenance mechanisms.
Evidence at d0949f291791c399f3e54409fcb92656a80597a4:
*
https://github.com/apache/ozone/blob/d0949f291791c399f3e54409fcb92656a80597a4/hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/snapshot/defrag/SnapshotDefragService.java#L265
*
https://github.com/apache/ozone/blob/d0949f291791c399f3e54409fcb92656a80597a4/hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/snapshot/defrag/SnapshotDefragService.java#L300
*
https://github.com/apache/ozone/blob/d0949f291791c399f3e54409fcb92656a80597a4/hadoop-ozone/ozone-manager/src/main/java/org/apache/hadoop/ozone/om/snapshot/SnapshotCache.java#L88
The SnapshotCache non-diff-table compaction option defaults to false; this
resolution does not claim automatic compaction is enabled by default.
> Enable compaction on all other snapshot tables excepting for
> KeyTable/FileTable/Directory
> -----------------------------------------------------------------------------------------
>
> Key: HDDS-12502
> URL: https://issues.apache.org/jira/browse/HDDS-12502
> Project: Apache Ozone
> Issue Type: Sub-task
> Reporter: Swaminathan Balachandran
> Priority: Major
>
> Leaving FileTable, DirectoryTable and keyTable we should enable compaction on
> all other tables. If there is a snapshot where there is a lot of keys that
> can be reclaimed there would be significant amount of data that would get
> added to deletedKeyTable and deletedDirectory table and would be eventually
> removed as well. We should allow compactions to reduce the rocksdb size.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
