[
https://issues.apache.org/jira/browse/CASSANDRA-19988?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sam Lightfoot updated CASSANDRA-19988:
--------------------------------------
Summary: Direct IO support for background reads (was: Direct IO support
for streaming)
> Direct IO support for background reads
> --------------------------------------
>
> Key: CASSANDRA-19988
> URL: https://issues.apache.org/jira/browse/CASSANDRA-19988
> Project: Apache Cassandra
> Issue Type: Improvement
> Reporter: Jon Haddad
> Priority: Normal
>
> * Extend Direct IO support for compaction reads to other background read
> operations
> * Introduce a new _background_read_disk_access_mode_ property to supercede
> _compaction_read_disk_access_mode_
>
> ||Component||Impact||Effort||
> |Streaming outgoing (writers)|Hundreds of GBs during bootstrap/repair|Medium
> — need to plumb {{DiskAccessMode}} through {{ChannelProxy}} and
> {{ComponentContext}}|
> |Scrub|Full SSTable scan|Low — pass {{DiskAccessMode}} to {{openDataReader}}|
> |Verify|Full SSTable scan|Low — pass {{DiskAccessMode}} to {{openDataReader}}|
> |Index rebuild (SAI/SASI/Accord)|Full SSTable scan per table|Low — pass
> {{DiskAccessMode}} to {{openDataReader}}|
> |Streaming incoming (mutation-based)|Re-scan of received SSTables|Low — pass
> {{DiskAccessMode}} to {{getScanner}}|
> |Expired SSTable compaction|Small SSTables|Low — pass {{DiskAccessMode}} to
> {{getScanner}}|
> |Accord journal scan|Background scan|Low — pass {{DiskAccessMode}} to
> {{getScanner}}|
> |Hints replay|Sequential file reads|Medium — different I/O path, not
> SSTable-based|
> |Compression dict sampling|Random reads|Low benefit — random access pattern|
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
