[
https://issues.apache.org/jira/browse/CASSANDRA-20421?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Andy Tolbert updated CASSANDRA-20421:
-------------------------------------
Fix Version/s: 5.1-alpha1
Source Control Link: https://github.com/apache/cassandra/pull/3598
Resolution: Fixed
Status: Resolved (was: Ready to Commit)
Merged into [trunk_cep_37|https://github.com/apache/cassandra/pull/3598] (set
version to 5.1-alpha1 since it was required but may not strictly be true as
this is contingent on CEP-37 being merged into trunk)
> Improve auto repair documentation and use more conservative default for
> bytes_per_assignment for full repair
> ------------------------------------------------------------------------------------------------------------
>
> Key: CASSANDRA-20421
> URL: https://issues.apache.org/jira/browse/CASSANDRA-20421
> Project: Apache Cassandra
> Issue Type: Task
> Components: Documentation
> Reporter: Andy Tolbert
> Assignee: Andy Tolbert
> Priority: Normal
> Fix For: 5.1-alpha1
>
>
> Doing another pass of the auto repair documentation I found a few issues I'd
> like to address.
> * Reduce full repair default {{bytes_per_assignment}} from 200GiB to 50GiB to
> be more conservative by default.
> * Move auto repair documentation under managing/operating and add it to site
> nav.
> * Document {{FixedSplitTokenRangeSplitter}} in cassandra.yaml and overview
> doc.
> * Document that it is desirable to run repair before gc_grace_seconds and why
> that is the case.
> * Remove special casing default of incremental_repair scheduling of 1h,
> default it to 24h like every other repair type. (Simplifies code, and we
> expect users to tune this anyways)
> * Set both incremental and preview_repaired 'enabled' setting as false in
> cassandra.yaml comments, preview_repaired 'enabled' false in
> cassandra_latest.yaml.
> * Set min_repair_interval: 24h in cassandra.yaml for incremental repair, 1h
> in cassandra_latest.yaml
> * Explain why running preview repair is useful in auto repair guide, and
> explain what it does in cassandra.yaml, and reference the metrics to monitor
> to observe for validation mismatches.
> * Make it clear that {{bytes_per_assignment}} and {{max_bytes_schedule}}
> account for compressed size of data instead of actual size to resolve bytes
> for ranges in SSTables.
> * Update existing Repair operating guide to change "repair is not run
> automatically by cassandra" to a 'Automated Repair Scheduling' section that
> cross references the Auto Repair overview doc.
> * Cleanup javadoc in RepairTokenRangeSplitter to avoid redundant
> documentation w/ overview.adoc, reference overview.adoc in the javadoc.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
