[
https://issues.apache.org/jira/browse/JCR-5149?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17952890#comment-17952890
]
Tom Duffey commented on JCR-5149:
---------------------------------
https://github.com/apache/jackrabbit/pull/243
> Allow disabling cluster node auto sync outside of testing
> ---------------------------------------------------------
>
> Key: JCR-5149
> URL: https://issues.apache.org/jira/browse/JCR-5149
> Project: Jackrabbit Content Repository
> Issue Type: Improvement
> Components: clustering, config
> Affects Versions: 2.23.1
> Reporter: Tom Duffey
> Priority: Minor
>
> Internally ClusterNode provides a `disableAutoSync` field that is currently
> only accessible to unit tests.
> We would like this field to be available during initialization so that a
> normal node can be configured that will sync at startup but not start its
> auto-sync thread. It should default to the current behavior (disableAutoSync
> = false) but allow us to specify a true value within the ClusterConfig
> supplied to the init() method.
> The use case here is our ongoing effort to provide a "hot backup" solution.
> Although not truly that, we can get very close by:
> # Launching a disposable "backup only" node say once a day
> # This node is configured to join the cluster, sync on startup but not
> process any subsequent journal updates
> # After successful startup and sync simply turn the node off, save the index
> and whatever revision ID was obtained during startup sync
> Compared to my other attempts at hot backup this seems like the most minimal
> and least risky change to Jackrabbit 2.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
