[
https://issues.apache.org/jira/browse/KUDU-1550?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Todd Lipcon updated KUDU-1550:
------------------------------
Component/s: tserver
consensus
> remote_bootstrap should avoid starting the entire process over again when
> `bootstrap_source` node crashes before remote_bootstrap complete.
> -------------------------------------------------------------------------------------------------------------------------------------------
>
> Key: KUDU-1550
> URL: https://issues.apache.org/jira/browse/KUDU-1550
> Project: Kudu
> Issue Type: Improvement
> Components: consensus, tserver
> Reporter: zhangsong
>
> In current design and implementation, remote_bootstrap will start the entire
> process again, when `bootstrap_source` node crashes before `remote_bootstrap`
> complete. Instead, it will be better to continue remote_bootstrap from last
> failed `position`.
> 1 it will save a lot of network and computation resource for remote_bootstrap.
> 2 it will save much time to recover when met `bootstrap_source` crash .
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
