[
https://issues.apache.org/jira/browse/HDFS-12979?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16837642#comment-16837642
]
Plamen Jeliazkov commented on HDFS-12979:
-----------------------------------------
I like the approach of giving more power to the Active NameNode here. It could
even be a lease system like we do with block reports.
Idea: When it's time (current Active checkpoint age is > X configured hours),
the Standbys all compete for some sort of lease to be the "primary
checkpointer", and then one is chosen to actually create and ship the
checkpoint. If the lease "expires" then other Standbys are allowed to produce
and ship their checkpoint image instead.
General question: Is it valuable for all the Standbys to produce checkpoint
images?
> StandbyNode should upload FsImage to ObserverNode after checkpointing.
> ----------------------------------------------------------------------
>
> Key: HDFS-12979
> URL: https://issues.apache.org/jira/browse/HDFS-12979
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: hdfs
> Reporter: Konstantin Shvachko
> Assignee: Chen Liang
> Priority: Major
> Attachments: HDFS-12979.001.patch, HDFS-12979.002.patch,
> HDFS-12979.003.patch
>
>
> ObserverNode does not create checkpoints. So it's fsimage file can get very
> old making bootstrap of ObserverNode too long. A StandbyNode should copy
> latest fsimage to ObserverNode(s) along with ANN.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
