[
https://issues.apache.org/jira/browse/HDFS-4166?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13493452#comment-13493452
]
Aaron T. Myers commented on HDFS-4166:
--------------------------------------
bq. The main reason why I am considering this is - you need to have a separate
scheduling system, that needs to be highly available, then it needs to deal
with non-availability of namenode etc
Good point. Perhaps Oozie, which already integrates tightly with Hadoop and I
believe intends to add HA soon, would be the right way to do this?
bq. So a daemon thread along the lines of replication factor may be a simpler
solution.
Yea, my main concern is that folks will want a rich scheduling interface (every
week, every night, every X hours, every other Tuesday, etc.) that would be
annoying for us to build into HDFS and already exists elsewhere.
> Add support for scheduled automatic snapshots
> ---------------------------------------------
>
> Key: HDFS-4166
> URL: https://issues.apache.org/jira/browse/HDFS-4166
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Components: name-node
> Affects Versions: Snapshot (HDFS-2802)
> Reporter: Suresh Srinivas
> Assignee: Suresh Srinivas
>
> This jira will track the work related to supporting automatic scheduled
> snapshots.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
