[
https://issues.apache.org/jira/browse/OOZIE-1715?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13914858#comment-13914858
]
Robert Kanter commented on OOZIE-1715:
--------------------------------------
This is for the Job IDs? As long as the Oozie servers don't all start at the
same exact time, this should work. Though I suppose it would be nice if the
IDs all matched.
Given that we have Curator, I think we should use either their [Shared
Counter|http://curator.apache.org/curator-recipes/shared-counter.html] or
[Distributed Atomic
Long|http://curator.apache.org/curator-recipes/distributed-atomic-long.html].
> Distributed ID sequence for HA
> ------------------------------
>
> Key: OOZIE-1715
> URL: https://issues.apache.org/jira/browse/OOZIE-1715
> Project: Oozie
> Issue Type: Bug
> Components: HA
> Reporter: purshotam shah
>
> Currently each host maintains their own sequencing. In case of HA, this will
> not work. We need to have distributed sequence generator. All server in HA
> will use it to get incremental sequence.
> One good option is to use ZK PERSISTENT_SEQUENTIAL, other options is to use
> Database sequence generator (not sure if all DB provides).
> There are few other discussed @
> http://srinathsview.blogspot.com/2012/04/generating-distributed-sequence-number.html
> http://stackoverflow.com/questions/2671858/distributed-sequence-number-generation/5685869
> Suggestion?
--
This message was sent by Atlassian JIRA
(v6.1.5#6160)
