[
https://issues.apache.org/jira/browse/FLINK-2288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14620572#comment-14620572
]
ASF GitHub Bot commented on FLINK-2288:
---------------------------------------
Github user tillrohrmann commented on the pull request:
https://github.com/apache/flink/pull/886#issuecomment-120000744
The web interface is, modulo some object which are not serializable,
already independent of the JobManager. It should not be a big problem to
only have one web server which also retrieves the leading JobManager from
ZooKeeper and then serves the information from the leader.
On Thu, Jul 9, 2015 at 4:24 PM, Max <notificati...@github.com> wrote:
> Yes that makes sense. So the user will always have to connect to the web
> interface of the leading job manager, right? We could only circumvent that
> by separating the web interface from the job manager.
>
> —
> Reply to this email directly or view it on GitHub
> <https://github.com/apache/flink/pull/886#issuecomment-120000117>.
>
> Setup ZooKeeper for distributed coordination
> --------------------------------------------
>
> Key: FLINK-2288
> URL: https://issues.apache.org/jira/browse/FLINK-2288
> Project: Flink
> Issue Type: Sub-task
> Components: JobManager, TaskManager
> Reporter: Ufuk Celebi
> Assignee: Ufuk Celebi
> Fix For: 0.10
>
>
> Having standby JM instances for job manager high availabilty requires
> distributed coordination between JM, TM, and clients. For this, we will use
> ZooKeeper (ZK).
> Pros:
> - Proven solution (other projects use it for this as well)
> - Apache TLP with large community, docs, and library with required "recipies"
> like leader election (see below)
> Related Wiki:
> https://cwiki.apache.org/confluence/display/FLINK/JobManager+High+Availability
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
