ignite-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Konstantin Boudnik (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (IGNITE-1203) Cluster Discovery via ZooKeeper
Date Sat, 15 Aug 2015 00:47:45 GMT

    [ https://issues.apache.org/jira/browse/IGNITE-1203?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14698005#comment-14698005
] 

Konstantin Boudnik commented on IGNITE-1203:
--------------------------------------------

Paul, is there an actual usecase for ZK in Ignite besides "it would be nice" and "Hadoop uses
it"?
Looking at ZK I really wonder why anyone is using it considering all the limitations, and
implementation shortcomings it has. From the description above the follow catches my attention
as well "ZooKeeper is a centralised service" - any centralized service is bad in a distributed
system because it introduces an SPF. You can argue that ZK itself is HA-ready, but the amount
of resources it requires to support that HA is troubling; and it doesn't work on clusters
connected via WAN. Hence the question: what's the real use-case behind this? Thanks!

> Cluster Discovery via ZooKeeper
> -------------------------------
>
>                 Key: IGNITE-1203
>                 URL: https://issues.apache.org/jira/browse/IGNITE-1203
>             Project: Ignite
>          Issue Type: New Feature
>          Components: general
>            Reporter: Raúl Kripalani
>            Assignee: Dmitriy Setrakyan
>              Labels: discovery
>
> From their site:
> bq. ZooKeeper is a centralised service for maintaining configuration information, naming,
providing distributed synchronisation, and providing group services. 
> It is used by Kafka, Hadoop Yarn, HBase and many other technologies as a distributed
service registry and discovery directory.
> It would be nice to support cluster discovery via ZK.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

Mime
View raw message