[
https://issues.apache.org/jira/browse/AMBARI-9158?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Myroslav Papirkovskyy updated AMBARI-9158:
------------------------------------------
Description:
Deployment Options
AMS Single Node, Local File System - This is the default configuration and
deployment option of AMS. It utilizes a single instance of AMS backed by local
file system. Based on performance data, this will serve up to 800 nodes.
AMS backed by HDFS - Possible deployment option, same as above, but where HBase
is backed by HDFS provided by the deployed cluster. In this option a service
daemon is configured similar to emit metrics to a collector, this happens
similar to datanode sending metrics to a gmond running on a specific host:port.
The client (Ambari Web/View) does simple round-robin for GET calls. No sharding
needed.
was:
Deployment Options
AMS Single Node, Local File System - This is the default configuration and
deployment option of AMS. It utilizes a single instance of AMS backed by local
file system. Based on performance data, this will serve up to 800 nodes.
AMS backed by HDFS - Possible deployment option, same as above, but where HBase
is backed by HDFS provided by the deployed cluster. In this option a service
daemon is configured similar to emit metrics to a collector, this happens
similar to datanode sending metrics to a gmond running on a specific host:port.
The client (Ambari Web/View) does simple round-robin for GET calls. No sharding
needed.
> Add support for AMS operation Modes
> -----------------------------------
>
> Key: AMBARI-9158
> URL: https://issues.apache.org/jira/browse/AMBARI-9158
> Project: Ambari
> Issue Type: Bug
> Components: ambari-metrics
> Affects Versions: 2.0.0
> Reporter: Myroslav Papirkovskyy
> Assignee: Myroslav Papirkovskyy
> Priority: Critical
> Fix For: 2.0.0
>
>
> Deployment Options
> AMS Single Node, Local File System - This is the default configuration and
> deployment option of AMS. It utilizes a single instance of AMS backed by
> local file system. Based on performance data, this will serve up to 800 nodes.
> AMS backed by HDFS - Possible deployment option, same as above, but where
> HBase is backed by HDFS provided by the deployed cluster. In this option a
> service daemon is configured similar to emit metrics to a collector, this
> happens similar to datanode sending metrics to a gmond running on a specific
> host:port. The client (Ambari Web/View) does simple round-robin for GET
> calls. No sharding needed.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
