[
https://issues.apache.org/jira/browse/YARN-1028?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13846319#comment-13846319
]
Karthik Kambatla commented on YARN-1028:
----------------------------------------
Summary of the patch and how configured failover works:
# Clients / AMs use ClientRMProxy, and NMs use ServerRMProxy to connect to the
ResourceManager. (Client|Server)RMProxy extend RMProxy. These proxies are
similar to the NameNodeProxies in HDFS, but have implementation differences.
# When HA is enabled, instead of using the RM address directly from the conf,
uses a FailoverProxyProvider that provides the proxy. The default is
ConfiguredFailoverProxyProvider. Users can plugin their own implementations.
# ConfiguredFailoverProxyProvider gets the list of RMs from the configuration
passed to it, and returns the ResourceManagers in round-robin (alternate for
two RMs) when asked for a proxy to the RM.
# All failover related connection configs are controlled by
{{yarn.client.failover.*}} configs. The delay between failovers
(yarn.client.failover.max-attempts) is exponential with a base of
yarn.client.failover.sleep-base-ms (default - 500 ms), and can be a maximum of
yarn.client.failover.sleep-max-ms (default - 15000 ms).
> Add FailoverProxyProvider like capability to RMProxy
> ----------------------------------------------------
>
> Key: YARN-1028
> URL: https://issues.apache.org/jira/browse/YARN-1028
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Bikas Saha
> Assignee: Karthik Kambatla
> Attachments: yarn-1028-1.patch, yarn-1028-2.patch, yarn-1028-3.patch,
> yarn-1028-4.patch, yarn-1028-5.patch, yarn-1028-6.patch,
> yarn-1028-draft-cumulative.patch
>
>
> RMProxy layer currently abstracts RM discovery and implements it by looking
> up service information from configuration. Motivated by HDFS and using
> existing classes from Common, we can add failover proxy providers that may
> provide RM discovery in extensible ways.
--
This message was sent by Atlassian JIRA
(v6.1.4#6159)
