[
https://issues.apache.org/jira/browse/YARN-1027?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13728992#comment-13728992
]
Bikas Saha commented on YARN-1027:
----------------------------------
RMHAServiceProtocol may be a wrapper that does what main() currently does. Upon
transitionToActive() it instantiates and starts ResourceManager object and upon
transitionToStandby() it stops the ResourceManager object. Potential issues may
occur in the stopping/shutdown logic in the RM and how it interacts with the
extra wrapping layer. Since RPC services are not up and listening, clients may
see long timeouts when connecting to standby RM's.
Alternatively, RMHAServiceProtocol can be added as another service in the RM.
All services may be started as they currently do. Modify RM service startup
logic to first start this service and maybe other RPC services. The
RMHAServiceProtocol service starts internal services when it
transitionToActive() and stops them when transitionToStandby().
Add haEnabled configuration which is false by default. When HA is off then RM
is always started in active state immediately.
> Implement RMHAServiceProtocol
> -----------------------------
>
> Key: YARN-1027
> URL: https://issues.apache.org/jira/browse/YARN-1027
> Project: Hadoop YARN
> Issue Type: Sub-task
> Reporter: Bikas Saha
>
> Implement existing HAServiceProtocol from Hadoop common. This protocol is the
> single point of interaction between the RM and HA clients/services.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
