[
https://issues.apache.org/jira/browse/YARN-1232?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13783170#comment-13783170
]
Bikas Saha commented on YARN-1232:
----------------------------------
yarn.resourcemanager.id sounds good. How about
yarn.resourcemanager.ha.rm-instances or rm-ids.
Lets just get rid of this. Dont see it used publicly anywhere
{code}
- private static InetSocketAddress getRmAddress(Configuration conf) {
- return conf.getSocketAddr(YarnConfiguration.RM_ADDRESS,
- YarnConfiguration.DEFAULT_RM_ADDRESS, YarnConfiguration.DEFAULT_RM_PORT);
+ private static InetSocketAddress getRmAddress(Configuration conf)
+ throws Exception {
+ return ClientRMProxy.getRMAddress(conf, ApplicationClientProtocol.class);
{code}
getConfValue() renamed to getConfValueForRMId() or something like that?
addSuffix() name doesnt clarity what this does? Does it need to be public?
Is setAllRpcAddresses() setting non rm-id indexed confs for other pieces of
code that try to access the RPC addresses?
Should we set the RM conf explicitly and update it, instead of passing conf as
an argument?
{code}
+ rm.createAndInitActiveServices(conf);
{code}
In HA enabled scenarios we need the explicit rm ids. How are we handling rm-id
in non-HA setups or in existing clusters where no rm-id is being set currently?
> Configuration to support multiple RMs
> -------------------------------------
>
> Key: YARN-1232
> URL: https://issues.apache.org/jira/browse/YARN-1232
> Project: Hadoop YARN
> Issue Type: Sub-task
> Components: resourcemanager
> Reporter: Karthik Kambatla
> Assignee: Karthik Kambatla
> Labels: ha
> Attachments: yarn-1232-1.patch, yarn-1232-2.patch, yarn-1232-3.patch,
> yarn-1232-4.patch, yarn-1232-5.patch
>
>
> We should augment the configuration to allow users specify two RMs and the
> individual RPC addresses for them.
--
This message was sent by Atlassian JIRA
(v6.1#6144)
