[
https://issues.apache.org/jira/browse/HDFS-14079?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16700039#comment-16700039
]
Yiqun Lin commented on HDFS-14079:
----------------------------------
Hi [~surendrasingh], I am thinking one further problem. This will maybe lead a
inconsistent Router state.
For example, we have three Routers, R1, R2 and R3 as the admin list.
Firstly, we use R1 as the target Router, then we add a mount table m1 for R1.
Then R1 becomes unavailable, then client fail over to R2. We won't get m1 info
from current Router. I mean current implementation isn't fully HA endpoint.
> RBF: RouterAdmin should have failover concept for router
> --------------------------------------------------------
>
> Key: HDFS-14079
> URL: https://issues.apache.org/jira/browse/HDFS-14079
> Project: Hadoop HDFS
> Issue Type: Sub-task
> Affects Versions: 3.1.1
> Reporter: Surendra Singh Lilhore
> Assignee: Surendra Singh Lilhore
> Priority: Major
> Attachments: HDFS-14079-HDFS-13891.01.patch
>
>
> Currenlty {{RouterAdmin}} connect with only one router for admin operation,
> if the configured router is down then router admin command is failing. It
> should allow to configure all the router admin address.
> {code}
> // Initialize RouterClient
> try {
> String address = getConf().getTrimmed(
> RBFConfigKeys.DFS_ROUTER_ADMIN_ADDRESS_KEY,
> RBFConfigKeys.DFS_ROUTER_ADMIN_ADDRESS_DEFAULT);
> InetSocketAddress routerSocket = NetUtils.createSocketAddr(address);
> client = new RouterClient(routerSocket, getConf());
> } catch (RPC.VersionMismatch v) {
> System.err.println(
> "Version mismatch between client and server... command aborted");
> return exitCode;
> }
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: hdfs-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: hdfs-issues-h...@hadoop.apache.org
