You are viewing a plain text version of this content. The canonical link for it is here.
Posted to yarn-issues@hadoop.apache.org by "Subru Krishnan (JIRA)" <ji...@apache.org> on 2016/06/18 00:50:05 UTC
[jira] [Resolved] (YARN-5119) Timeout HA issue: Enable IPC ping for
all calls by default
[ https://issues.apache.org/jira/browse/YARN-5119?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Subru Krishnan resolved YARN-5119.
----------------------------------
Resolution: Duplicate
I looked into the problem and found that the root cause was that the default timeout for the rpc clients was infinite and moreover it is not configurable. This has been fixed as [~iwasakims] has made the rpc client timeout configurablein a later version, so I am closing this jira as a duplicate. Thanks to [~brookz] for cherry-picking HADOOP-11252 and validating that it resolves our issue.
> Timeout HA issue: Enable IPC ping for all calls by default
> ----------------------------------------------------------
>
> Key: YARN-5119
> URL: https://issues.apache.org/jira/browse/YARN-5119
> Project: Hadoop YARN
> Issue Type: Bug
> Components: client, nodemanager
> Reporter: Giovanni Matteo Fumarola
> Assignee: Subru Krishnan
>
> We have RM work preserving HA setup with 2 RM instances – RM1 and RM2. RM1 is initially active and so clients connect successfully to a RM1. RM1 subsequently hangs (for any reason) and RM2 takes over as active but clients wait indefinitely on RM1.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: yarn-issues-unsubscribe@hadoop.apache.org
For additional commands, e-mail: yarn-issues-help@hadoop.apache.org