[ https://issues.apache.org/jira/browse/YARN-3327?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
sandflee updated YARN-3327: --------------------------- Summary: if NMClientAsync stopContainer failed because of IOException, there's no chance to stopContainer again (was: if NMClientAsync stopContainer failed because of IOException, there's no change to stopContainer again) > if NMClientAsync stopContainer failed because of IOException, there's no > chance to stopContainer again > ------------------------------------------------------------------------------------------------------- > > Key: YARN-3327 > URL: https://issues.apache.org/jira/browse/YARN-3327 > Project: Hadoop YARN > Issue Type: Bug > Components: api, client > Affects Versions: 2.6.0 > Reporter: sandflee > > In AM, I use NMClientAsync to control containers, when I use stopContainer to > kill container ,IOException happens, then If I stopContainer again, nothing > happens, because container Managed by NMClientAsync came to FAILED state. > In my opinion, IOException just means temporary error,it shouldn't came to > FAILED state especially when NM Restart is enabled。 -- This message was sent by Atlassian JIRA (v6.3.4#6332)