[ https://issues.apache.org/jira/browse/YARN-1521?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13946246#comment-13946246 ]
Zhijie Shen commented on YARN-1521: ----------------------------------- Browsed through the newest, and have the following comments: 1. If the application is already in the RM cache, we shouldn't log success. Otherwise, there may be multiple logs for one submission. {code} + RMAuditLogger.logSuccess(user, AuditConstants.SUBMIT_APP_REQUEST, + "ClientRMService", applicationId); {code} 2. Should it sleep a while before retry? {code} + try { + client.getApplications(); + return; + } catch (Exception e) { + LOG.error(e.getMessage()); + } finally { + client.stop(); + } {code} 3. Again, sleep 1ms before next try, yielding to the thread of API methods' invoking. And, have max retry when exception? {code} while (keepRunning) { + if (cluster.getStartFailoverFlag()) { + try { + explicitFailover(); + keepRunning = false; + } catch (Exception e) { + // Do Nothing + } finally { + keepRunning = false; + } + } + } {code} > Mark appropriate protocol methods with the idempotent annotation or > AtMostOnce annotation > ----------------------------------------------------------------------------------------- > > Key: YARN-1521 > URL: https://issues.apache.org/jira/browse/YARN-1521 > Project: Hadoop YARN > Issue Type: Sub-task > Reporter: Xuan Gong > Assignee: Xuan Gong > Attachments: YARN-1521.0.patch, YARN-1521.1.patch > > > After YARN-1028, we add the automatically failover into RMProxy. This JIRA is > to identify whether we need to add idempotent annotation and which methods > can be marked as idempotent. -- This message was sent by Atlassian JIRA (v6.2#6252)