[ https://issues.apache.org/jira/browse/EAGLE-381?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]
Huizhi Lu updated EAGLE-381: ---------------------------- Description: About the attempt_id, assuming attempt_id starting from 1 may not be completely correct. This will not work in HDP Sandbox 2.4. Please see the attempt id in sandbox. Spark version is 1.6.0. I am confused why the attempt id are in different naming rule. We need to fix this issue. I will take care of it. hdp-2.4 [ { "id" : "application_1468625664674_0003", "name" : "Spark Pi", "attempts" : [ { "attemptId" : "appattempt_1468625664674_0003_000001", "startTime" : "2016-07-16T00:13:16.320GMT", "endTime" : "2016-07-16T00:13:24.279GMT", "sparkUser" : "spark", "completed" : true } ] }, { "id" : "application_1468625664674_0002", "name" : "Spark Pi", "attempts" : [ { "startTime" : "2016-07-16T00:12:01.386GMT", "endTime" : "2016-07-16T00:12:18.181GMT", "sparkUser" : "spark", "completed" : true } ] cluster. "id" : "application_1464382345557_264985", "name" : "xxxx", "attempts" : [ { "attemptId" : "1", "startTime" : "2016-07-15T09:01:25.032GMT", "endTime" : "2016-07-15T09:58:29.941GMT", "sparkUser" : "xxx", "completed" : true } ] solution: get resource manager version from api/conf and set attempt id depending on the version . was: About the attempt_id, assuming attempt_id starting from 1 may not be completely correct. This will not work in HDP Sandbox 2.4. Please see the attempt id in sandbox. Spark version is 1.6.0. I am confused why the attempt id are in different naming rule. We need to fix this issue. I will take care of it. hdp-2.4 [ { "id" : "application_1468625664674_0003", "name" : "Spark Pi", "attempts" : [ { "attemptId" : "appattempt_1468625664674_0003_000001", "startTime" : "2016-07-16T00:13:16.320GMT", "endTime" : "2016-07-16T00:13:24.279GMT", "sparkUser" : "spark", "completed" : true } ] }, { "id" : "application_1468625664674_0002", "name" : "Spark Pi", "attempts" : [ { "startTime" : "2016-07-16T00:12:01.386GMT", "endTime" : "2016-07-16T00:12:18.181GMT", "sparkUser" : "spark", "completed" : true } ] Ares. "id" : "application_1464382345557_264985", "name" : "xxxx", "attempts" : [ { "attemptId" : "1", "startTime" : "2016-07-15T09:01:25.032GMT", "endTime" : "2016-07-15T09:58:29.941GMT", "sparkUser" : "xxx", "completed" : true } ] > Get resource manager version and improve app attempt id fetcher for spark job > feeder > ------------------------------------------------------------------------------------ > > Key: EAGLE-381 > URL: https://issues.apache.org/jira/browse/EAGLE-381 > Project: Eagle > Issue Type: Bug > Affects Versions: v0.5.0 > Reporter: Huizhi Lu > Assignee: Huizhi Lu > Fix For: v0.5.0 > > > About the attempt_id, assuming attempt_id starting from 1 may not be > completely correct. This will not work in HDP Sandbox 2.4. Please see the > attempt id in sandbox. Spark version is 1.6.0. > I am confused why the attempt id are in different naming rule. > We need to fix this issue. I will take care of it. > hdp-2.4 > [ { > "id" : "application_1468625664674_0003", > "name" : "Spark Pi", > "attempts" : [ { > "attemptId" : "appattempt_1468625664674_0003_000001", > "startTime" : "2016-07-16T00:13:16.320GMT", > "endTime" : "2016-07-16T00:13:24.279GMT", > "sparkUser" : "spark", > "completed" : true > } ] > }, { > "id" : "application_1468625664674_0002", > "name" : "Spark Pi", > "attempts" : [ { > "startTime" : "2016-07-16T00:12:01.386GMT", > "endTime" : "2016-07-16T00:12:18.181GMT", > "sparkUser" : "spark", > "completed" : true > } ] > cluster. > "id" : "application_1464382345557_264985", > "name" : "xxxx", > "attempts" : [ { > "attemptId" : "1", > "startTime" : "2016-07-15T09:01:25.032GMT", > "endTime" : "2016-07-15T09:58:29.941GMT", > "sparkUser" : "xxx", > "completed" : true > } ] > solution: > get resource manager version from api/conf and set attempt id depending on > the version . -- This message was sent by Atlassian JIRA (v6.3.4#6332)