[ https://issues.apache.org/jira/browse/HBASE-16775?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15553215#comment-15553215 ]
Appy commented on HBASE-16775: ------------------------------ yeah, gave it a bunch of time and now I am of the opinion that this test doesn't make sense because it's not testing any feature of 'ExportSnapshot' tool. It's basically trying to test if mapreduce frameworks retires a mapper if the "mapreduce.map.maxattempts" is set, nothing HBasey! Also not that the test itself is wrong. It injects failure only some times (3% probability for each mapper). So many runs don't even get the failure injected, and those are precisely the only ones where it passes. {noformat} if (testFailures) { if (context.getConfiguration().getBoolean(CONF_TEST_RETRY, false)) { if (random == null) { random = new Random(); } // FLAKY-TEST-WARN: lower is better, we can get some runs without the // retry, but at least we reduce the number of test failures due to // this test exception from the same map task. if (random.nextFloat() < 0.03) { throw new IOException("TEST RETRY FAILURE: Unable to copy input=" + inputInfo + " time=" + System.currentTimeMillis()); } } else { context.getCounter(Counter.COPY_FAILED).increment(1); throw new IOException("TEST FAILURE: Unable to copy input=" + inputInfo); } } } {noformat} > Flakey test with TestExportSnapshot#testExportRetry and > TestMobExportSnapshot#testExportRetry > ---------------------------------------------------------------------------------------------- > > Key: HBASE-16775 > URL: https://issues.apache.org/jira/browse/HBASE-16775 > Project: HBase > Issue Type: Bug > Affects Versions: 2.0.0 > Reporter: huaxiang sun > Assignee: huaxiang sun > > The root cause is that conf.setInt("mapreduce.map.maxattempts", 10) is not > taken by the mapper job, so the retry is actually 0. Debugging to see why > this is the case. -- This message was sent by Atlassian JIRA (v6.3.4#6332)