[
https://issues.apache.org/jira/browse/HDFS-9919?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15185637#comment-15185637
]
Andrew Wang commented on HDFS-9919:
-----------------------------------
LGTM +1, will commit shortly
> TestStandbyCheckpoints#testNonPrimarySBNUploadFSImage waitForCheckpoint
> incorrectly
> -----------------------------------------------------------------------------------
>
> Key: HDFS-9919
> URL: https://issues.apache.org/jira/browse/HDFS-9919
> Project: Hadoop HDFS
> Issue Type: Bug
> Components: test
> Reporter: Lin Yiqun
> Assignee: Lin Yiqun
> Priority: Minor
> Attachments: HDFS-9919.001.patch
>
>
> In HDFS-9787, it sloved the problem that standby NNs can upload FSImage to
> ANN after become non-primary standby NN. But in its unit test, it seems there
> is a small problem. when ANN change state to standby and standby nns should
> do a checkpoint. And in test, it makes a checpoint check as follow:
> {code}
> for (int i = 0; i < NUM_NNS; i++) {
> // Once the standby catches up, it should do a checkpoint
> // and save to local directories.
> HATestUtil.waitForCheckpoint(cluster, 1, ImmutableList.of(12));
> }
> {code}
> And in these code, the nnIdx for waitForCheckpoint is always {{1}}, not
> {{i}}. It seems there is no need to check one standby nn's checkpoint
> {{NUM_NNS}} times.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
