[
https://issues.apache.org/jira/browse/HBASE-20475?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16467225#comment-16467225
]
Zheng Hu commented on HBASE-20475:
----------------------------------
Let me explain the changes in UT refactor :
1. Let TestReplicationDroppedTables#setUpBase override the
TestReplicationBase#setUpBase, because when I read the log, I found that in
testEditsBehindDroppedTableTiming, there were some replaying WALs which was
stuck in its previous UT testEditsDroppedWithDroppedTable. The original
version, we TestReplicationBase#setUpBase firstly, then
TestReplicationDroppedTables#setUp, so the peer creation is ahead of WAL
rolling, so the newly created peer in testEditsBehindDroppedTableTiming would
still catch the WALs from previous UT.
2. In those UTs, we would shutdown the source mini cluster to keep only one
RS, but in the following operations, we were still using the htable1 & admin
which was initialized for the previous mini cluster, so I fixed those .
3. The verifyReplicationProceeded & verifyReplicationStuck should not only
check the lastRowkey, as I said above.... So I fixed those too..
4. Some minor change, such as we used the deprecated HTableDescriptor, I
changed them to builder...
> Fix the flaky TestReplicationDroppedTables unit test.
> -----------------------------------------------------
>
> Key: HBASE-20475
> URL: https://issues.apache.org/jira/browse/HBASE-20475
> Project: HBase
> Issue Type: Bug
> Affects Versions: 2.1.0
> Reporter: Zheng Hu
> Assignee: Zheng Hu
> Priority: Major
> Fix For: 3.0.0, 2.1.0
>
> Attachments:
> 0001-HBASE-20475-TestReplicationDroppedTables-refactor.patch,
> HBASE-20475-addendum-v2.patch, HBASE-20475-addendum-v3.patch,
> HBASE-20475-addendum.patch, HBASE-20475.patch
>
>
> See
> https://builds.apache.org/job/HBASE-Find-Flaky-Tests/lastSuccessfulBuild/artifact/dashboard.html
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
