[ 
https://issues.apache.org/jira/browse/HBASE-18069?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16017021#comment-16017021
 ] 

Chia-Ping Tsai edited comment on HBASE-18069 at 5/19/17 7:35 AM:
-----------------------------------------------------------------

bq. The enableTableReplication will sink the table schema change to peer 
cluster.
It seems to me that it happens only if the table doesn't exist in the peer 
cluster. If the table exists, the both of enableTableReplication and 
disableTableReplication don't change the descriptor in peer cluster.


was (Author: chia7712):
bq, The enableTableReplication will sink the table schema change to peer 
cluster.
It seems to me that it happens only if the table doesn't exist in the peer 
cluster. If the table exists, the both of enableTableReplication and 
disableTableReplication don't change the descriptor in peer cluster.

> Fix flaky test 
> TestReplicationAdminWithClusters#testDisableAndEnableReplication
> -------------------------------------------------------------------------------
>
>                 Key: HBASE-18069
>                 URL: https://issues.apache.org/jira/browse/HBASE-18069
>             Project: HBase
>          Issue Type: Bug
>          Components: test
>            Reporter: Chia-Ping Tsai
>            Priority: Trivial
>              Labels: beginner
>             Fix For: 2.0.0, 1.4.0
>
>
> If we run testDisableAndEnableReplication, we will get the following error 
> message.
> {code}
> testDisableAndEnableReplication(org.apache.hadoop.hbase.client.replication.TestReplicationAdminWithClusters)
>   Time elapsed: 2.046 sec  <<< FAILURE!
> java.lang.AssertionError: expected:<1> but was:<0>
>         at org.junit.Assert.fail(Assert.java:88)
>         at org.junit.Assert.failNotEquals(Assert.java:834)
>         at org.junit.Assert.assertEquals(Assert.java:645)
>         at org.junit.Assert.assertEquals(Assert.java:631)
>         at 
> org.apache.hadoop.hbase.client.replication.TestReplicationAdminWithClusters.testDisableAndEnableReplication(TestReplicationAdminWithClusters.java:160)
> {code}
> The critical code is shown below.
> {code}
>     admin1.disableTableReplication(tableName);
>     HTableDescriptor table = admin1.getTableDescriptor(tableName);
>     for (HColumnDescriptor fam : table.getColumnFamilies()) {
>       assertEquals(fam.getScope(), HConstants.REPLICATION_SCOPE_LOCAL);
>     }
>     table = admin2.getTableDescriptor(tableName);
>     for (HColumnDescriptor fam : table.getColumnFamilies()) {
>       assertEquals(fam.getScope(), HConstants.REPLICATION_SCOPE_LOCAL);
>     }
> {code}
> Is HTD got from admin2 affected by admin1? I don't think so. We should remove 
> the related assertion.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)

Reply via email to