big-data/hadoop
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

HDFS-14742. RBF: TestRouterFaultTolerant tests are flaky

Browse Source
This commit is contained in:
Akira Ajisaka 2020-04-21 00:41:18 +09:00 committed by GitHub
parent 1fdfaebd98
commit e069a06137
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
2 changed files with 5 additions and 1 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
hadoop-hdfs-project/hadoop-hdfs-rbf/src/test/java/org/apache/hadoop/hdfs/server/federation/FederationTestUtils.java
View File

@ -474,7 +474,10 @@ public static RouterClient getAdminClient(
/**
* Add a mount table entry in some name services and wait until it is
* available.
* available. If there are multiple routers,
* {@link #createMountTableEntry(List, String, DestinationOrder, Collection)}
* should be used instead because the method does not refresh
* the mount tables of the other routers.
* @param router Router to change.
* @param mountPoint Name of the mount point.
* @param order Order of the mount table entry.

1
hadoop-hdfs-project/hadoop-hdfs-rbf/src/test/java/org/apache/hadoop/hdfs/server/federation/router/TestRouterFaultTolerant.java
View File

@ -248,6 +248,7 @@ private void testWriteWithFailedSubcluster(final DestinationOrder order)
LOG.info("Setup {} with order {}", mountPoint, order);
createMountTableEntry(
getRandomRouter(), mountPoint, order, namenodes.keySet());
refreshRoutersCaches(routers);
LOG.info("Write in {} should succeed writing in ns0 and fail for ns1",
mountPath);