Apache
/
hadoop
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

YARN-8558. NM recovery level db not cleaned up properly on container finish. Contributed by Bibin A Chundatt. 

(cherry picked from commit 3d586841ab)
This commit is contained in:
bibinchundatt 2018-07-28 20:52:39 +05:30
parent c2c3eee69c
commit 2e7876a725
2 changed files with 17 additions and 4 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/main/java/org/apache/hadoop/yarn/server/nodemanager/recovery/NMLeveldbStateStoreService.java
View File

@ -143,9 +143,9 @@ public class NMLeveldbStateStoreService extends NMStateStoreService {
NM_TOKENS_KEY_PREFIX + PREV_MASTER_KEY_SUFFIX;
private static final String CONTAINER_TOKENS_KEY_PREFIX =
"ContainerTokens/";
private static final String CONTAINER_TOKENS_CURRENT_MASTER_KEY =
private static final String CONTAINER_TOKEN_SECRETMANAGER_CURRENT_MASTER_KEY =
CONTAINER_TOKENS_KEY_PREFIX + CURRENT_MASTER_KEY_SUFFIX;
private static final String CONTAINER_TOKENS_PREV_MASTER_KEY =
private static final String CONTAINER_TOKEN_SECRETMANAGER_PREV_MASTER_KEY =
CONTAINER_TOKENS_KEY_PREFIX + PREV_MASTER_KEY_SUFFIX;
private static final String LOG_DELETER_KEY_PREFIX = "LogDeleters/";
@ -658,6 +658,12 @@ public class NMLeveldbStateStoreService extends NMStateStoreService {
batch.delete(bytes(keyPrefix + CONTAINER_KILLED_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_EXIT_CODE_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_UPDATE_TOKEN_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_START_TIME_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_LOG_DIR_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_VERSION_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_REMAIN_RETRIES_KEY_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_RESTART_TIMES_SUFFIX));
batch.delete(bytes(keyPrefix + CONTAINER_WORK_DIR_KEY_SUFFIX));
List<String> unknownKeysForContainer = containerUnknownKeySuffixes
.removeAll(containerId);
for (String unknownKeySuffix : unknownKeysForContainer) {
@ -1169,13 +1175,13 @@ public class NMLeveldbStateStoreService extends NMStateStoreService {
@Override
public void storeContainerTokenCurrentMasterKey(MasterKey key)
throws IOException {
storeMasterKey(CONTAINER_TOKENS_CURRENT_MASTER_KEY, key);
storeMasterKey(CONTAINER_TOKEN_SECRETMANAGER_CURRENT_MASTER_KEY, key);
}
@Override
public void storeContainerTokenPreviousMasterKey(MasterKey key)
throws IOException {
storeMasterKey(CONTAINER_TOKENS_PREV_MASTER_KEY, key);
storeMasterKey(CONTAINER_TOKEN_SECRETMANAGER_PREV_MASTER_KEY, key);
}
@Override

7
hadoop-yarn-project/hadoop-yarn/hadoop-yarn-server/hadoop-yarn-server-nodemanager/src/test/java/org/apache/hadoop/yarn/server/nodemanager/recovery/TestNMLeveldbStateStoreService.java
View File

@ -28,7 +28,9 @@ import static org.junit.Assert.assertTrue;
import static org.mockito.Matchers.any;
import static org.mockito.Mockito.isNull;
import static org.mockito.Mockito.mock;
import static org.mockito.Mockito.spy;
import static org.mockito.Mockito.timeout;
import static org.mockito.Mockito.times;
import static org.mockito.Mockito.verify;
import static org.mockito.Mockito.when;
@ -382,6 +384,11 @@ public class TestNMLeveldbStateStoreService {
restartStateStore();
recoveredContainers = stateStore.loadContainersState();
assertTrue(recoveredContainers.isEmpty());
// recover again to check remove clears all containers
restartStateStore();
NMStateStoreService nmStoreSpy = spy(stateStore);
nmStoreSpy.loadContainersState();
verify(nmStoreSpy,times(0)).removeContainer(any(ContainerId.class));
}
private void validateRetryAttempts(ContainerId containerId)