Browse Source

Cache HadoopUtils instance with specific days expire time (#2181)

* Cache HadoopUtils instance with 7 days expire time

* solve sonar issue

* add kerberos expire time config

* move KERBEROS_EXPIRE_TIME to Constants.java
pull/2/head
tswstarplanet 5 years ago committed by GitHub
parent
commit
e0fc3171da
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
  1. 5
      dolphinscheduler-common/src/main/java/org/apache/dolphinscheduler/common/Constants.java
  2. 44
      dolphinscheduler-common/src/main/java/org/apache/dolphinscheduler/common/utils/HadoopUtils.java
  3. 2
      dolphinscheduler-common/src/main/resources/common.properties

5
dolphinscheduler-common/src/main/java/org/apache/dolphinscheduler/common/Constants.java

@ -813,6 +813,11 @@ public final class Constants {
*/ */
public static final String KERBEROS = "kerberos"; public static final String KERBEROS = "kerberos";
/**
* kerberos expire time
*/
public static final String KERBEROS_EXPIRE_TIME = "kerberos.expire.time";
/** /**
* java.security.krb5.conf * java.security.krb5.conf
*/ */

44
dolphinscheduler-common/src/main/java/org/apache/dolphinscheduler/common/utils/HadoopUtils.java

@ -16,6 +16,9 @@
*/ */
package org.apache.dolphinscheduler.common.utils; package org.apache.dolphinscheduler.common.utils;
import com.google.common.cache.CacheBuilder;
import com.google.common.cache.CacheLoader;
import com.google.common.cache.LoadingCache;
import org.apache.dolphinscheduler.common.Constants; import org.apache.dolphinscheduler.common.Constants;
import org.apache.dolphinscheduler.common.enums.ExecutionStatus; import org.apache.dolphinscheduler.common.enums.ExecutionStatus;
import org.apache.dolphinscheduler.common.enums.ResUploadType; import org.apache.dolphinscheduler.common.enums.ResUploadType;
@ -37,6 +40,7 @@ import java.security.PrivilegedExceptionAction;
import java.util.Collections; import java.util.Collections;
import java.util.List; import java.util.List;
import java.util.Map; import java.util.Map;
import java.util.concurrent.TimeUnit;
import java.util.stream.Collectors; import java.util.stream.Collectors;
import java.util.stream.Stream; import java.util.stream.Stream;
@ -48,24 +52,31 @@ public class HadoopUtils implements Closeable {
private static final Logger logger = LoggerFactory.getLogger(HadoopUtils.class); private static final Logger logger = LoggerFactory.getLogger(HadoopUtils.class);
private static HadoopUtils instance = new HadoopUtils(); private static final String HADOOP_UTILS_KEY = "HADOOP_UTILS_KEY";
private static final LoadingCache<String, HadoopUtils> cache = CacheBuilder
.newBuilder()
.expireAfterWrite(PropertyUtils.getInt(Constants.KERBEROS_EXPIRE_TIME, 7), TimeUnit.DAYS)
.build(new CacheLoader<String, HadoopUtils>() {
@Override
public HadoopUtils load(String key) throws Exception {
return new HadoopUtils();
}
});
private static Configuration configuration; private static Configuration configuration;
private static FileSystem fs; private static FileSystem fs;
private String hdfsUser; private static String hdfsUser = PropertyUtils.getString(Constants.HDFS_ROOT_USER);
private HadoopUtils() { private HadoopUtils() {
hdfsUser = PropertyUtils.getString(Constants.HDFS_ROOT_USER);
init(); init();
initHdfsPath(); initHdfsPath();
} }
public static HadoopUtils getInstance() { public static HadoopUtils getInstance() {
// if kerberos startup , renew HadoopUtils
if (CommonUtils.getKerberosStartupState()){ return cache.getUnchecked(HADOOP_UTILS_KEY);
return new HadoopUtils();
}
return instance;
} }
/** /**
@ -88,10 +99,7 @@ public class HadoopUtils implements Closeable {
/** /**
* init hadoop configuration * init hadoop configuration
*/ */
private void init() { private static void init() {
if (configuration == null) {
synchronized (HadoopUtils.class) {
if (configuration == null) {
try { try {
configuration = new Configuration(); configuration = new Configuration();
@ -161,10 +169,6 @@ public class HadoopUtils implements Closeable {
} catch (Exception e) { } catch (Exception e) {
logger.error(e.getMessage(), e); logger.error(e.getMessage(), e);
} }
}
}
}
} }
/** /**
@ -203,7 +207,6 @@ public class HadoopUtils implements Closeable {
} }
/** /**
* cat file on hdfs * cat file on hdfs
* *
@ -258,7 +261,7 @@ public class HadoopUtils implements Closeable {
/** /**
* the src file is on the local disk. Add it to FS at * the src file is on the local disk. Add it to FS at
* the given dst name. * the given dst name.
*
* @param srcFile local file * @param srcFile local file
* @param dstHdfsPath destination hdfs path * @param dstHdfsPath destination hdfs path
* @param deleteSource whether to delete the src * @param deleteSource whether to delete the src
@ -307,7 +310,6 @@ public class HadoopUtils implements Closeable {
} }
/** /**
*
* delete a file * delete a file
* *
* @param hdfsFilePath the path to delete. * @param hdfsFilePath the path to delete.
@ -351,10 +353,11 @@ public class HadoopUtils implements Closeable {
/** /**
* Renames Path src to Path dst. Can take place on local fs * Renames Path src to Path dst. Can take place on local fs
* or remote DFS. * or remote DFS.
*
* @param src path to be renamed * @param src path to be renamed
* @param dst new path after rename * @param dst new path after rename
* @throws IOException on failure
* @return true if rename is successful * @return true if rename is successful
* @throws IOException on failure
*/ */
public boolean rename(String src, String dst) throws IOException { public boolean rename(String src, String dst) throws IOException {
return fs.rename(new Path(src), new Path(dst)); return fs.rename(new Path(src), new Path(dst));
@ -400,7 +403,6 @@ public class HadoopUtils implements Closeable {
} }
/** /**
*
* @return data hdfs path * @return data hdfs path
*/ */
public static String getHdfsDataBasePath() { public static String getHdfsDataBasePath() {

2
dolphinscheduler-common/src/main/resources/common.properties

@ -91,4 +91,4 @@ yarn.resourcemanager.ha.rm.ids=192.168.xx.xx,192.168.xx.xx
# If it is a single resourcemanager, you only need to configure one host name. If it is resourcemanager HA, the default configuration is fine # If it is a single resourcemanager, you only need to configure one host name. If it is resourcemanager HA, the default configuration is fine
yarn.application.status.address=http://ark1:8088/ws/v1/cluster/apps/%s yarn.application.status.address=http://ark1:8088/ws/v1/cluster/apps/%s
kerberos.expire.time=7

Loading…
Cancel
Save