sodonnel commented on a change in pull request #1147:
URL: https://github.com/apache/hadoop-ozone/pull/1147#discussion_r456372422
##########
File path:
hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/common/utils/ContainerCache.java
##########
@@ -117,30 +119,57 @@ public ReferenceCountedDB getDB(long containerID, String
containerDBType,
throws IOException {
Preconditions.checkState(containerID >= 0,
"Container ID cannot be negative.");
- lock.lock();
+ ReferenceCountedDB db;
+ Lock containerLock = rocksDBLock.get(containerDBPath);
+ containerLock.lock();
try {
- ReferenceCountedDB db = (ReferenceCountedDB) this.get(containerDBPath);
+ lock.lock();
+ try {
+ db = (ReferenceCountedDB) this.get(containerDBPath);
+ if (db != null) {
+ db.incrementReference();
+ return db;
+ }
+ } finally {
+ lock.unlock();
+ }
- if (db == null) {
+ try {
MetadataStore metadataStore =
MetadataStoreBuilder.newBuilder()
- .setDbFile(new File(containerDBPath))
- .setCreateIfMissing(false)
- .setConf(conf)
- .setDBType(containerDBType)
- .build();
+ .setDbFile(new File(containerDBPath))
+ .setCreateIfMissing(false)
+ .setConf(conf)
+ .setDBType(containerDBType)
+ .build();
db = new ReferenceCountedDB(metadataStore, containerDBPath);
- this.put(containerDBPath, db);
+ } catch (Exception e) {
+ LOG.error("Error opening DB. Container:{} ContainerPath:{}",
+ containerID, containerDBPath, e);
+ throw e;
+ }
+
+ lock.lock();
+ try {
+ ReferenceCountedDB currentDB =
+ (ReferenceCountedDB) this.get(containerDBPath);
+ if (currentDB != null) {
+ // increment the reference before returning the object
+ currentDB.incrementReference();
+ // clean the db created in previous step
+ db.cleanup();
Review comment:
This cleanup() call may be expensive, and ideally it should be outside
the lock. However, I think it will be very rare where you hit this scenario and
therefore I think the logic is OK as it is now.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: ozone-issues-unsubscr...@hadoop.apache.org
For additional commands, e-mail: ozone-issues-h...@hadoop.apache.org
