runzhiwang commented on a change in pull request #709: HDDS-3244. Improve write
efficiency by opening RocksDB only once
URL: https://github.com/apache/hadoop-ozone/pull/709#discussion_r396509770
##########
File path:
hadoop-hdds/container-service/src/main/java/org/apache/hadoop/ozone/container/keyvalue/helpers/KeyValueContainerUtil.java
##########
@@ -73,14 +71,6 @@ public static void createContainerMetaData(File
containerMetaDataPath, File
throw new IOException("Unable to create directory for metadata storage."
+
" Path: " + containerMetaDataPath);
}
- MetadataStore store = MetadataStoreBuilder.newBuilder().setConf(conf)
- .setCreateIfMissing(true).setDbFile(dbFile).build();
Review comment:
@mukul1987 This is in HddsDispatcher.WriteData, if add to the cache here, 2
follower datanodes will not execute RocksDB.open in their
HddsDispatcher.WriteData until the leader datanode finish
HddsDispatcher.WriteData including RocksDB.open, so leader and follower open
RocksDB in sequential. But if in apply transaction phase, RocksDB.open in
parallel, it save 200ms. I think RocksDB was not used until
HddsDispatcher.PutBlock, so add to cache in apply transaction phase does not
matter. Please fix me if I am wrong.
----------------------------------------------------------------
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:
[email protected]
With regards,
Apache Git Services
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
