danny0405 commented on code in PR #12927: URL: https://github.com/apache/hudi/pull/12927#discussion_r1986195060
########## rfc/rfc-91/rfc-91.md: ########## @@ -0,0 +1,145 @@ +<!-- + Licensed to the Apache Software Foundation (ASF) under one or more + contributor license agreements. See the NOTICE file distributed with + this work for additional information regarding copyright ownership. + The ASF licenses this file to You under the Apache License, Version 2.0 + (the "License"); you may not use this file except in compliance with + the License. You may obtain a copy of the License at + + http://www.apache.org/licenses/LICENSE-2.0 + + Unless required by applicable law or agreed to in writing, software + distributed under the License is distributed on an "AS IS" BASIS, + WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + See the License for the specific language governing permissions and + limitations under the License. +--> +# RFC-91: Storage-based lock provider using conditional writes + +## Proposers + +- @alexr17 + +## Approvers + + - @yihua + - @danny0405 + +## Status + +JIRA: [HUDI-9122](https://issues.apache.org/jira/browse/HUDI-9122) + +## Abstract + +Currently in Hudi, distributed locking relies on external systems like Zookeeper, which add complexity and extra dependencies. This RFC introduces a storage-based implementation of the `LockProvider` interface that utilizes conditional writes in cloud storage platforms (such as GCS and AWS S3) to implement a native distributed locking mechanism for Hudi. By directly integrating lock management with cloud storage, this solution reduces operational overhead, and ensures robust coordination during concurrent writes. + +## Background + +AWS S3 recently introduced conditional writes, and GCS and Azure storage already support them. This RFC leverages these features to implement a distributed lock provider for Hudi using a leader election algorithm. In this approach, each process attempts an atomic conditional write to a file calculated using the table base path. The first process to succeed is elected leader and takes charge of exclusive operations. This method provides a straightforward, reliable locking mechanism without the need for external lock providers. + +## Implementation + +This design implements a leader election algorithm for Apache Hudi using a single lock file per table stored in cloud storage. Each table’s lock is represented by a JSON file with the following fields: Review Comment: So the design mainly focuses on cloud storage, how about the HDFS which also has a large user base. -- 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. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
