[HADOOP-16830] Add Public IOStatistics API

• Type: New Feature
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 3.3.0
• Fix Version/s: 3.3.1
• Component/s: fs, fs/s3
• Labels: pull-request-available
• Target Version/s: 3.4.0
• Assignee: Steve Loughran
• Reporter: Steve Loughran
• Created: 27/Jan/20 20:34
• Updated: 24/Apr/23 17:31
• Resolved: 14/Jan/21 14:07

Description

Applications like to collect the statistics which specific operations take, by collecting exactly those operations done during the execution of FS API calls by their individual worker threads, and returning these to their job driver

• S3A has a statistics API for some streams, but it's a non-standard one; Impala &c can't use it
• FileSystem storage statistics are public, but as they aren't cross-thread, they don't aggregate properly

Proposed

1. A new IOStatistics interface to serve up statistics
2. S3A to implement
3. other stores to follow
4. Pass-through from the usual wrapper classes (FS data input/output streams)

It's hard to think about how best to offer an API for operation context stats, and how to actually implement.

ThreadLocal isn't enough because the helper threads need to update on the thread local value of the instigator

My Initial PoC doesn't address that issue, but it shows what I'm thinking of

Issue Links

• blocks: HADOOP-13551 Collect AwsSdkMetrics in S3A FileSystem IOStatistics
• blocks: IMPALA-8544 Expose additional S3A / S3Guard metrics
• is a parent of: MAPREDUCE-7315 LocatedFileStatusFetcher to collect/publish IOStatistics
• is related to: HADOOP-17281 Implement FileSystem.listStatusIterator() in S3AFileSystem
• is superceded by: HADOOP-17469 IOStatistics Phase II
• links to: GitHub Pull Request #1820
• links to: GitHub Pull Request #2069
• links to: GitHub Pull Request #2323
• links to: GitHub Pull Request #2577
• links to: GitHub Pull Request #2616

Sub-Tasks

• S3A statistics to support IOStatistics: Resolved
• ABFS Streams to support IOStatistics API: Resolved
• renaming S3A Statistic DELEGATION_TOKENS_ISSUED to DELEGATION_TOKEN_ISSUED broke tests downstream: Resolved
• hadoop-common to add IOStatistics API: Resolved
• refactoring the signature of S3ClientFactory breaks hboss compile: Resolved
• Intermittent failure of S3A tests which make assertions on statistics/IOStatistics: Resolved
• AbstractContractStreamIOStatisticsTest fails if read buffer not full: Open
• S3A ITestPartialRenamesDeletes.testPartialDirDelete[bulk-delete=true] failure: Resolved
• [s3a] Intermittent failure of ITestS3ADeleteCost.testDeleteSingleFileInDir: Resolved