You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@flink.apache.org by "Piotr Nowojski (Jira)" <ji...@apache.org> on 2019/12/06 14:16:00 UTC

[jira] [Comment Edited] (FLINK-15104) Performance regression on 4.12.2019 in RocksDB benchmarks

    [ https://issues.apache.org/jira/browse/FLINK-15104?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16989786#comment-16989786 ] 

Piotr Nowojski edited comment on FLINK-15104 at 12/6/19 2:15 PM:
-----------------------------------------------------------------

I agree that this version bumping looks like the prime suspect. 

I'm pretty sure the performance regression is real. As you can see those benchmarks are pretty stable (especially non incremental one). 

It might be small, however the problem is that the benchmark coverage for real world usage of RocksDB is very small. It could well be that this change is affecting other use cases in much more profound way, so I wouldn't ignore it without first trying to understand what has happened/changed.


was (Author: pnowojski):
I'm pretty sure the performance regression is real. As you can see those benchmarks are pretty stable (especially non incremental one). 

It might be small, however the problem is that the benchmark coverage for real world usage of RocksDB is very small. It could well be that this change is affecting other use cases in much more profound way, so I wouldn't ignore it without first trying to understand what has happened/changed.

> Performance regression on 4.12.2019 in RocksDB benchmarks
> ---------------------------------------------------------
>
>                 Key: FLINK-15104
>                 URL: https://issues.apache.org/jira/browse/FLINK-15104
>             Project: Flink
>          Issue Type: Bug
>          Components: Benchmarks, Runtime / State Backends
>            Reporter: Piotr Nowojski
>            Priority: Blocker
>             Fix For: 1.10.0
>
>
> A potential small performance regression that happened either on December 4th or 5th:
> http://codespeed.dak8s.net:8000/timeline/?ben=stateBackends.ROCKS&env=2
> http://codespeed.dak8s.net:8000/timeline/?ben=stateBackends.ROCKS_INC&env=2
> it doesn't seem to be visible in the states benchmarks suites, which might be an issue on its own.



--
This message was sent by Atlassian Jira
(v8.3.4#803005)