You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@ozone.apache.org by "Attila Doroszlai (Jira)" <ji...@apache.org> on 2023/05/31 19:06:00 UTC

[jira] [Assigned] (HDDS-8731) Standardize JVM pause monitor

     [ https://issues.apache.org/jira/browse/HDDS-8731?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Attila Doroszlai reassigned HDDS-8731:
--------------------------------------

    Assignee: Attila Doroszlai

> Standardize JVM pause monitor
> -----------------------------
>
>                 Key: HDDS-8731
>                 URL: https://issues.apache.org/jira/browse/HDDS-8731
>             Project: Apache Ozone
>          Issue Type: Improvement
>            Reporter: Attila Doroszlai
>            Assignee: Attila Doroszlai
>            Priority: Major
>
> Ozone uses JvmPauseMonitor from both Hadoop and Ratis.  I think we should standardize on the one from Ratis:
> * It is automatically started for Ratis servers.
> * It has better output:
>    ** more compact (single line)
>    ** includes ID information
> * Ratis has more frequent releases, so it may be easier to make improvements or fixes, if needed, at the source.
> {code:title=Ratis}
> scm_1       | 2023-04-22 01:58:49,875 [JvmPauseMonitor0] INFO util.JvmPauseMonitor: JvmPauseMonitor-58db5753-a6a0-4ffc-aac9-e0d1221cc5e9: Started
> ...
> scm_1       | 2023-04-22 01:59:26,776 [JvmPauseMonitor0] WARN util.JvmPauseMonitor: JvmPauseMonitor-58db5753-a6a0-4ffc-aac9-e0d1221cc5e9: Detected pause in JVM or host machine (eg GC): pause of approximately 1393794624ns. No GCs detected.
> {code}
> {code:title=Hadoop}
> scm_1       | 2023-04-22 01:58:57,031 [org.apache.hadoop.util.JvmPauseMonitor$Monitor@45e931a9] INFO util.JvmPauseMonitor: Starting JVM pause monitor
> ...
> scm_1       | 2023-04-22 01:59:26,933 [org.apache.hadoop.util.JvmPauseMonitor$Monitor@45e931a9] INFO util.JvmPauseMonitor: Detected pause in JVM or host machine (eg GC): pause of approximately 1393ms
> scm_1       | No GCs detected
> {code}
> Proposed changes:
> # Remove usage of Hadoop JvmPauseMonitor
> # Add explicit usage of Ratis JvmPauseMonitor
> #* for services without any Ratis server (HttpFS, Recon, S3 Gateway)
> #* for services that have optional Ratis server if Ratis is disabled (OM, SCM)
> CC [~szetszwo]



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@ozone.apache.org
For additional commands, e-mail: issues-help@ozone.apache.org