[jira] [Resolved] (KUDU-2187) Don't hold threadpool lock when creating additional workers

["Todd Lipcon (JIRA)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/KUDU-2187?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Todd Lipcon resolved KUDU-2187.
-------------------------------
       Resolution: Fixed
    Fix Version/s: 1.6.0

Fixed thisin f5e1203a344c031e357075c900740da33f4d736e

> Don't hold threadpool lock when creating additional workers
> -----------------------------------------------------------
>
>                 Key: KUDU-2187
>                 URL: https://issues.apache.org/jira/browse/KUDU-2187
>             Project: Kudu
>          Issue Type: Improvement
>          Components: perf, util
>    Affects Versions: 1.6.0
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
>            Priority: Major
>             Fix For: 1.6.0
>
>
> I've noticed in a lot of clusters that creating new threads often takes a relatively long time, for whatever reason. Currently we create new threads while holding the threadpool's lock, which means that, even if other existing worker threads finish their current work items and are ready to handle newly enqueued work, everything is frozen while the thread is being created. This creates potentially long stalls resulting in increased queue time at the high percentiles.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)