[jira] [Resolved] (KUDU-1777) glog drop_log_memory causes a lot of IO to log dir

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

     [ https://issues.apache.org/jira/browse/KUDU-1777?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Todd Lipcon resolved KUDU-1777.
-------------------------------
       Resolution: Duplicate
    Fix Version/s: n/a

Going to resolve this as a duplicate of KUDU-695, since the fix for KUDU-695 will prevent this from causing issues.

> glog drop_log_memory causes a lot of IO to log dir
> --------------------------------------------------
>
>                 Key: KUDU-1777
>                 URL: https://issues.apache.org/jira/browse/KUDU-1777
>             Project: Kudu
>          Issue Type: Improvement
>          Components: util
>    Affects Versions: 1.1.0
>            Reporter: Todd Lipcon
>            Assignee: Todd Lipcon
>            Priority: Critical
>             Fix For: n/a
>
>
> glog is configured by default to call posix_fadvise with DONTNEED every time it flushes, and also configured such that any WARN or above causes such a flush. Every such fadvise call generates new IO to the log directory, journal commits, etc, and I think this may be affecting latency and stability adversely.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)