zookeeper-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Mocheng Guo (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (ZOOKEEPER-3432) Improving zookeeper trace for performance and scalability
Date Tue, 18 Jun 2019 20:38:00 GMT

     [ https://issues.apache.org/jira/browse/ZOOKEEPER-3432?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Mocheng Guo updated ZOOKEEPER-3432:
-----------------------------------
    Description: Current server trace goes into normal local log files which does not scale
and has negative impact on server performance when turned on. The proposed improvement is
to write traces asynchronously with configurable in memory buffer to a separate process
which can be on different hardware, so that large volume of traces would be processed and
persisted without affecting zookeeper server performance.  (was: Current server trace goes
into normal local log files which does not scale and has negative impact on server performance
when turned on. The proposed improvement is to write traces asynchronously with configurable
in memory buffer to a separate process which can be on different hardware, so that large
volume of traces would be processed and persisted without affect zookeeper server performance.)

> Improving zookeeper trace for performance and scalability
> ---------------------------------------------------------
>
>                 Key: ZOOKEEPER-3432
>                 URL: https://issues.apache.org/jira/browse/ZOOKEEPER-3432
>             Project: ZooKeeper
>          Issue Type: Improvement
>          Components: server
>    Affects Versions: 3.5.6
>            Reporter: Mocheng Guo
>            Priority: Major
>
> Current server trace goes into normal local log files which does not scale and has negative
impact on server performance when turned on. The proposed improvement is to write traces
asynchronously with configurable in memory buffer to a separate process which can be on different hardware,
so that large volume of traces would be processed and persisted without affecting zookeeper
server performance.



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

Mime
View raw message