[jira] [Closed] (RATIS-1231) Improve discovery ducktape test to research small timeouts and behavior on large cluster.

["Vladimir Steshin (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/RATIS-1231?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Vladimir Steshin closed RATIS-1231.
-----------------------------------

Wrong project.

> Improve discovery ducktape test to research small timeouts and behavior on large cluster.
> -----------------------------------------------------------------------------------------
>
>                 Key: RATIS-1231
>                 URL: https://issues.apache.org/jira/browse/RATIS-1231
>             Project: Ratis
>          Issue Type: Improvement
>            Reporter: Vladimir Steshin
>            Priority: Major
>
> * Prepared separated, centralized place for JVM settings for further extending and development. Can be used for handy test experiments.
> * Node id is now read from the log, not by remote JMX call. Faster, simpler, more reliable. JMX fails sometimes on small timeouts because node id is not published yet. Started node always prints its node id on info level.
> * failure_detection_timeout is now global variable with default value. To run cluster experiments with different settings and on various node count.
> * ZooKeeperSPI got parameter 'session_timeout' according to the documentation.
> * To reduce node failure detection time in 2.8, restored 'soLinger' setting in TCPDiscoverySPI which depends on the Ignite version. No need in > 2.9.0. Impossible for < 2.8.0.
> * Restored logging of the ipfilter crashing nodes.
> * Introduced 'forceSync' setting for ZooKeeper server to avoid slow disk problem.
> * Simplified await timeout on 'Node failure event'. Discovered sensitive delays on ssh, log reads and other non-manageable timeouts



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