[
https://issues.apache.org/jira/browse/PHOENIX-7986?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Viraj Jasani resolved PHOENIX-7986.
-----------------------------------
Fix Version/s: 5.4.0
5.3.3
Resolution: Fixed
> Replace IndexCDCConsumer's exponential idle backoff with a fixed poll interval
> ------------------------------------------------------------------------------
>
> Key: PHOENIX-7986
> URL: https://issues.apache.org/jira/browse/PHOENIX-7986
> Project: Phoenix
> Issue Type: Improvement
> Reporter: Palash Chauhan
> Assignee: Palash Chauhan
> Priority: Major
> Fix For: 5.4.0, 5.3.3
>
>
> When an eventually consistent index consumer finds no new changes, it backs
> off exponentially, up to 40 seconds between polls after about a minute and a
> half of quiet. A write to an idle region can therefore wait ~45 seconds
> before the consumer even looks for it. That backoff is borrowed from HBase's
> RPC retry logic which exists to relieve a struggling server. That is not the
> case in case of the EC index consumer.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
