[jira] [Resolved] (HBASE-24636) Increase default Normalizer interval

["Nick Dimiduk (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HBASE-24636?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Nick Dimiduk resolved HBASE-24636.
----------------------------------
    Resolution: Won't Fix

Let's leave this at 5mins for now. The reason being is that the normalizer should be responsive to an active cluster. After HBASE-24628, the operator now has a rate limiting configuration they can use to control the overall impact of these actions. If the work queue gets backed up, the uniqueness property of the work queue will protect against a table being scheduled over and over without others getting a chance.

Watchers, speak up if you think otherwise.

> Increase default Normalizer interval
> ------------------------------------
>
>                 Key: HBASE-24636
>                 URL: https://issues.apache.org/jira/browse/HBASE-24636
>             Project: HBase
>          Issue Type: Task
>          Components: Normalizer
>            Reporter: Nick Dimiduk
>            Priority: Major
>
> Our current default interval for the normalizer chore is 5 minutes. I think that's super aggressive for a background process that's intended to nudge a cluster toward healthy state, considering there's no rate-limiting in place at all (HBASE-24628). IIRC, the default used to be 30 minutes. Maybe we decide on a new default after we decide on a rate limit.



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