[
https://issues.apache.org/jira/browse/HBASE-29245?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Junegunn Choi resolved HBASE-29245.
-----------------------------------
Fix Version/s: 3.0.0-beta-2
2.6.3
Resolution: Fixed
> Region reopening batch size not increased when backoff is zero
> --------------------------------------------------------------
>
> Key: HBASE-29245
> URL: https://issues.apache.org/jira/browse/HBASE-29245
> Project: HBase
> Issue Type: Bug
> Reporter: Junegunn Choi
> Assignee: Junegunn Choi
> Priority: Minor
> Labels: pull-request-available
> Fix For: 3.0.0-beta-2, 2.6.3
>
> Attachments: image-2025-04-08-15-46-35-759.png,
> image-2025-04-08-15-46-44-588.png
>
>
> h2. Problem
> {{hbase.reopen.table.regions.progressive.batch.size.max}} added in
> HBASE-28215 is for making HBase reopen regions in batches on an alter
> operation. The batch size is expected to be gradually increased from 1 to the
> specified maximum. (i.e. 1, 2, 4, 8, 16, ..., max) However, when
> {{hbase.reopen.table.regions.progressive.batch.backoff.ms}} is set to 0,
> which is the default, the batch size is not increased. So whatever the
> maximum value is, HBase will reopen only one region at a time.
> h2. Workaround
> Use a non-zero backoff value, such as 1, even if you don't need backoff.
> h2. Solution
> Make a slight modification to the code so that the batch size is increased
> even when backoff is 0.
> h2. Test
> The existing test case was updated to validate the progressive growth of the
> batch size.
> h2. Manual testing
> Tested the fix locally, and confirmed that the batch size increases as
> expected.
> !image-2025-04-08-15-46-35-759.png|width=600!
> !image-2025-04-08-15-46-44-588.png|width=600!
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
