[
https://issues.apache.org/jira/browse/HBASE-28585?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17885250#comment-17885250
]
Nihal Jain commented on HBASE-28585:
------------------------------------
Ohk no worries let me explain a bit.
In HBase we have multiple active branches where we push code.
Whenever we commit a change in master it will only go in 4.x release. Hence for
us to push a change to other branches we need to checkout to the particular
branch and raise PRs specific to it.
In general we merge changes in following order:
* master
* branch-3
* branch-2
* branch-2.6 and so on.
Master being the next unreleased 4.x branch. Similarly we will have branch-3
for 3.x release, branch-2 for 2.7.x release, branch-2.6 for 2.6.x release and
so on.
To backport you basically need to switch to specific branch. Checkout a new
branch say:
{code:java}
git checkout branch-3
git checkout -b HBASE-28585_branch-3
git cherry-pick fdd8fa5d083b95318a94f7f105d7a5f231f02987 # this should be the
commit id of the PR you merged in master
{code}
Shared this for your knowledge/ future reference. You can find more info at
[https://hbase.apache.org/book.html#committing.patches] (note: info might be
stale and not fully accurate based on how we do it today)
Anyways, let me raise backports on your behalf since this should be a clean
apply. No actions needed from you end. Thank you so much for raising the issue
and the fix!
> copy_tables_desc.rb script should handle scenarios where the namespace does
> not exist in the target cluster
> -----------------------------------------------------------------------------------------------------------
>
> Key: HBASE-28585
> URL: https://issues.apache.org/jira/browse/HBASE-28585
> Project: HBase
> Issue Type: Improvement
> Components: jruby, scripts
> Affects Versions: 2.4.17
> Reporter: wenhao
> Assignee: wenhao
> Priority: Minor
> Labels: pull-request-available
> Fix For: 4.0.0-alpha-1
>
>
> When utilizing the {{copy_tables_desc.rb}} script to duplicate tables to a
> target cluster, if the specified table's namespace is nonexistent in the
> target cluster, the script fails to execute successfully. It is recommended
> to incorporate logic within the script for detecting and handling scenarios
> where the namespace does not exist.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
