[
https://issues.apache.org/jira/browse/HIVE-26930?focusedWorklogId=857020&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-857020
]
ASF GitHub Bot logged work on HIVE-26930:
-----------------------------------------
Author: ASF GitHub Bot
Created on: 14/Apr/23 08:46
Start Date: 14/Apr/23 08:46
Worklog Time Spent: 10m
Work Description: subhasisgorai opened a new pull request, #4236:
URL: https://github.com/apache/hive/pull/4236
### What changes does this PR contain?
To support the Planned/Unplanned Failover, we need the capability to
increase the retention period for both the Notification Logs and Change Manager
entries until the successful reverse replication is done (i.e. the Optimized
Bootstrap). A database-level property 'repl.db.under.failover.sync.pending' was
introduced to signify this state. This PR contains the changes for,
1. selective deletion of notification events that are not relevant to the
database(s) in failover
2. skipping the CM clearer thread execution until the time the Optimized
Bootstrap is not done
### Why this change is needed?
The change is needed to make the Optimized Bootstrap and Point-in-time
consistency possible. If the relevant Notification logs and Change Manager
entries are not retained, we can't perform the Optimized Bootstrap.
### Were the changes tested?
I have included relevant unit tests in this PR, and will also perform manual
verification after deploying the changes on a cluster.
### Does this PR introduce _any_ user-facing change?
No
Issue Time Tracking
-------------------
Worklog Id: (was: 857020)
Time Spent: 40m (was: 0.5h)
> Support for increased retention of Notification Logs and Change Manager
> entries
> -------------------------------------------------------------------------------
>
> Key: HIVE-26930
> URL: https://issues.apache.org/jira/browse/HIVE-26930
> Project: Hive
> Issue Type: Improvement
> Components: Standalone Metastore
> Reporter: Subhasis Gorai
> Assignee: Subhasis Gorai
> Priority: Major
> Labels: pull-request-available
> Time Spent: 40m
> Remaining Estimate: 0h
>
> In order to support the Planned/Unplanned Failover use cases, we need the
> capability to increase the retention period for both the Notification Logs
> and Change Manager entries until the successful reverse replication is done
> (i.e. the Optimized Bootstrap).
> If the relevant Notification logs and Change Manager entries are not
> retained, we can't perform a successful Optimized Bootstrap.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
