[
https://issues.apache.org/jira/browse/HBASE-29124?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hernan Gelaf-Romer updated HBASE-29124:
---------------------------------------
Description:
Restoring to original splits using incremental backups that contain bulkload
edits will fail if the target FS is different than the FS underpinning HBase
itself. This was detected when backups were stored on S3A, but can also be
replicated via the LocalFileSystem. The patch updates a test that exposes this
issue.
Link to a similar issue: https://issues.apache.org/jira/browse/HBASE-29114
was:
When the backups are stored on a location that is not the DistributedFilesystem
underpinning HBase itself, restoring incremental backups with bulkloaded data
with original splits fails. Detected with backups stored on S3A, but can be
reproduced with any other (like LocalFilesystem).
More context here: https://issues.apache.org/jira/browse/HBASE-29114
> Restoring to original splits fails if bulkloads are on separate FileSystem
> --------------------------------------------------------------------------
>
> Key: HBASE-29124
> URL: https://issues.apache.org/jira/browse/HBASE-29124
> Project: HBase
> Issue Type: Bug
> Reporter: Hernan Gelaf-Romer
> Priority: Major
> Labels: pull-request-available
>
> Restoring to original splits using incremental backups that contain bulkload
> edits will fail if the target FS is different than the FS underpinning HBase
> itself. This was detected when backups were stored on S3A, but can also be
> replicated via the LocalFileSystem. The patch updates a test that exposes
> this issue.
>
> Link to a similar issue: https://issues.apache.org/jira/browse/HBASE-29114
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
