[
https://issues.apache.org/jira/browse/HBASE-22075?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16799257#comment-16799257
]
Vladimir Rodionov commented on HBASE-22075:
-------------------------------------------
{quote}
Is this something that we can show in a unit test?
{quote}
I have one, but it works only without patch, [~elserj].
Scenario is the following:
0. set "hbase.bulkload.retries.number" to 2
1. Create MOB table (1 region)
2. Load some of MOB data
3. Flush
4. Load MOB data again
5. Flush
6. Now we have 2 store files and 2 MOB file
7. Split table up to at least 4 regions
8. Trigger MOB compaction. Compaction should fail with some data loaded
partially
9. Verify, that we have data missing.
To reproduce this with the patch we will need some special test-related hacks
in the code.
> Potential data loss when MOB compaction fails
> ---------------------------------------------
>
> Key: HBASE-22075
> URL: https://issues.apache.org/jira/browse/HBASE-22075
> Project: HBase
> Issue Type: Bug
> Components: mob
> Affects Versions: 2.1.0, 2.0.0, 2.0.1, 2.1.1, 2.0.2, 2.0.3, 2.1.2, 2.0.4,
> 2.1.3
> Reporter: Vladimir Rodionov
> Assignee: Vladimir Rodionov
> Priority: Critical
> Labels: mob
> Fix For: 2.2.0, 2.0.5, 2.1.4
>
> Attachments: HBASE-22075-v1.patch
>
>
> When MOB compaction fails during last step (bulk load of a newly created
> reference file) there is a high chance of a data loss due to partially loaded
> reference file, cells of which refer to (now) non-existent MOB file. The
> newly created MOB file is deleted automatically in case of a MOB compaction
> failure, but some cells with the references to this file might be loaded to
> HBase.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
