hbase-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jiajia Li (JIRA)" <j...@apache.org>
Subject [jira] [Commented] (HBASE-12673) Add a UT to read mob file when the mob hfile moving from the mob dir to the archive dir
Date Tue, 16 Dec 2014 09:52:13 GMT

    [ https://issues.apache.org/jira/browse/HBASE-12673?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14247995#comment-14247995

Jiajia Li commented on HBASE-12673:

hi, [~jon@cloudera.com], as you know when hbase reads mob cell, it has two steps.
# Read the ref cell from the HBase, and get the cell value which is the mob file name.
# HBase has two possible locations to read the mob cell, one is the mobWorkingDir/fileName,
archiveDir/fileName. When the mob file is not in the mobWorkingDir, HBase will try the second
location.But now we only retry after the FileNotFoundException(https://github.com/apache/hbase/blob/hbase-11339/hbase-server/src/main/java/org/apache/hadoop/hbase/regionserver/HMobStore.java#L312)
Do you means a table deletion on the original table happens in the middle of the read operation
will throw other IOExceptions? 
I don't know the HFileLink how to guarantee the case in the MOB? Can you please give a more
detailed description? Thanks

> Add a UT to read mob file when the mob hfile moving from the mob dir to the archive dir
> ---------------------------------------------------------------------------------------
>                 Key: HBASE-12673
>                 URL: https://issues.apache.org/jira/browse/HBASE-12673
>             Project: HBase
>          Issue Type: Sub-task
>          Components: regionserver, Scanners
>    Affects Versions: hbase-11339
>            Reporter: Jiajia Li
>            Assignee: Jiajia Li
>             Fix For: hbase-11339
>         Attachments: HBASE-12673.patch
> add a unit test to scan the cloned table when deleting the original table, and the steps
as following:
> 1) create a table with mobs, 
> 2) snapshot it, 
> 3) clone it as a a different table
> 4) have a read workload on the snapshot
> 5) delete the original table

This message was sent by Atlassian JIRA

View raw message