[ 
https://issues.apache.org/jira/browse/HDFS-15812?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17316524#comment-17316524
 ] 

Satya Gaurav commented on HDFS-15812:
-------------------------------------

Hi Team,

We are using the Hbase MOB files and we are facing an issue of hdfs size.

when we are calculating the size on phoenix and hdfs it's coming as different 
size.

when we are running phoenix query we are getting 3 TB size.

jdbc:phoenix:> SELECT /*+ NO_INDEX */ 
DOC_CONTENT_TYPE,sum(DOC_COMPRESSED_SIZE)/1024/1024/1024/1024 "Doc_Size as TB" 
from obst.document_metadata group by DOC_CONTENT_TYPE; 
+-------------------+-----------------+ | DOC_CONTENT_TYPE | Doc_Size as TB | 
+-------------------+-----------------+ | MOB | 3,668 | 
+-------------------+-----------------+

But when we have checked at hdfs level then it cam 14 TB

hdfs dfs -du -h -s /apps/hbase/data/mobdir/data/OBST/DOCUMENT_CONTENT/

14.1 T 42.3 T /apps/hbase/data/mobdir/data/OBST/DOCUMENT_CONTENT/

so why it is exactly like this no idea even after running the major compaction.

As a workaround when we are manually inserting a fake row and deleting that 
particular row and after deletion when we ran compaction then it's reducing the 
size.

 

flush 'OBST:DOCUMENT_CONTENT' compact 'OBST:DOCUMENT_CONTENT' major_compact 
'OBST:DOCUMENT_CONTENT', 'cfDocContent', 'MOB' major_compact 
'OBST:DOCUMENT_CONTENT'

so what could be the possible reason ?

is there any specific way to delete the hbase mob files when the TTL is forever 
and why it's showing the different size at HDFS level compare to phoenix level.

 

Regards,

Satya

> after deleting data of hbase table hdfs size is not decreasing
> --------------------------------------------------------------
>
>                 Key: HDFS-15812
>                 URL: https://issues.apache.org/jira/browse/HDFS-15812
>             Project: Hadoop HDFS
>          Issue Type: Bug
>          Components: hdfs
>    Affects Versions: 2.0.2-alpha
>         Environment: HDP 3.1.4.0-315
> Hbase 2.0.2.3.1.4.0-315
>            Reporter: Satya Gaurav
>            Priority: Major
>
> I am deleting the data from hbase table, it's deleting from hbase table but 
> the size of the hdfs directory is not reducing. Even I ran the major 
> compaction but after that also hdfs size didn't reduce. Any solution for this 
> issue?



--
This message was sent by Atlassian Jira
(v8.3.4#803005)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to