[ 
https://issues.apache.org/jira/browse/OMID-131?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16816843#comment-16816843
 ] 

Lars Hofhansl commented on OMID-131:
------------------------------------

Here are two otherwise identical table. For TEST_T I used Tephra and for TEST_O 
I used Omid:
{code:java}
27825496  27825496  /hbase/data/default/TEST_O
14622248  14622248  /hbase/data/default/TEST_T
{code}
The storage is nearly doubled. This also reflects in the scan times:
{code:java}
0: jdbc:phoenix:localhost> select count(*) from test_o;
+-----------+
| COUNT(1)  |
+-----------+
| 262144    |
+-----------+
1 row selected (2.563 seconds)

0: jdbc:phoenix:localhost> select count(*) from test_t;
+-----------+
| COUNT(1)  |
+-----------+
| 262144    |
+-----------+
1 row selected (0.72 seconds)
{code}

Both tables are fully flushed and compacted.
Not saying that Tephra is better (it's not, it has other issues), just that 
there's a lot of room for improvement.

> Enhance row level conlict detection to reduce client memory footprint.
> ----------------------------------------------------------------------
>
>                 Key: OMID-131
>                 URL: https://issues.apache.org/jira/browse/OMID-131
>             Project: Apache Omid
>          Issue Type: Improvement
>    Affects Versions: 1.0.1
>            Reporter: Yonatan Gottesman
>            Assignee: Yonatan Gottesman
>            Priority: Major
>          Time Spent: 20m
>  Remaining Estimate: 0h
>
> Currently even if omid is working in row level conflict detection, all the 
> cells are kept by the client until after commit for writing shadow cells or 
> removing them after abort.
> Instead the client should store only the rows or only the family.
> If we choose to store only the row, the client memory footprint will be 
> smaller, but if a region server wants to query a shadow cell for a cell it 
> might have to query a different region.
> On the other hand client can store the famelies of a writeset. This will have 
> a bigger footprint if we have lots of families but if a region queries a 
> shadow cell it will be on the same region.
>  
>  



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Reply via email to