hbase-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Todd Lipcon" <t...@cloudera.com>
Subject Re: Review Request: HBASE-2670. Memstore should retain multiple versions of a row when memstore TS differs
Date Tue, 15 Jun 2010 22:33:49 GMT


> On 2010-06-15 15:03:40, stack wrote:
> > src/main/java/org/apache/hadoop/hbase/KeyValue.java, line 558
> > <http://review.hbase.org/r/180/diff/1/?file=1345#file1345line558>
> >
> >     Is this going to be correct always?  Cloning, we don't want the src memstoreTS?

yea, this clone is necessary so that incrementColumnValue() doesn't blow up to create a new
memstore entry for every increment. We could put it in that code, but I think it's cleaner
to just be part of clone. I think it's best in clone so that a foo.compareTo(foo.clone())
== 0 as an invariant.


> On 2010-06-15 15:03:40, stack wrote:
> > src/main/java/org/apache/hadoop/hbase/KeyValue.java, line 1306
> > <http://review.hbase.org/r/180/diff/1/?file=1345#file1345line1306>
> >
> >     What we going to do about N versions all of same r/f/q/ts but of different memstoreTS?
 We're not going to suppress them just yet?  We're going to punt till hbase-1485?  The multiple
versions make it out to store files too?

Yea, leaving that off since the problem already exists and it's a bit of a heavy change.


> On 2010-06-15 15:03:40, stack wrote:
> > src/test/java/org/apache/hadoop/hbase/TestAcidGuarantees.java, line 140
> > <http://review.hbase.org/r/180/diff/1/?file=1348#file1348line140>
> >
> >     Prefix w/ 'TODO' to make this work-to-do more findable.

ah, I mean that we don't need to verify at this point because the writer hasn't actually written
any rows yet! will clarify


- Todd


-----------------------------------------------------------
This is an automatically generated e-mail. To reply, visit:
http://review.hbase.org/r/180/#review229
-----------------------------------------------------------


On 2010-06-14 16:30:30, Todd Lipcon wrote:
> 
> -----------------------------------------------------------
> This is an automatically generated e-mail. To reply, visit:
> http://review.hbase.org/r/180/
> -----------------------------------------------------------
> 
> (Updated 2010-06-14 16:30:30)
> 
> 
> Review request for hbase and Ryan Rawson.
> 
> 
> Summary
> -------
> 
> This changes the memstore comparator, improves the atomicity tests, and also fixes ICV
to continue to have the right behavior even with the comparator change.
> 
> 
> This addresses bug hbase-2670.
>     http://issues.apache.org/jira/browse/hbase-2670
> 
> 
> Diffs
> -----
> 
>   src/main/java/org/apache/hadoop/hbase/KeyValue.java 71284cf 
>   src/main/java/org/apache/hadoop/hbase/regionserver/Store.java 2a0dcee 
>   src/test/java/org/apache/hadoop/hbase/MultithreadedTestUtil.java 7c062d7 
>   src/test/java/org/apache/hadoop/hbase/TestAcidGuarantees.java 75f3c8b 
>   src/test/java/org/apache/hadoop/hbase/regionserver/TestMemStore.java 9833d76 
>   src/test/java/org/apache/hadoop/hbase/regionserver/TestStoreScanner.java 0566af7 
> 
> Diff: http://review.hbase.org/r/180/diff
> 
> 
> Testing
> -------
> 
> Unit tests.
> 
> 
> Thanks,
> 
> Todd
> 
>


Mime
View raw message