hbase-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Jonathan Gray <jg...@facebook.com>
Subject RE: Update 2 versions of a cell : same issue as HBASE-1485 ?
Date Fri, 20 Aug 2010 21:19:44 GMT
Seems to me this would be fixed by HBASE-1485.  There should be no difference whether you are
updating one version or multiple, a fix for HBASE-1485 would address either.

> -----Original Message-----
> From: jdcryans@gmail.com [mailto:jdcryans@gmail.com] On Behalf Of Jean-
> Daniel Cryans
> Sent: Friday, August 20, 2010 10:47 AM
> To: user@hbase.apache.org
> Subject: Re: Update 2 versions of a cell : same issue as HBASE-1485 ?
> 
> I think that this is stretching the usage of timestamps... but would
> you be able to write a small unit test that demonstrates it? Also did
> you try it on both 0.20 and 0.89?
> 
> Thx,
> 
> J-D
> 
> On Fri, Aug 20, 2010 at 6:18 AM, Evert Arckens <evert@outerthought.org>
> wrote:
> > Hi,
> >
> > I have a scenario of which I'm not sure if it is just another example
> of
> > issue HBASE-1485, or if it is another but related issue :
> >
> > Write 3 cells of a column, at timestamp 1, 2 and 3.
> > Then update the cells at timestamp 1 and 2 (more or less at the same
> > moment).
> >
> > Result is that the update of the cell at timestamp 1 gets lost.
> > If I would perform a major compaction between both updates, the
> update on
> > timestamp 1 wouldn't get lost.
> >
> > The reason I'm not sure if it is exactly the same issue as HBASE-1485
> is
> > because that one just talks about updates on 1 cell, not on a
> combination
> > updates on 2 cells at different timestamps.
> >
> > Regards,
> > Evert Arckens.
> >

Mime
View raw message