Mailing-List: contact general-help@hadoop.apache.org; run by ezmlm
Precedence: bulk
Reply-To: general@hadoop.apache.org
Received-SPF: neutral (athena.apache.org: local policy)
Content-Type: text/plain; charset="us-ascii"
MIME-Version: 1.0 (Apple Message framework v1082)
Subject: Re: [DISCUSS] Hadoop Security Release off Yahoo! patchset
From: Eric Baldeschwieler <eric14@yahoo-inc.com>
In-Reply-To: <AANLkTik+oTnp69ZNF7Hgtom1pTTBy+J_m=R8cQ3X+1W0@mail.gmail.com>
Date: Mon, 17 Jan 2011 15:29:10 -0800
Content-Transfer-Encoding: quoted-printable
Message-ID: <1BBA9772-9F30-44A1-A979-387D7891903E@yahoo-inc.com>
References: <516684F5-0052-4381-805D-760B61DECB16@yahoo-inc.com>
 <366A9E58-5BD7-497D-9AE1-229959ED4065@apache.org>
 <18C5C999-4680-4684-BC55-A430C40FD746@yahoo-inc.com>
 <AANLkTim+y7CBi8WeV9W+FVckn-1D1SAV2hSc6RnEebSn@mail.gmail.com>
 <F8414465-24F2-4F40-97F1-B13048820F24@yahoo-inc.com>
 <D7FA8ED2-D9C6-4B45-AC3B-9E965ADD9F5D@apache.org>
 <2A07F1E6-7096-493B-B92E-89938689DD50@yahoo-inc.com>
 <5CDDF962-5828-459F-87C3-5033EC21E9BF@mac.com>
 <075308A1-129B-4BF7-8924-C04EC6106D3E@yahoo-inc.com>
 <AANLkTimbXjRTTmEXiY1wweFfNGKmc5aWKw9D03J7gjBg@mail.gmail.com>
 <388582DF-FC85-49D1-A89C-1F36CE34A0E2@yahoo-inc.com>
 <AANLkTim=+8_WBNbxU43xwH50WT8UQwbac_mR8FQ+0jjz@mail.gmail.com>
 <04705B3C-49A9-46B3-8AA9-5673EFBDE544@yahoo-inc.com>
 <AANLkTinAk+8YNLHb580uFibMWtNc=oYdqq+sOdBc+=6U@mail.gmail.com>
 <74BDFA74-DB12-4109-89DF-B353FC7296C4@yahoo-inc.com>
 <AANLkTimLLrfYbaaOzbEV9dcsaMFLeMmdgx73FasdxGhA@mail.gmail.com>
 <B4C8E6B2-A343-449B-AE63-DE5E3F883ABA@yahoo-inc.com>
 <00F14279-2802-4CC3-91AC-481AE257FD8B@yahoo-inc.com>
 <AANLkTikqbRrU+3+y2nctu7RzKNb7+P8kd=4az+eQHzKW@mail.gmail.com>
 <58FF3525-B734-44A0-A5C6-45282A23F06F@yahoo-inc.com>
 <AANLkTik-=ZDac0iEbYuDHeBrRosQ3JNGvRbf7MxexQND@mail.gmail.com>
 <7D4397E2-B618-4BD1-8E1E-08B1C598B76F@yahoo-inc.com>
 <10BA9684-51FF-4332-A2FD-5F648E0AAF8C@Holsman.NET>
 <4E3C214A-16C6-46FF-908B-8EFB075A24E0@yahoo-inc.com>
 <AANLkTik+oTnp69ZNF7Hgtom1pTTBy+J_m=R8cQ3X+1W0@mail.gmail.com>
To: "general@hadoop.apache.org" <general@hadoop.apache.org>

Hi Stack,

I feel your pain.  We're running a 700 node HBASE cluster containing a =
HUGE collections of all web pages.  Both versions of append were started =
by engineers working at yahoo and we've put A LOT of investment into =
both.  I really, really want to see the append issue solved for HBASE!! =20=


My point is simply that we need to separate our concerns.  I would 300% =
support a community of folks building a 0.20 derived version of hadoop =
with append and we know that any new release post 0.20 will contain an =
append solution.  This branch is more backwards facing.  We are simply =
trying to share our last two years of 0.20 experience with the =
community, so that a) folks can use it if they find value in it, b) this =
work can be merged into future hadoop releases (that will have append).

We want to share what we have tested, since we believe that the testing =
is a good chunk of our contribution.

Thanks,

E14

On Jan 16, 2011, at 2:57 PM, Stack wrote:

> On Fri, Jan 14, 2011 at 10:25 AM, Eric Baldeschwieler
> <eric14@yahoo-inc.com> wrote:
>> 2) append is hard. It is so hard we rewrote the entire write pipeline =
(5 person-years work) in trunk after giving up on the codeline you are =
suggesting we merge in. That work is what distinguishes all post 20 =
releases from 20 releases in my mind. I dont trust the 20 append code =
line. We've been hurt badly by it.  We did the rewrite only after losing =
a bunch of production data a bunch of times with the previous code line. =
 I think the various 20 append patch lines may be fine for specialized =
hbase clusters, but they doesn't have the rigor behind them to bet your =
business in them.
>>=20
>=20
> Eric:
>=20
> A few comments on the above:
>=20
> + Append has had a bunch of work done on it since the Y! dataloss of a
> few years ago on an ancestor of the branch-0.20-append codebase (IIRC
> the issue you refer to in particular -- the 'dataloss' because
> partially written blocks were done up in tmp dirs, and on cluster
> restart, tmp data was cleared -- has been fixed in
> branch-0.20.append).
> + You may not trust 0.20-append (or its close cousin over in CDH) but
> a bunch of HBasers do. On the one hand, we have little choice.  Until
> the *new* append becomes available in a stable Hadoop the HBase
> project has had to sustain itself (What you think?, 3-6 months before
> we see 0.22?  HBase project can't hold its breath that long).  On
> other hand, the branch-0.20-append work has been carried out by lads
> (and lasses!) who know their HDFS.  Its true that it will not have
> been tested with Y! rigor but near-derivatives -- CDH or the FB
> branches -- already do HDFS-200-based append in production.
>=20
> St.Ack
> P.S. Don't get me wrong.  HBase is looking forward to *new* append.
> We just need something to suck on meantime.