X-MimeOLE: Produced By Microsoft Exchange V6.5
Received: by onstor-exch02.onstor.net 
	id <01C81A76.7CD6503F@onstor-exch02.onstor.net>; Mon, 29 Oct 2007 14:55:58 -0700
MIME-Version: 1.0
Content-Type: text/plain;
	charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Content-class: urn:content-classes:message
Subject: RE: Online eek functional spec
Date: Mon, 29 Oct 2007 14:55:58 -0700
Message-ID: <BB375AF679D4A34E9CA8DFA650E2B04E0648D189@onstor-exch02.onstor.net>
In-Reply-To: <20071029142926.7ea256ca@ripper.onstor.net>
X-MS-Has-Attach: 
X-MS-TNEF-Correlator: 
Thread-Topic: Online eek functional spec
Thread-Index: AcgacsfqqB2ghkXxSsSfEvlTy5TpaAAA0Dgw
References: <BB375AF679D4A34E9CA8DFA650E2B04E062AD31F@onstor-exch02.onstor.net><BB375AF679D4A34E9CA8DFA650E2B04E063842D9@onstor-exch02.onstor.net><BB375AF679D4A34E9CA8DFA650E2B04E06384346@onstor-exch02.onstor.net><BB375AF679D4A34E9CA8DFA650E2B04E0648D079@onstor-exch02.onstor.net><BB375AF679D4A34E9CA8DFA650E2B04E0648D08A@onstor-exch02.onstor.net><BB375AF679D4A34E9CA8DFA650E2B04E0648D12B@onstor-exch02.onstor.net> <20071029142926.7ea256ca@ripper.onstor.net>
From: "Eric Barrett" <eric.barrett@onstor.com>
To: "Andy Sharp" <andy.sharp@onstor.com>,
	"Brian DeForest" <brian.deforest@onstor.com>
Cc: "Jobi Ariyamannil" <jobi.ariyamannil@onstor.com>,
	"Amit Bothra" <amit.bothra@onstor.com>,
	"Sandrine Boulanger" <sandrine.boulanger@onstor.com>,
	"dl-Design Review" <dl-designreview@onstor.com>,
	"Narayan Venkat" <narayan.venkat@onstor.com>,
	"Sudheesh Nair" <sudheesh.nair@onstor.com>,
	"Huy Duong" <huy.duong@onstor.com>

> We've been lazy.  Perpetrating the evil is just more evil.=20

I agree with this, strongly.

It doesn't mean the online eek procedure is the place to do it -- I
don't have a horse in that race -- but any and all Engineering projects
should move away from using 'root' in the field, even for exceptional
situations.

In fact, running a file system check, even if we had a flawless FS, is
still NOT an exceptional situation, because you would want to do it
periodically for maintenance; disk arrays and switches and the wires
that connect them all are not perfect, and never will be.  Especially if
it were easy to do because it didn't cause downtime.



-----Original Message-----
From: Andy Sharp=20
Sent: Monday, October 29, 2007 2:29 PM
To: Brian DeForest
Cc: Jobi Ariyamannil; Amit Bothra; Sandrine Boulanger; dl-Design Review;
Narayan Venkat; Sudheesh Nair; Eric Barrett; Huy Duong
Subject: Re: Online eek functional spec

I can't think of a reason why this should not have a standard user
interface like all other features.  That means available through the
GUI and all the rest.  Only available through root login?  Please.
We've been lazy.  Perpetrating the evil is just more evil.

Adding it to the nfxsh and whatever all else should be part of the spec
and the scope.

Just my opinion.

Cheers,

a

On Mon, 29 Oct 2007 14:20:19 -0700 "Brian DeForest"
<brian.deforest@onstor.com> wrote:

> The issue/question is whether it should be in nfxsh.   If the
> consensus is that it doesn't need to be in nfxsh, then the issue is
> closed. Maybe others have an opinion.
>=20
> > _____________________________________________=20
> > From: 	Jobi Ariyamannil =20
> > Sent:	Monday, October 29, 2007 1:28 PM
> > To:	Brian DeForest; Amit Bothra; Sandrine Boulanger;
> > dl-Design Review
> > Subject:	RE: Online eek functional spec
> >=20
> > Eek was never part of nfxsh.  We did not plan to make it part of
> > nfxsh in R98.
> > Online eek implemention itself is going to take rest of the R98 time
> > (Nov 15 feature complete).
> >=20
> > Regards,
> > Jobi
> >=20
> > _____________________________________________
> > From: Brian DeForest=20
> > Sent: Monday, October 29, 2007 1:21 PM
> > To: Amit Bothra; Sandrine Boulanger; dl-Design Review
> > Subject: RE: Online eek functional spec
> >=20
> > Maybe I'm missing something, but the incremental time it takes to
> > implement an admin command vs. diag command is maybe 2 more days at
> > most -- documentation (1 day) and internal review (1 day).
> >=20
> > Additionally, this is a highly requested feature so it presumably
> > should be exposed as an admin command.
> >=20
> > _____________________________________________=20
> > From: 	Amit Bothra =20
> > Sent:	Monday, October 29, 2007 12:03 PM
> > To:	Sandrine Boulanger; dl-Design Review
> > Subject:	RE: Online eek functional spec
> >=20
> > Hi Sandrine,
> >=20
> > I agree that its probably nice to have a CLI interface at admin
> > level for this, but given the time frame its not do-able.
> >=20
> > Thanks,
> > Amit
> >=20
> > _____________________________________________
> > From: Sandrine Boulanger=20
> > Sent: Monday, October 29, 2007 11:24 AM
> > To: Amit Bothra; dl-Design Review
> > Subject: RE: Online eek functional spec
> >=20
> > If we want to give customers the ability to run online EEK,
> > shouldn't we provide a CLI interface at admin level instead of
> > having to login as root and go to bsd to start online eek?
> >=20
> > _____________________________________________
> > From: Amit Bothra=20
> > Sent: Tuesday, October 23, 2007 11:26 PM
> > To: dl-Design Review
> > Subject: Online eek functional spec
> >=20
> > Please review the attached online eek functional specification. Any
> > feedback is highly appreciated.
> >=20
> > Thanks,
> > Amit
> >=20
> >  << File: Online EEK - Func  Spec-v4.doc >>=20
