AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20080627195318.543b06e3@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:onstor-exch02.onstor.net
NSV:
SSH:
R:<brian.stark@onstor.com>,<dl-CougarCore@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
X-Sylpheed-End-Special-Headers: 1
Date: Fri, 27 Jun 2008 19:56:34 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: Brian Stark <brian.stark@onstor.com>, dl-Cougar Core Team
 <dl-CougarCore@onstor.com>
Subject: submittal #28 beta - list of inclusions.latest
Message-ID: <20080627195634.798dd3aa@ripper.onstor.net>
Organization: Onstor
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

Team,


Here is the list of bugs/changes now proposed for sub28beta.  The
filesystem/corruption changes have been pulled from the list.  All the
changes currently on this list have been integrated to the beta branch
as of now.

Once all the changes were made to integrate the filesystem changes, Jon
and Jobi had no confidence in the viability of such a result.  The
amount of divergence from dev branch to cg_beta branch in the
filesystem code during the past several weeks has made such a three
way merge a risky proposition.

Software development's recommendation is to take the list of items
below, and then switch to dev/release branch from here on out.  The QA
testing strategy for that would be a week of general functional testing
and regression of the fixed issues utilizing 2 clusters.  Consider that
option (A).

There are only two other options worth putting out for team
consideration:

B) Merge all of the nfx-tree/code/sm-fs/ directory into the cg_beta
branch, and go from there.  Dev considers this to be more risky than
(A) but doable.  QA test results might be informative.

C) Try to put together a frankestein three way merge utilizing 3-ish
changelists from the dev branch to try to address the possibility of
corruption on a beta customer machine if they hit the right combination
of unclean volume shutdowns and log replays.  I believe I heard the
word "crazy" used when analyzing this option after seeing the first
attempt at creating the merged code.  QA test results would likely be
useless because the actual testing requirements would be unbounded.
This creature has not been compiled or unit tested.


Cheers,

a


It will contain fixes for at least the following issues:

> TED00024346 NFS perf. occassionally drops to 0.

Change 29506 by jobia@jobi:jobi2 on 2008/06/03 10:26:26

> TED00024174 (8394 - Onstor) Several volume exceptions due
> to array issue      

LSI firmware WAR
Change 29760 by billn@billn-dev on 2008/06/18 15:40:58

> TED00024253     Csoak: Too many exceptions due to
> fs_volReadWriteQueue: Unable to send the read requests to the storage

Send early response to ea for long log replays.
Change 29798 by jobia@jobi:jobi on 2008/06/20 15:30:03

> TED23467: Clean up messages at boot time  

Change 29427 by larrys@larrys-r14-dmip on 2008/05/27 11:42:00
Change 29566 by rendellf@rendellf-test on 2008/06/05 14:52:34
Change 29570 by rendellf@rendellf-test on 2008/06/05 21:06:15


Pending a small amount of testing by Raj on the nightly build of dev
branch, the following change will also go in:

Change 29844 by billn@billn-dev on 2008/06/25 17:22:53

	Protection code added to make sure there is never a 
	false positive on an I/O read or write operation.


> TED00024445 - Cougar read data performance problem

Change 29855 by maximk@maximk-13 on 2008/06/26 11:23:46
	24445 Do not hold spinlock while copying the data.
