AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20080629182911.55aa1c30@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:onstor-exch02.onstor.net
NSV:
SSH:
R:<sandrine.boulanger@onstor.com>,<jonathan.goldick@onstor.com>,<vikas.saini@onstor.com>,<ed.kwan@onstor.com>,<paul.hammer@onstor.com>,<brian.stark@onstor.com>,<maxim.kozlovsky@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#imap/andys@onstor.net@onstor-exch02.onstor.net/INBOX	0	BB375AF679D4A34E9CA8DFA650E2B04E0AAC5A9E@onstor-exch02.onstor.net
X-Sylpheed-End-Special-Headers: 1
Date: Sun, 29 Jun 2008 18:33:57 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: "Sandrine Boulanger" <sandrine.boulanger@onstor.com>
Cc: "Jonathan Goldick" <jonathan.goldick@onstor.com>, "Vikas Saini"
 <vikas.saini@onstor.com>, "Ed Kwan" <ed.kwan@onstor.com>, "Paul Hammer"
 <paul.hammer@onstor.com>, "Brian Stark" <brian.stark@onstor.com>, "Maxim
 Kozlovsky" <maxim.kozlovsky@onstor.com>
Subject: Re: sub28beta tarball locations
Message-ID: <20080629183357.0d74615e@ripper.onstor.net>
In-Reply-To: <BB375AF679D4A34E9CA8DFA650E2B04E0AAC5A9E@onstor-exch02.onstor.net>
References: <BB375AF679D4A34E9CA8DFA650E2B04E06CD8B7F@onstor-exch02.onstor.net>
	<BB375AF679D4A34E9CA8DFA650E2B04E0AAC5A9E@onstor-exch02.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

On Sun, 29 Jun 2008 13:17:15 -0700 "Sandrine Boulanger"
<sandrine.boulanger@onstor.com> wrote:

> I can try. Do we have a 28dev build?

We do now:

bobcat

R3.3.0.DEV-BCDBG-062908.tar.gz
R3.3.0.DEV-BC-062908.tar.gz

bobcat-linux

R4.0.0.DEV-BLDBG-062908.tar.gz
R4.0.0.DEV-BL-062908.tar.gz

cougar

R4.0.0.DEV-CGDBG-062908.tar.gz
R4.0.0.DEV-CG-062908.tar.gz

cheetah

R3.3.0.DEV-DBG-062908.tar.gz
R3.3.0.DEV-062908.tar.gz

buily against perforce label cougar-sub28dev:

Change 29893 on 2008/06/27 by jans@jans-devwork 'file system stress test changes'

> -----Original Message-----
> From: Jonathan Goldick 
> Sent: Sunday, June 29, 2008 12:53 PM
> To: Sandrine Boulanger; Andy Sharp; Vikas Saini
> Cc: Paul Hammer; Brian Stark; Maxim Kozlovsky
> Subject: Re: sub28beta tarball locations
> 
> Rcon to an fp or txrx core, run gprof_start, run the test for a few
> minutes, run gprof_stop and then gprof_dump 
> 
> I'm not sure if the above is going to be that useful since this has to
> be a bad integration.  If you run sub28dev do you see good results on
> the same machine?
> 
> -----Original Message-----
> From: Sandrine Boulanger
> To: Jonathan Goldick; Andy Sharp; Vikas Saini
> CC: Paul Hammer; Brian Stark; Maxim Kozlovsky
> Sent: Sun Jun 29 12:07:45 2008
> Subject: RE: sub28beta tarball locations
> 
> How can I get that? 
> 
> -----Original Message-----
> From: Jonathan Goldick 
> Sent: Sunday, June 29, 2008 11:44 AM
> To: Sandrine Boulanger; Andy Sharp; Vikas Saini
> Cc: Paul Hammer; Brian Stark; Maxim Kozlovsky
> Subject: Re: sub28beta tarball locations
> 
> We likely need the gprof output for the various cores.  I can only
> assume that we have an integration problem as dev runs very fast.
> 
> -----Original Message-----
> From: Sandrine Boulanger
> To: Andy Sharp; Vikas Saini; Jonathan Goldick
> CC: Paul Hammer; Brian Stark
> Sent: Sun Jun 29 11:38:52 2008
> Subject: RE: sub28beta tarball locations
> 
> I've upgraded a Cougar with the OPT build and started running
> nfsperftest.
> 
> This seems very very slow. The seqread has not completed yet but it
> seem to take forever.
> 
> The kpi for the reads are attached. Highest seems to be
> DCACHE_READ_BLOCK.
> 
> This is the vsvr stats I get for the reads:
> 
> ------------------------------------------------------------------------
> --------------------------
> 
> | G7R204-1       |    148 |      0 |      0 |      0 |   13.06K |
> 0 |        0 |        0 |
> 
> | G7R204-T1      |      0 |      0 |      0 |      0 |        0 |
> 0 |        0 |        0 |
> 
> | VS_MGMT_1883   |      0 |      0 |      0 |      0 |        0 |
> 0 |        0 |        0 |
> 
> ------------------------------------------------------------------------
> --------------------------
> 
> | G7R204-1       |    199 |      0 |      0 |      0 |   17.54K |
> 0 |        0 |        0 |
> 
> | G7R204-T1      |      0 |      0 |      0 |      0 |        0 |
> 0 |        0 |        0 |
> 
> | VS_MGMT_1883   |      0 |      0 |      0 |      0 |        0 |
> 0 |        0 |        0 |
> 
> ------------------------------------------------------------------------
> ----------   
> 
>  
> 
> Let me know if there's something else I can try or if the test is
> wrong.
> 
>  
> 
>  
> 
> TXRX1:1 > nfsperftest -p clean -c 32 -v
> vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> nfsperftest -p clean -c 32 -v vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> Start cleanup test with 32 clients per volume, and 8 volumes
> 
>  
> 
> Tests with 255 or more total clients can fail due
> 
> to mount storm issues.  Run the 'cleanup' test with the
> 
> same -c and -v values until it works to get the
> 
> mounts set up.
> 
> TXRX1:2 > Using volume 'vol1'
> 
> Using volume 'vol2'
> 
> Using volume 'vol3'
> 
> Using volume 'vol4'
> 
> Using volume 'vol5'
> 
> Using volume 'vol6'
> 
> Using volume 'vol7'
> 
> Using volume 'vol8'
> 
> 13: auth_rpc_open: 532: Auth-rpc RMC session opened again for session
> :0x100381bb00, prev_session: 0x0
> 
> NFS Perf: 256 workflows completed in 0.636571 seconds, avg/workflow
> 2486 usecs
> 
>         0 waits for memory count
> 
>         0 waits for allocatable request/memory count
> 
>  
> 
> TXRX1:3 > nfsperftest  -p seqwr -f 100000 -I 1 -c 32 -v
> vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> nfsperftest  -p seqwr -f 100000 -I 1 -c 32 -v
> vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> Rounding the file size up to 100032 KB to match the io size of 64 KB.
> 
> Start seqwrite test with:
> 
>         4 IO(s) per client
> 
>         32 clients per volume
> 
>         8 volumes
> 
>         File Size is 100032 KB
> 
>         IO Size is 64 KB
> 
>         Total IO(s) that will be issued per file is 1563
> 
>         IO Alignment Size is 8 bytes
> 
>         Commit run after writing all of the data.
> 
>         Write stability is unstable
> 
>         A unique pattern will be written to every file and block.
> 
>  
> 
> Tests with 255 or more total clients can fail due
> 
> to mount storm issues.  Run the 'cleanup' test with the
> 
> same -c and -v values until it works to get the
> 
> mounts set up.
> 
> TXRX1:4 > Using volume 'vol1'
> 
> Using volume 'vol2'
> 
> Using volume 'vol3'
> 
> Using volume 'vol4'
> 
> Using volume 'vol5'
> 
> Using volume 'vol6'
> 
> Using volume 'vol7'
> 
> Using volume 'vol8'
> 
>  
> 
>  
> 
> TXRX1:4 > 
> 
>  
> 
> TXRX1:4 > NFS Perf: 256 workflows completed in 1812.058581 seconds,
> avg/workflow 7078353 usecs
> 
>         roughly 25 GB    transferred at 14.471 MBytes/sec
> 
>         0 waits for memory count
> 
>         0 waits for allocatable request/memory count
> 
>  
> 
>  
> 
> TXRX1:4 > nfsperftest  -p seqread -f 100000 -I 1 -c 32 -v
> vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> nfsperftest  -p seqread -f 100000 -I 1 -c 32 -v
> vol1,vol2,vol3,vol4,vol5,vol6,vol7,vol8
> 
> Start seqread test with:
> 
>         4 IO(s) per client
> 
>         32 clients per volume
> 
>         8 volumes
> 
>         File Size is 100000 KB
> 
>         IO Size is 64 KB
> 
>         Total IO(s) that will be issued per file is 1562
> 
>         IO Alignment Size is 8 bytes
> 
>         All data read will be validated.
> 
>  
> 
> Tests with 255 or more total clients can fail due
> 
> to mount storm issues.  Run the 'cleanup' test with the
> 
> same -c and -v values until it works to get the
> 
> mounts set up.
> 
> TXRX1:5 > Using volume 'vol1'
> 
> Using volume 'vol2'
> 
> Using volume 'vol3'
> 
> Using volume 'vol4'
> 
> Using volume 'vol5'
> 
> Using volume 'vol6'
> 
> Using volume 'vol7'
> 
> Using volume 'vol8'
> 
>  
> 
>  
> 
> TXRX1:5 > 
> 
>  
> 
> TXRX1:5 > 
> 
>  
> 
> TXRX1:5 >
> 
>  
> 
> -----Original Message-----
> From: Andy Sharp 
> Sent: Saturday, June 28, 2008 5:42 PM
> To: Sandrine Boulanger; Vikas Saini
> Subject: sub28beta tarball locations
> 
>  
> 
>  
> 
> Here is some info:
> 
>  
> 
> Source tree location:
> 
>  
> 
> /n/Build-Trees/R4.0.0.0/R4.0.0.0-062808-sub28/nfx-tree
> 
>  
> 
> /n/Build-Trees/R4.0.0.0/R4.0.0.0-062808-sub28/linux
> 
>  
> 
>  
> 
> Tarballs for upgrades:
> 
>  
> 
> http://10.2.0.21/upgrade/R4.0.0.0CG-062808.tar.gz
> 
> http://10.2.0.21/upgrade/R4.0.0.0CGDBG-062808.tar.gz
> 
>  
> 
> The list of defects addresses I sent in another email.
> 
>  
> 
> Larry's instructions for doing the AT testing: Beg Vikas to do it!
> 
>  
> 
> So I guess that's what I have to do since I don't have a clue.
> 
>  
> 
> a
> 
