Is the 2844 dedicated to tape or is it swinging disk arms, too? 2844 is on a PCI bus. What's the throughput of the PCI bus? Is the PCI bus shared with other controllers? How does the PCI bus connect to memory? Is the memory fast enough? 280D is supposed to be capable of 1G, 2G, or 4G depending on what's on the other end of the pipe and how far away it is. Only 1G if over 150 meters, for instance. Still, your calculations show much less than 1G throughput.
I believe that "sustained data rate" for tape means the head and the tape drive motor are capable of laying data on the tape that fast. Is the processor in the tape drive capable of managing protocol and taking data off the fibre and feeding it to the head that fast? Probably. Sustained data rate may actually refer to reading from tape and writing to fibre.
Looking at your numbers a little closer, 165Mbps is pretty darned close to 240Mbps rated speed. If you expect the tape to have a catalog, you're probably not going to achieve full sustained rate through the entire 6 hours.
Still, I think you should look at the connections between memory and the fibre port.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Thursday, June 13, 2013 3:25 PM
To: Midrange Systems Technical Discussion
Subject: RE: LTO4 vs LTO3 speed
Yes, there is a Cisco switch between the drive and the card. But wouldn't it have to be set pretty low to not exceed these times?
3741731MB of user libraries saved in 6 hours and 18 minutes is, what, (6hours * 60min/hour * 60 sec/min) + (18min * 60sec/min) =
21600+1080=22,680sec
3741731MB/22,680sec=165MB/s
Comparing that with:
LTO4 Sustained data transfer rate* Up to 240 Mbps compressed*; 120
Mbps native
What's next, figure out if the BRMS save is compressed? Figure that the "sustained data rate" is for saving one big library natively compressed and not with the overhead of BRMS being able to tell you each individual object on each tape, etc?
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From: Dan Kimmel <dkimmel@xxxxxxxxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>,
Date: 06/13/2013 04:11 PM
Subject: RE: LTO4 vs LTO3 speed
Sent by: midrange-l-bounces@xxxxxxxxxxxx
At this point we have to look to the systems guys. Larry, Jim, Pete? Could
be either the 2844 or the 280D or the stuff that connects them. Could also
be something completely off the box (switches, tape drive internals, ???).
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [
mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Thursday, June 13, 2013 3:02 PM
To: Midrange Systems Technical Discussion
Subject: RE: LTO4 vs LTO3 speed
WRKHDWRSC *STG
Resource Type-model Status Text
CMB04 2844-001 Operational Storage Controller
DC05 280D-001 Operational Tape Controller
TAPGRT01 3573-040 Operational Tape Library
2844 Location : U0595.001.10782A9-CB1-C03
280D Location : U0595.001.10782A9-CB1-C04
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail
to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From: Dan Kimmel <dkimmel@xxxxxxxxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>,
Date: 06/13/2013 03:56 PM
Subject: RE: LTO4 vs LTO3 speed
Sent by: midrange-l-bounces@xxxxxxxxxxxx
I doubt if it's CPU. At least not main CPU. There's a lot of plumbing
between the main CPU and I/O devices. The CPU probably loads a block into
memory, then passes the address to I/O controller processor which copies
memory to the I/O port and handles protocol. So you're dependent on the
speed of that I/O processor and all the channels and busses that connect
it to memory and to the I/O port.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [
mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Thursday, June 13, 2013 2:51 PM
To: Midrange Systems Technical Discussion
Subject: RE: LTO4 vs LTO3 speed
Perhaps the CPU? Perhaps a Power 6 9117-MMA P30 just ain't man enough to
drive this bad boy?
Resource Type-model Status Text
CEC01 9117-MMA Operational Main Card Enclosure
PN03 28D4 Operational System Control Panel
MP13 53CF Operational System Processor Card
MP14 53CF Operational System Processor Card
MP15 53CF Operational System Processor Card
MP16 53CF Operational System Processor Card
MP17 53CF Operational System Processor Card
MP18 53CF Operational System Processor Card
MP19 53CF Operational System Processor Card
MP20 53CF Operational System Processor Card
PV01 52AD Operational Processor Capacity Card
SP04 294E Operational Service Processor Card
SP05 294E Operational Service Processor Card
...
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail
to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
From: Dan Kimmel <dkimmel@xxxxxxxxxxxxxxx>
To: Midrange Systems Technical Discussion <midrange-l@xxxxxxxxxxxx>,
Date: 06/13/2013 03:42 PM
Subject: RE: LTO4 vs LTO3 speed
Sent by: midrange-l-bounces@xxxxxxxxxxxx
Not much difference. Leads me to believe (or suspect, anyway) your
performance bottleneck is not the speed of the tape, drive, or connection
media.
-----Original Message-----
From: midrange-l-bounces@xxxxxxxxxxxx [
mailto:midrange-l-bounces@xxxxxxxxxxxx] On Behalf Of rob@xxxxxxxxx
Sent: Thursday, June 13, 2013 2:23 PM
To: midrange-l@xxxxxxxxxxxx
Subject: LTO4 vs LTO3 speed
I read a chart that says basically:
LTO1: Up to 40MB/s
LTO2: Up to 80MB/s
LTO3: Up to 160MB/s
LTO4: Up to 240MB/s
LTO5: Up to 360MB/s
LTO6: Up to 540MB/s
http://www-304.ibm.com/services/weblectures/dlv/partnerworld/online/ltu29864/mod%208%20ibm%20system%20storage%20tape%20part%203.ppt
And another chart that says the drive makes no difference - it's totally
dependent on the interface (we use 280D or 5774) (something smells rotten
about this claim).
http://www-01.ibm.com/support/docview.wss?uid=nas1b69de8ff8246ff2c8625688700566393
This chart says LTO4 Sustained data transfer rate* Up to 240 Mbps
compressed*; 120 Mbps native
http://www-03.ibm.com/systems/storage/tape/oem/lto4/specifications.html
Another chart says LTO3 Sustained data transfer rate1: Up to 60 Mbps
native
We have the LTO4 drives in question mounted in a 3573-L4U (aka TS3200)
http://www-03.ibm.com/systems/storage/tape/ts3200/index.html
Moved from scsi LTO3 to fiber LTO4.
Full system save completed with BRMS:
03/06/13 06/05/13
LTO3 LTO4
SAVSYS 00:07 00:07
*IBM 00:06 00:05
*ALLUSR 06:06 06:18
*LINK 00:20 00:19
Another system:
LTO3 LTO4
SAVSYS 00:18 00:19
*IBM 00:09 00:08
*ALLUSR 02:28 02:34
*LINK 01:30 01:23
Granted, there was a three month time difference in between there.
Utilization may have changed.
System 1:
Information collected . . . . . . . . . : 04/11/13 04:00:02
% of Size in
Description Disk 1,000,000 bytes
User libraries 77.11 3604127.46
User directories 1.50 70014.59
Folders and documents .00 .14
QSYS .10 4481.52
Other IBM libraries .39 18235.97
Licensed Internal Code .11 5182.31
Temporary space .39 18319.04
Unused space 20.41 953978.86
System internal objects .16 7509.85
Objects not in a library .00 .00
TOTAL 100.17 4681849.74
Information collected . . . . . . . . . : 06/13/13 04:00:02
% of Size in
Description Disk 1,000,000 bytes
User libraries 80.05 3741731.66
User directories 1.56 72747.25
Folders and documents .00 .14
QSYS .10 4596.45
Other IBM libraries .38 17809.87
Licensed Internal Code .11 5199.08
Temporary space .31 14563.16
Unused space 17.57 821259.54
System internal objects .20 9196.28
Objects not in a library .00 .00
TOTAL 100.28 4687103.43
System 2:
Information collected . . . . . . . . . : 04/11/13 04:00:04
% of Size in
Description Disk 1,000,000 bytes
User libraries 71.73 650784.59
User directories 7.37 66912.19
Folders and documents .00 16.75
QSYS .53 4825.70
Other IBM libraries 1.67 15128.99
Licensed Internal Code .54 4913.39
Temporary space 2.45 22214.63
Unused space 14.51 131641.09
System internal objects .98 8872.16
Objects not in a library .00 .01
TOTAL 99.78 905309.50
Information collected . . . . . . . . . : 05/30/13 04:00:08
% of Size in
Description Disk 1,000,000 bytes
User libraries 70.35 638249.19
User directories 7.36 66821.73
Folders and documents .00 16.75
QSYS .77 6973.45
Other IBM libraries 1.90 17223.31
Licensed Internal Code .54 4913.39
Temporary space 2.49 22568.05
Unused space 15.78 143215.09
System internal objects 1.02 9224.12
Objects not in a library .00 .00
TOTAL 100.21 909205.08
Rob Berendt
--
IBM Certified System Administrator - IBM i 6.1 Group Dekko Dept 1600 Mail
to: 2505 Dekko Drive
Garrett, IN 46738
Ship to: Dock 108
6928N 400E
Kendallville, IN 46755
http://www.dekko.com
--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing
list To post a message email: MIDRANGE-L@xxxxxxxxxxxx To subscribe,
unsubscribe, or change list options,
visit:
http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx Before posting, please take a
moment to review the archives at
http://archive.midrange.com/midrange-l.
midrange.com
