Hi Pierre, et al.
We are in the process of adding a new VTL to our TSM server, and we are trying
to understand if there is a way to configure the drives that will spread the
I/O load across the available fcsi ports. The AIX server has 4 4GB fibre
connections to the VTL appliance. We are emulating a Quantum P3000 library
with 64 drives. When we initially configure the library with all 64 drives,
they are divided across the 4 fibre ports with
port 1 having drive1-drive16,
port2 drive17-drive32,
port3 drive33-drive48,
port4 drive49-drive64.
We thought this may not be an optimal setup if TSM uses a round-robin
algorithm. So, we recreated the configuration. This time we defined 16 drives
initially, which were then spread across the 4 fibre ports with port1
drive1-drive4,
port2 drive5-drive8,
port3 drive9-drive12,
port4 drive13-drive16.
Then we went through the process 3 more times until we had 64 drives defined.
This results in the having the drives defined across the 4 ports as:
port1 drive1-drive4, drive17-drive20, drive33-drive36, drive49-drive52
port2 drive5-drive8, drive21-drive24, drive37-drive40, drive53-drive56
port3 drive9-drive12, drive25-drive28, drive41-drive44, drive57-drive60
port4 drive13-drive16, drive29-drive32, drive45-drive48, drive61-drive64
This seems like a better configuration for a round-robin drive assignment. It
was at this point that we thought we should try to learn how the drives are
assigned.
Thanks for any suggestions or pointers.
Larry
-----Original Message-----
From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] On Behalf Of
CAYE PIERRE
Sent: Thursday, August 23, 2007 5:36 AM
To: ADSM-L AT VM.MARIST DOT EDU
Subject: Re: How does a library choose a drive?
Hi,
>From my observations, it depend of the action TSM is doing.
For example, when TSM is in initialisation state, there is a validation of all
the drives, in element addresses order.
When TSM is doing an audit libr checkl=y, it work the same
For that cases, and others, concerning ALL drives, TSM seems to act in element
adresses order
BUT when doing recurent task, as backup and so, TSM SEEMS to act as Richard
explain, asking the OS for a drive, so it's the OS which make the choice, based
on it own algorythm.
Larry, what are the reason why you want to know that point ?
Pierre
> -----Message d'origine-----
> De : ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU] De
> la part de Kelly Lipp
> Envoyé : mercredi 22 août 2007 20:56
> À : ADSM-L AT VM.MARIST DOT EDU
> Objet : Re: [ADSM-L] How does a library choose a drive?
>
> I'm with Paul: TSM chooses the drive and based on my
> observations always does round robin. Not LRU as I don't
> think it keeps track of that. So it will choose the next
> drive in order as long at that drive isn't in use. If in
> use, skip to the next and keep going around.
>
> You see this on label libvol commands in particular.
>
>
> Kelly J. Lipp
> VP Manufacturing & CTO
> STORServer, Inc.
> 485-B Elkton Drive
> Colorado Springs, CO 80907
> 719-266-8777
> lipp AT storserver DOT com
>
> -----Original Message-----
> From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]
> On Behalf Of Paul Zarnowski
> Sent: Wednesday, August 22, 2007 12:00 PM
> To: ADSM-L AT VM.MARIST DOT EDU
> Subject: Re: [ADSM-L] How does a library choose a drive?
>
> Sorry to disagree, but I don't see how AIX can choose the
> drive - I believe that TSM chooses it. I don't know if it's
> LRU or round-robin, but I do know that it attempts to spread
> the load evenly.
>
> At 01:56 PM 8/22/2007, Mcnutt, Larry E. wrote:
> >Thanks Kelly, Richard and Andy for the quick responses. So, as I
> >understand it in my scenario, it is the AIX host that
> chooses the drive
>
> >that is the "least recently used" one. At the very
> beginning, it will
> >start at the first drive and sequentially go through them
> all and then
> >it will begin picking the "least recently used" drive to
> keep the usage
>
> >spread evenly across all the drives.
> >
> >-----Original Message-----
> >From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]
> On Behalf
> >Of Andy Huebner
> >Sent: Wednesday, August 22, 2007 12:26 PM
> >To: ADSM-L AT VM.MARIST DOT EDU
> >Subject: Re: How does a library choose a drive?
> >
> >"Least Recently Used" I think is the management level explanation of
> >how it is done.
> >
> >Andy Huebner
> >
> >-----Original Message-----
> >From: ADSM: Dist Stor Manager [mailto:ADSM-L AT VM.MARIST DOT EDU]
> On Behalf
> >Of Richard Sims
> >Sent: Wednesday, August 22, 2007 10:57 AM
> >To: ADSM-L AT VM.MARIST DOT EDU
> >Subject: Re: [ADSM-L] How does a library choose a drive?
> >
> >On Aug 22, 2007, at 11:32 AM, Mcnutt, Larry E. wrote:
> >
> > > TSM Server 5.3.3.0
> > > AIX 5.3
> > >
> > > This is probably not a TSM specific question, but
> hopefully someone
> > > can point me to the right place to look.
> > > We would like to understand how the library decides which
> available
> > > drive to use. ...
> >
> >It doesn't decide: the host determines which drive to use.
> >TSM uses them rotationally, to balance wear.
> >
> >For a 3494 library, see the MTIOCLM (Library Mount) library
> system call
>
> >in the IBM SCSI Device Drivers Programming Reference for the
> >nitty-gritty. For a SCSI library, where the host has to
> engage in more
>
> >intricate control of library mechanics, I believe it's the
> >SMCIOC_MOVE_MEDIUM call (Transport a cartridge from one element to
> >another element).
> >
> > Richard Sims
> >
> >
> >This e-mail (including any attachments) is confidential and may be
> >legally privileged. If you are not an intended recipient or an
> >authorized representative of an intended recipient, you are
> prohibited
> >from using, copying or distributing the information in this
> e-mail or
> >its attachments. If you have received this e-mail in error, please
> >notify the sender immediately by return e-mail and delete
> all copies of
>
> >this message and any attachments.
> >Thank you.
> >
> >-----------------------------------------
> >This message and any attachments are intended for the individual or
> >entity named above. If you are not the intended recipient, please do
> >not forward, copy, print, use or disclose this communication
> to others;
>
> >also please notify the sender by replying to this message, and then
> >delete it from your system. The Timken Company / The Timken
> Corporation
>
>
> --
> Paul Zarnowski Ph: 607-255-4757
> Manager, Storage Services Fx: 607-255-8521
> 719 Rhodes Hall, Ithaca, NY 14853-3801 Em: psz1 AT cornell DOT edu
>
|