ADSM-L

Re: LTO restore performance - seek problem?

2001-05-10 08:58:09
Subject: Re: LTO restore performance - seek problem?
From: Rejean Larivee/Quebec/IBM <rlarivee AT CA.IBM DOT COM>
Date: Thu, 10 May 2001 09:58:24 -0300
Hello,
you need the PGA3 microcode which should be
available early next week, date subject to change.
This is the microcode that should address the
performance issue you are seeing.
The 12U7.fmr microcode listed below addresses
another issue (read error) with the LTO drive but does
not address the performance issue yet.
The PGA3 microcode, when it comes out, should also
include the 12U7 microcode changes.

-----------------------------------------------------------------
Rejean Larivee / IBM
Rejean Larivee / IBM
TSM/ADSM Level 2 Support





                    Ray
                    <ray AT OPS DOT SELU.       To:     ADSM-L AT VM.MARIST DOT 
EDU
                    EDU>                 cc:
                    Sent by:             Subject:     Re: LTO restore 
performance - seek problem?
                    "ADSM: Dist
                    Stor Manager"
                    <ADSM-L AT VM DOT MAR
                    IST.EDU>


                    05/10/01 05:32
                    AM
                    Please respond
                    to "ADSM: Dist
                    Stor Manager"





Hello all,

I applied OCE1 microcode to 2 3583 drives recently, and i think
performance is suffering now!  Backup seems fine, but tape-to-tape
operations is taking forever...  perhaps restore is suffering as well, i
have not tried any large restores.

I'm seeing the same behavior Rejean describes below.  A few files will be
copied, then a pause for 1-2 minutes, then a few more files.  In backing
up a primary storage pool to copy storage pool, it only backed up 60 gigs
in 9 hours. Does this seem too slow? (it does to me...)

I noticed at ftp://service.boulder.ibm.com/storage/358x new microcode
12U7.fmr was released on 5/7.  Can anyone confirm if this will fix my
problem? Also...what is the 12U7.ro file?

--Ray
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Ray DeJean                                       http://www.r-a-y.org
Systems Administrator               Southeastern Louisiana University
IBM Certified Specialist              AIX Administration, AIX Support
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=


On Thu, 3 May 2001, Rejean Larivee/Quebec/IBM wrote:

> Hello John,
> I have worked with one customer getting the same
> behaviour on restore, even with the 0C1 microcode
> applied. Small bursts of files being restored at a time.
>
> We have captured TSM and Atape traces and found
> TSM to be waiting for LOCATE commands to complete.
> We found this to be an LTO  microcode problem.
> A new microcode has been developed and is currently
> going through final stages of testing.
>
> >From my last conversation with the hardware engineers
> involved, our customer should get this new microcode
> early next week.
>
> I will update the list as soon as our customer gets
> the microcode installed and has a chance to test it.
> I will also let you know what the "official" microcode
> number is once I have it.
>
> Regards,
>
> -----------------------------------------------------------------
> Rejean Larivee / IBM
> TSM/ADSM Level 2 Support