× The internal search function is temporarily non-functional. The current search engine is no longer viable and we are researching alternatives.
As a stop gap measure, we are using Google's custom search engine service.
If you know of an easy to use, open source, search engine ... please contact support@midrange.com.


  • Subject: RE: DSKBAL on V4R3
  • From: "Graap, Ken" <keg@xxxxxxxxxxxxxxxxxx>
  • Date: Wed, 7 Apr 1999 12:28:45 -0700

>I don't know whether it still works but a simple way to do this is to
>RGZPFM.  Do this on some of your bigger files and see if the new disks
>aren't picking up some of the load.  The good thing about this is that it
>can be done with the system running.

Lurton, 

You need an exclusive lock on a file to reorganize it. Other users can't use
the file while it is be reorganized. But it is better that having the whole
system down, as you need to do with DSKBAL....

The possible solution to all of our problems may be the new disk balancing
commands available in V4R4...

Here is some information about what is coming soon to an AS/400 near you...


Disk Balancing on the AS/400 (In V4R4)

(1) I have included the syntax diagram and other information for the
   STRASPBAL command at the end of this note fyi.
(2) Notice that the STRASPBAL command can be run in 3 different fashions
   to do the following:

Balance the AMOUNT of data on each HDA (eg all disks will wind up x% full)
(*CAPACITY)

Balance the UTILIZATION of each HDA (eg each arm will wind up with the same
disk-busy (which drives performance), although % full could vary widely)
(*USAGE)

 Balance the HOT/COLD data so that the most frequently used data is on
non-compressed disks, and least frequently used data is on compressed disks
(*HSM)

 (3) Notice that the *USAGE and *HSM versions of STRASPBAL must be
   preceded by a run of the TRCASPBAL command to gather disk statistics so
   OS/400 will know how to juggle the data

 (4) The TRCASPBAL command behaves a bit like the STRPFRMON command ....
   you tell it to watch your disks for some period of time and gather
   stats.  You then feed those stats into the STRASPBAL command to take
   action.  A typical customer would probably run the TRCASPBAL command for
   an hour or two during a peak disk utilization period, then run the
   STRASPBAL command shortly thereafter while the trace stats still
   accurately reflect the contents of the disk HDA's

(5) The STRASPBAL command  can be stopped and restarted as required in
   order to get all the work done.  We anticipate that for a typical
   customer, it will take a few hours for this command to complete.

(6) The V4R4 implementation of STRASPBAL is not able to move the
   following types of data: journals, temporary data, IPCS (FSIOP) data, or
   data that is pinned in main memory

(7) There was a previous tool called DSKBAL that was available via PTF
   in V4R1, V4R2, and V4R3.  It accomplished the same functions as the
   STRASPBAL *CAPACITY command and the DST/SST "add unit and balance"
   option.  However, the new V4R4 function can run concurrently with normal
   operations and is much faster.

(8) When OS/400 stores data on disk, it is normally divided into
   "chunks" called "extents" that vary in size by object type .... for
   *FILE objects, the desired extent size is 1MB.  The TRCASPBAL collects
   IO counts for 1 MB regions of the disk (it could contain multiple
   extents although for DB it would normally be the one extent) .  The  IO
   counts are used to determine the hot/cold  level of  1MB  region.  A
   customer at COMMON asked what would happen if you had a one-billion
   record file, that only had 100,000 active records .... the answer is
   that any 1MB chunk that did not contain an active record at the time the
   TRCASPBAL function was run would be treated as "cold" data, and moved
   accordingly when the STRASPBAL command was run.  (this is good news ...
   HSM within a file!!)

(9) The V4R4 implementation of STRASPBAL *HSM only considers 2 different
   "types" of storage media .... compressed dasd and non-compressed dasd.
   Thus, at the present time, OS/400 would not differentiate between
   newer/faster standard (non-compressed) HDA's and older/slower standard
   (non-compressed) HDA's (eg a 9335 dasd would be treated the same as a fc
   6607 dasd).  Note that for STRASPBAL *USAGE, OS/400 WOULD wind up
   differentiating between the older/slower dasd, since the percent busy
   numbers would reflect the slower speed dasd.
 
(10) Our developers realize that this is a new function, and we have
   lots to learn about how our customers can use it most effectively.  We
   are eager to hear how customers have benefitted from the command, and
   how they would like to see it enhanced over time.  Some initial ideas we
   are entertaining include:

 different algorithms for selecting data to move

 more differentiation among media types

 able to move more data types (eg journals, temp data, IPCS data)

 knobs and settings that customers can use to customize and control
       their ASP balancing

 (11) One caveat to beware of is that if you use *HSM on an ASP with only
   a small number of standard (non-compressed) disk arms, concentrating all
   the "hot" data onto those arms could possibly hurt performance
 

Here is the command syntax:

 STRASPBAL (START ASP BALANCE) COMMAND
 _____________________________________
Job: B,I  Pgm: B,I  REXX: B,I Exec

Threadsafe: No

>>--STRASPBAL--ASP(--.-*ALL----------------------------------.--)------->
                       | <-------------------------------< (1) |
                       '---auxiliary-storage-pool-number-------'

                                                          (P)

>--TYPE(--.-*CAPACITY-.--)--TIMLMT(--.-*NOMAX-----.--)---------------><
            |-*USAGE----|              '-time-limit-'
            '-*HSM------'

  NOTES:

(1)  A maximum of 16 repetitions.

(P) All parameters preceding this point can be specified in positional
  form.


----------------------------------------------------------------------------

  PURPOSE

  The Start ASP Balance (STRASPBAL) command allows the user to start the
auxiliary storage pool (ASP) balancing function for one or more auxiliary
storage pools (ASPs).  Three types of ASP balancing can be started:

  1.  CAPACITY BALANCING - The data on all the units within the ASP will be
      balanced so each unit has an equal percentage of used and unused
space.
      This is useful when new units are added to an ASP.  Instead of having
      several units with the majority of the data and several new ones with

      no data, the user can spread the data evenly across all the units.

  2.  USAGE BALANCING -  The high-use and low-use data on each unit in the
ASP is redistributed to balance the arm utilization of each unit within
The specified ASP.  *USAGE balancing cannot be done until the trace
controlled by the Trace ASP Balance (TRCASPBAL) command has collected 

Usage statistics.  TRCASPBAL starts the trace function that will identify
the high-use and low-use data on each unit.  This information will be used
when the Start ASP Balance (STRASPBAL) TYPE(*USAGE) command is run.

      After this balance activity has run to completion the trace
Information will be cleared.  Usage balancing is useful when the ASP
contains large capacity disk units.

  3.  HIERARCHICAL STORAGE MANAGEMENT BALANCING -  The high-use and low-use
      data on each unit in the ASP is redistributed so that the high-use
      data resides on high performance units and the low-use data resides on
low performance units.  The ASP being balanced in this manner must contain
the correct mixture of slow and fast units to perform this type of balance.
This allows the low-use data on the high performance units
to be moved to low performance units.  Data that is high-use that resides on
slow units will be moved to the fast disk units. *HSM balancing cannot be
done until the trace controlled by the Trace ASP Balance (TRCASPBAL) command
has collected usage statistics.  TRCASPBAL starts the trace function that
will identify the high-use and low-use data on each unit. This information
will be used when the Start ASP Balance (STRASPBAL) TYPE(*HSM) command is
run. After this balance activity has run to completion the trace information
will be cleared.  *HSM balancing is useful when the ASP contains compressed
disk units.

        The user may specify a time limit that the function is to run for
each
ASP being balanced or the balance can be set to run to completion.  If the
balance function needs to be ended, use the End ASP Balance (ENDASPBAL)
command.  A message will be sent to the system history (QHST) log when the
balancing function is started for each ASP.  A message will also be sent to
the QHST log when the balancing function completes or is ended.

If the balance function is run for a few hours and then stopped, it will
continue from where it left off when the balance function restarts.  This
allows the balancing to be run during off hours over a several day period.

        For more information about ASP balancing see the AS/400 Hierarchical
Storage Management guide.

        RESTRICTIONS:  You must have *ALLOBJ special authority to use this
command.


  REQUIRED PARAMETERS

  ASP
      Specifies the auxiliary storage pools for which the ASP balancing
function is to be started.

      *ALL:  ASP balancing will be started for all ASPs defined to the
system.

      auxiliary-storage-pool-number:  Specify the ASP for which ASP
balancing is to be started.  Up to sixteen ASP numbers may be specified.

  TYPE
      Specifies the type of ASP balancing that is to be done.

      *CAPACITY:  The capacity balance function will be started.

      *USAGE:  The usage balance function will be started.

      *HSM:  The hierarchical storage management balance function will be
      started.

  TIMLMT
      Specifies the amount of time, in minutes, that the ASP balancing
Function will be allowed to run. When the time limit is reached the function
will end.  For example, if ASP(*ALL) is specified and the machine has four
ASPs configured and TIMLMT(60) is specified, four balance functions are
started and each can run 60 minutes.  If the balancing of any ASP has not
completed after 60 minutes, it will be forced to end.  This allows you to do
ASP balancing incrementally.  The balance function will not run across an
IPL.

      *NOMAX:  There is no time limit for the balance function.  For a
large ASP the balance function can take a long time to complete.  If you
start the balance function with *NOMAX and you want to force the function to
end, you can use the End ASP Balance (ENDASPBAL) command.

      time-limit:  Specify the time limit that the balance function will be
allowed to run.  Valid values range from 1 to 9999 minutes.


  EXAMPLES

  EXAMPLE 1:  STARTING ASP BALANCING FOR ASP 1

    STRASPBAL  ASP(1) TIMLMT(*NOMAX) TYPE(*CAPACITY)

This command allows the user to start the ASP balance function for ASP1. The
balance function will run until the capacity for each of the units has been
balanced.

  EXAMPLE 2:  STARTING ASP BALANCING WITH A TIME LIMIT

    STRASPBAL  ASP(*ALL) TIMLMT(60) TYPE(*USAGE)

  This command allows the user to start the ASP balance function for each
ASP on the system for which the TRCASPBAL command has been run.  Each
balance
function will have a time limit of sixty minutes.  After sixty minutes,
any balance functions which have not completed will be ended.




Kenneth

--
********************************
       Kenneth  E.  Graap
    IBM Certified Specialist
          AS/400   Professional 
          System  Administrator
 NW Natural - Information Services
           System Services
        503 226 4211 X5537
          FAX  503 721 2521
      keg@nwnatural.com
********************************


-----Original Message-----
From: Lurton Keel [mailto:Lurton.Keel@SERX.COM]
Sent: Wednesday, April 07, 1999 10:52 AM
To: 'MIDRANGE-L@midrange.com'
Subject: RE: DSKBAL on V4R3
+---
| This is the Midrange System Mailing List!
| To submit a new message, send your mail to MIDRANGE-L@midrange.com.
| To subscribe to this list send email to MIDRANGE-L-SUB@midrange.com.
| To unsubscribe from this list send email to MIDRANGE-L-UNSUB@midrange.com.
| Questions should be directed to the list owner/operator: david@midrange.com
+---


As an Amazon Associate we earn from qualifying purchases.

This thread ...


Follow On AppleNews
Return to Archive home page | Return to MIDRANGE.COM home page

This mailing list archive is Copyright 1997-2024 by midrange.com and David Gibbs as a compilation work. Use of the archive is restricted to research of a business or technical nature. Any other uses are prohibited. Full details are available on our policy page. If you have questions about this, please contact [javascript protected email address].

Operating expenses for this site are earned using the Amazon Associate program and Google Adsense.