6 Replies Latest reply on Mar 27, 2014 10:57 AM by Fred Novakov

    Mainview History Datasets: How to control and inspect?

    Sieghart Seith
      Share This:

      Sometimes I have to think about space-used in history datasets (HDS) of BMC MAINVIEW (like MVDB2) - and I wonder WHY some of these files are "overused". (Recenct case: Page-IO-Statistics "flush" some of my MVDB2 HDS.)

       

      My idea is: Whats about a tool that tell me WHAT kind of MV(DB2) data fills up my HDS?

       

      Nowadays we have to create a problem ticket (issue) and to transfer (big!) files to BMC. Too much efforts - I think.

        • 1. Re: Mainview History Datasets: How to control and inspect?

          Discussion successfully moved from Products to MainView

          • 2. Re: Mainview History Datasets: How to control and inspect?

            This is an excellent idea!    As a MainView customer you can open a case with us and enter this as an RFE (request for enhancement).    I would encourage you to do that. 

            But in the mean time understanding your DB2, CICS, MQ or IMS activity(workloads) would provide insight into what is being stored in the MainView history datasets.    These datasets store the interval data from all views that have history capability (the TIME command can be used on them).    So if you have a spike in workload/activity these datasets will fill more quickly.     I hope you are archiving these datasets, because they can be backed up and recovered and viewed online.    This is detailed in the MainViewAdministration Guide.    And I am guessing you are familiar    

            My final plug is for MainView Threshold Advisor which is a feature that comes with all MainView products and helps you understand your workloads and can automatically set thresholds that could be used to alert you of anomalies which would impact the history datasets.

            Thanks for your participation in the community and I hope this has been of some help.

            1 of 1 people found this helpful
            • 3. Re: Mainview History Datasets: How to control and inspect?
              Sieghart Seith

              OK - I will do this RFE (I still have a opened issue).

               

              NO - we are NOT familiar with archiving HDS. As I understand there is NO bmc tool for archive job. So we have to build our own automation (trigger is msg BBMEA018I for example), haven't?

              • 4. Re: Mainview History Datasets: How to control and inspect?

                You are correct that automation must be used to set up the archive process.    I have documented this for some of my customers and I will provide that documentation here.

                 

                      The following messages indicate the switch of a history dataset.   You can use AutoOPERATOR or another automation package to pick up the message from the console and trigger a job like the REPRO below to copy the history dataset to an archive dataset.   I have also provided information on how to allocate a previously archived history dataset to MAINVIEW and the process BMC recommends to use for archiving and retrieving history datasets.

                BBMEA018I End of File on Historical Data Set CSGI.MAINVIEW.IMAGSYSB.HISTDS02 (0)  *** this should be the one you trigger on**

                BBMEA002I Historical Data Recording switched to Data Set CSGI.MAINVIEW.IMAGSYSB.HISTDS03(0)

                 

                 

                This is using BMC AutoOperator.

                 

                Using variables IMFOJOB(jobname that generated the message) IMFODATE(date) IMFOTIME(time) you should be able to generate a meaningful output dataset name.

                 

                 

                //JOBNAME JOB (3892),'USERNAME',CLASS=F,MSGCLASS=R,

                // NOTIFY=&SYSUID,

                // MSGLEVEL=(1,1) *** INSTALLATION STANDARDS ***

                //COPY EXEC PGM=IDCAMS,REGION=4M

                //SYSPRINT DD SYSOUT=*

                //HISTDS DD UNIT=disk,DISP=SHR,

                // DSN=vsam file name from BBMEA018I message

                //HISTBACK DD DSN=backup of history,DISP=(,CATLG,DELETE),

                // UNIT=????

                //SYSIN DD *

                REPRO INFILE(HISTDS) -

                OUTFILE(HISTBACK)

                /*

                //

                 

                 

                 

                This information was taken out of the MAINVIEW Administration Guide (Chapter 13 Managing Historical Datasets)

                 

                 

                 

                To allocate a historical data set dynamically so that it is added to the

                historical database and appears in DSLIST, enter the ADD command. The

                ADD command can be entered as a primary command or as a line command.

                When you use the ADD command, you have a choice of permanently or

                temporarily allocating the historical data set. To permanently save a

                dynamically allocated historical data set so that it is retained when the PAS is

                recycled, issue the SAVE command while adding a historical data set (see

                page 13-22 for more information); otherwise, the data set is allocated

                1. temporarily.

                To add a historical data set dynamically:

                Step 1 Create the data set by following the instructions for your type of PAS (MVS

                or BBI-SS) in the MAINVIEW Common Customization Guide.

                Step 2 Display the DSLIST view for the product for which you want to add a

                historical data set by following the instructions on page 13-9.

                Step 3 Enter the ADD command in one of the following ways:

                • On the COMMAND line, type ADD and press Enter.

                • In the line command column, type A and press Enter.

                Step 4 Enter the name of the historical data set you want to allocate in the Dataset

                name field.

                Step 5 Specify whether you want this data set to be eligible for recording by typing

                YES or NO in the Record? field and then pressing Enter.

                The message DATA SET STARTED is displayed.

                Note: Use this field to set recording eligibility of a data set initially; use the

                O line command to control recording eligibility once the data set is

                allocated (see “Changing Recording Eligibility” on page 13-11 for

                more information).

                Step 6 Optionally, add information for additional data sets, if necessary, and press

                Enter for each data set to add it dynamically.

                Step 7 Press END to return to the DSLIST view.

                The message DATA SET ALLOCATED is displayed for each newly allocated

                data set and the data set name is added to the DSLIST view. Dynamic

                allocation is complete but does not make the data set a permanent member of

                the historical database. When the PAS is recycled, the data set name no

                longer appears on DSLIST and must be added again.

                To permanently add a dynamically allocated data set to the historical

                database, issue the SAVE command (see page 13-22 for more information).

                Figure 13-15 Add Historical Data Set Panel

                ------------------------- Add Historical Data Set --------------------------

                COMMAND ===> SCROLL ===> CSR

                Enter dataset name and recording eligibility.

                -Enter a fully qualified dataset name

                -Do not enclose dataset name in quotes

                -Dataset must have already been created

                Dataset name:

                Record? May be used for recording? Yes/No

                END to exit adding historical dataset

                ADD to add historical dataset and continue add dialog

                CANCEL to exit the add dialog

                HELP to view related help

                RESET to reset to initial data

                 

                 

                Archiving and Retrieving Historical Data

                You can use the REPRO or EXPORT function of IBM’s IDCAMS utility or

                the standard utility your site uses to archive historical data to tape or to

                offline data sets. At a later time, you can use the same utility to retrieve the

                data from the archive into a historical data set for viewing.

                Archiving Data

                As soon as a historical data set fills up, MAINVIEW sends a status message

                to the operator console. BMC Software recommends setting up an

                automation product that uses this operator console message to trigger the

                IDCAMS REPRO or EXPORT function to automatically offload historical

                data to the archive each time a data set reaches capacity. For example, you

                can use one of these automation products:

                • MAINVIEW AutoOPERATOR

                • the standard utility that your site uses

                Retrieving Data

                When you restore archived historical data, the data set you move the data to

                must be currently allocated and must not be eligible for recording.

                When you complete the restoration of a historical data set from tape, there

                are either no time and date stamps displayed or invalid date and time stamps

                displayed in the From Date, To Date, and Time fields for the data set.

                The date and time data is not correct because the PAS reads the date and time

                of historical data sets only at data set initialization time or when data sets are

                added dynamically.

                To restore archived historical data and display the correct date and time of a

                restored historical data set on the DSLIST view, BMC Software recommends

                this procedure:

                Step 1 Use the D line command to deallocate the historical data set (see

                page 13-23).

                Step 2 Restore data from the archive using an IDCAMS or other standard utility.

                Archiving and Retrieving Historical Data

                Step 3 Use the ADD command to add the historical data set (see page 13-20) to the

                historical database.

                Step 4 Optionally use the SAVE command (see page 13-22) if you want to

                permanently save the data set allocations so they are not removed at the next

                PAS initialization.

                 

                1 of 1 people found this helpful
                • 5. Re: Mainview History Datasets: How to control and inspect?
                  Sieghart Seith

                  Moreover: the reason for "flooding" our mvdb2 HDS is:

                  PAGESET statistics.

                  These are written for EVERY pagesets (which have io counts) for EVERY interval and let no room (space) for other (actual important) data!

                   

                  So I do a RFE to can SWITCH OFF this type of history recording.

                  • 6. Re: Mainview History Datasets: How to control and inspect?

                    Yes the best way to approach this is via a RFE.     You could make it more generic, by saying control what history records are recorded to the history datasets.   Example is turning off the PAGESET statistics  which flooded my history datasets.   

                    Thank you for your enhancement ideas.