SMFACCT classes 2 & 3

[Former Member]

Hello

according to recommendations the accounting trace classes 1,2 and 3 should be started but 2&3 can produce significant CPU overhead. We had one case when in consolidation systems there was heavy traffic and we stopped the traces. After a short time the traces were started again - I suppose as a consequence of RSCOLL00 program running every hour. How can we circumvent this , has anyone an idea ? I am asking as z/OS DBA with little basis knowledge. Do you start these traces in development or consolidation systems at all ?

Thank you

Martin Ketterer

[KayKa · Accepted solution]

Hi Martin,

sorry, i don't know exactly how to prevent the restart of this traces.

In production our traces looks like this:

|TNO TYPE   CLASS        DEST QUAL IFCID                          |
|01  STAT   01,03,04,05, SMF  NO                                  |
|01         06                                                    |
|02  ACCTG  01,02,03     SMF  NO                                  |
|03  STAT   31           OP1  NO   156,172,196,225,258,313,330,337|
|04  PERFM  30           SMF  NO   318                            |
|05  STAT   08           SMF  NO   199                            |

In development there are some additional traces

|TNO TYPE   CLASS        DEST QUAL IFCID                              |
|01  STAT   01,03,04,05, SMF  NO                                      |
|01         06                                                        |
|02  ACCTG  01,02,03     SMF  NO                                      |
|03  PERFM  30           OP1  NO   003,006,007,008,009,010,020,023,024|
|03                                025,031,044,045,055,068,069,070,071|
|03                                072,073,074,075,083,084,085,086,087|
|03                                088,089,095,096,105,106,107,112,113|
|03                                172,196,213,214,215,216,226,227,239|
|04  AUDIT  30           OP1  NO   140,141,142,143,144,145,169        |
|05  ACCTG  02,03,05,07, OP1  NO                                      |
|05         08                                                        |
|06  STAT   31           OP2  NO   156,172,196,197,225,258,313,330,337|
|07  STAT   08           SMF  NO   199                                |
|08  PERFM  30           SMF  NO   318                                |

Gruß

Kay

[Former Member · Accepted solution]

Hello Kay

thank you for your answer. The output of the display command looks familiar to me - well I will wait and see if someone has an idea how to prevent SAP from starting traces that were stopped.

Best regards

Martin

[brian_walker · Accepted solution]

We've seen the same behavior in our production systems as well. We were told during a going-live check to turn off some of the SMF traces in production (which we did) but then an hour later they were running again. As we are moving away from the zSeries platform, we gave up worrying about it.

I would suggest opening an OSS message and then sharing the results as this appears to be either a common problem or incorrect documentation about which traces should be enabled.

Brian

[hans_baessler · Accepted solution]

the traces are started and checked by an hourly CCMS job

To turn off the automatic start of DB traces, proceed as follows:

1. Call transaction DBACOCKPIT

2. Choose Configuration SAP Collector Settings

3. Enter TOFF in the (Enter) command field of the DBA Cockpit

Is is described in the SAP DBA Guide chapter 6.3.10 Automatic Start of DB Traces

http://service.sap.com/~sapidb/011000358700000375092007E

cheers, Hans

[brian_walker · Accepted solution]

Thanks for that information! It is good to see the admin guides have been updated with screenshots and information for the later releases of the DBACOCKPIT.

Brian

[KayKa · Accepted solution]

Hi,

sometimes in the past i used the TON to let start the the IFCID(318) automatically in some systems where the trace wasn't running after restart.

After your posting yesterday i tried the TOFF.

That's my result:

28.09.2010 15:34 TOFF

28.09.2010 15:34
-----------------------------------------------------------------------
|TNO TYPE   CLASS        DEST QUAL IFCID                              |
|01  STAT   01,03,04,05, SMF  NO                                      |
|01         06                                                        |
|02  ACCTG  01,02,03     SMF  NO                                      |
|03  STAT   31           OP1  NO   156,172,196,197,225,258,313,330,337|
|04  PERFM  30           SMF  NO   318                                |
|05  STAT   08           SMF  NO   199                                |
-----------------------------------------------------------------------

28.09.2010 15:41 System restart

28.09.2010 15:43
-------------------------------------------------------------
|TNO TYPE   CLASS        DEST QUAL IFCID                    |
|01  STAT   01,03,04,05, SMF  NO                            |
|01         06                                              |
|02  ACCTG  01,02,03     SMF  NO                            |
-------------------------------------------------------------

29.09.2010 07:44
-----------------------------------------------------------------------
|TNO TYPE   CLASS        DEST QUAL IFCID                              |
|01  STAT   01,03,04,05, SMF  NO                                      |
|01         06                                                        |
|02  ACCTG  01,02,03     SMF  NO                                      |
|03  STAT   31           OP1  NO   156,172,196,197,225,258,313,330,337|
|04  PERFM  30           SMF  NO   318                                |
-----------------------------------------------------------------------

So i think there must be little bit more that starts our traces.

regards

Kay

[Former Member · Accepted solution]

Hello Hans

it is very very late for my answer - sorry but we had lots of other things to do and yes : your answer was correct. We have now succeeded in switching off these accounting traces exactly as you mentioned it. I don't know why there was a situation when it did not work.

Thanks a lot and best regards

Martin Ketterer