MICS

Expand all | Collapse all

EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condition

  • 1.  EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condition

    Posted 03-28-2014 04:35 PM

    Hello CA SMF DIRECTOR Boarders:

    Looking for others' input here on how an SMF DIRECTOR EXTRACT treats a CAF3103W (NO DATA SELECTED - a "warning" type message) condition as a COND CODE 0004, where there are also other instances where a COND CODE 0004 occurs, mostly when CAF2417W (....AN OVERLAP IN INDEX...) can and does occur occasionally, even though we have specified the max-value of OVERLAP(300).

    Currently, a circumvention is used (at this site) by having a 'dummy' DFSORT job-step that detects an empty file and generates an ABEND -- more accurately raising the "NO DATA" condition.  Unfortunately, this alternative has to be explained.

    From experience with running SMF DIRECTOR EXTRACT production-job requests, the CAF3103_ condition should more likely be treated as an ERROR, not a WARNING, while percolating a higher COND CODE at job-step end.

    Any other sites / admins out there have this same challenge with interpreting a COND CODE 0004, when it is more about being an ERROR not a WARNING (CAF3103E and maybe a COND CODE 0008, instead)?

    Looking for feedback and hopefully ammunition to help sway CA Support to change CAF3103_ to "E" or (less desired) provide a means to suppress the COND CODE 0004 for CAF2417W  conditions which are mostly normal and can/do occur.

    Regards,

    Scott Barry
    SBBWorks, Inc.



  • 2.  RE: EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condi

    Community Manager
    Posted 03-31-2014 10:04 AM

    Does anyone have any feedback/suggestions for Scott?

     

    sbb:

    Hello CA SMF DIRECTOR Boarders:

    Looking for others' input here on how an SMF DIRECTOR EXTRACT treats a CAF3103W (NO DATA SELECTED - a "warning" type message) condition as a COND CODE 0004, where there are also other instances where a COND CODE 0004 occurs, mostly when CAF2417W (....AN OVERLAP IN INDEX...) can and does occur occasionally, even though we have specified the max-value of OVERLAP(300).

    Currently, a circumvention is used (at this site) by having a 'dummy' DFSORT job-step that detects an empty file and generates an ABEND -- more accurately raising the "NO DATA" condition.  Unfortunately, this alternative has to be explained.

    From experience with running SMF DIRECTOR EXTRACT production-job requests, the CAF3103_ condition should more likely be treated as an ERROR, not a WARNING, while percolating a higher COND CODE at job-step end.

    Any other sites / admins out there have this same challenge with interpreting a COND CODE 0004, when it is more about being an ERROR not a WARNING (CAF3103E and maybe a COND CODE 0008, instead)?

    Looking for feedback and hopefully ammunition to help sway CA Support to change CAF3103_ to "E" or (less desired) provide a means to suppress the COND CODE 0004 for CAF2417W  conditions which are mostly normal and can/do occur.

    Regards,

    Scott Barry
    SBBWorks, Inc.



  • 3.  RE: EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condi

    Broadcom Employee
    Posted 03-31-2014 11:10 AM

    Hi Scott,

    The issuing of a return code of 4 when no data has been received on an EXTRACT with SMF Director dates back to the origins of the product.  Were we to change the condition code, there might be issues with other customers' production runs where they are looking for a condition code of 0 or 4 in the EXTRACT to determine whether or not a subsequent step should be run.  If suddenly the condition code were to be changed to 8, their production set up would need to be changed. 

    I have looked over your email that you sent privately along about this and I will be replying to it shortly with a suggestion we may implement in the not-too-distant future. It is something that will provide greater flexibility for the no-record condition.

    Mike

     

     



  • 4.  RE: EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condi

    Posted 03-31-2014 01:50 PM

    Hoping here that CA will look forward (CA SMF DIRECTOR of today/tomorrow) instead of backward (legacy CA SMF JARS product) - that to mean, consider that maybe the "NO DATA" condition in production-mode never should have been COND CODE 0004 (a warning, rather than an error).

    And if there are SMF DIRECTOR clients out there on this channel, consider this is an opportunity to share experiences if there are SMFD EXTRACT jobs that cannot check for GE COND CODE 0004 because it's too broad a abnormal/warning category.

    Okay, so maybe your next job-step fails when it doesn't have data to process -- does that make sense?  Doesn't to me.

    As a circumvention, we forced to use DFSORT COPY to detect an empty file and then generate a hard-error -- doesn't happen often, but it can happen.

    And so one might say well there are possible condition where NO DATA is acceptable -- then the approach to take is add SMF type 23 to you SELECT(....) parameter - then you will get a COND CODE 0000 and your subsequent SMF data post-processing job-step will properly detect no-data and act accordingly.

    Scott Barry
    SBBWorks, Inc.



  • 5.  RE: EXTRACT, NO DATA, and COND CODE 0004 warning rather than an error condi

    Posted 04-30-2014 02:54 PM

    Hello again -

    An update on this situation with CA SMF DIRECTOR and EXTRACT behavior...

    CA has agreed to change the COND CODE from 0004 to 0008 (higher severity, not really a WARNING) for message "CAF3103W  NO SMF RECORDS WERE EXTRACTED" condition, at some future release-level (hopefully the next one and within a reasonable time-period!)  I expect the message will change from "W" to "E" as well for effect.

    Fortunately, in the meantime, CA has agreed to provide interim PTF-relief by increasing the maximum-value allowed for the OVERLAP parameter (to 86400) -- basically eliminating any COND CODE 0004 that occurs today with message CAF2417W. 

    Obviously this is a circumvention to the problem where CAF3103 should never have been a subtle warning anyway.  Fortunately, CA listened when the evidence was demonstrated clearly....and for that we say thank you!

    Regards,

    Scott Barry
    SBBWorks, Inc.

    P. S.  With the upcoming CA COMMUNITIES overhaul/migration to JIVE platform, we have been soliciting to get CA SMF DIRECTOR raised up out of  the "CA MICS GLOBAL USER COMMUNITY" since it is not MICS-related nor is it anymore a CA JARS component.....  Hopefully that change will happen sooner rather than later.   Bu-Bye for now.  SBB