View previous topic :: View next topic
|
Author |
Message |
Manu SK
New User
Joined: 20 Mar 2020 Posts: 3 Location: India
|
|
|
|
Hi All,
We have a number of workload automation solutions available, on Mainframe platform, from various vendors. I would like to understand how these products/solutions get notified about various events like start of job, end of job, start/end of a job step, new dataset cataloged etc.
Thanks |
|
Back to top |
|
|
enrico-sorichetti
Superior Member
Joined: 14 Mar 2007 Posts: 10879 Location: italy
|
|
|
|
quoting the IBM manual
Quote: |
IBM Tivoli Workload Scheduler for z/OS tracks the progress of jobs and started
tasks through the z/OS system by using JES and SMF exit points. Add all these
exits on each z/OS system where you will start IBM Tivoli Workload Scheduler for z/OS |
the other products behave in the same way |
|
Back to top |
|
|
Apoorva
New User
Joined: 28 Jan 2020 Posts: 49 Location: India
|
|
|
|
Manu SK wrote: |
Hi All,
We have a number of workload automation solutions available, on Mainframe platform, from various vendors. I would like to understand how these products/solutions get notified about various events like start of job, end of job, start/end of a job step, new dataset cataloged etc.
Thanks |
Yes as @enrico-sorichetti told us, vendors normally use exits (mainly SMF's IEFU83/IEFU84) to get notified about various events occurring in the system. They might also use ENF exits (Event notification facility), JES exits, Sub-system interfaces (SSI interfaces) depending on events they want to be notified to.
Though exits are widely used, sometimes they might modify IBM modules/control blocks as well in order to intercept certain events for which IBM hasn't provided an exit. For example if a Scheduler has to trigger a job when a new record with particular key gets inserted into a DB2 table (This is just an example, and I am taking DB2 example here because many people are familiar with DB2) then vendors may have to modify DB2 modules/programs in order to accomplish this as DB2 doesn't provide an exit for this. Moreover it's impossible to provide an exit for each and every event occurring in the system. Modifying IBM modules to intercept certain functions is normally called hooking/front-ending/ZAPing. Though this is not the preferred approach, vendors still do it if they don't find any exits.
Another example from CICS perspective is if a batch job needs to be triggered if a transaction, fired by a particular user ends with a particular RC. CICS may not have an exit (as far as I know) for this.
So, in general schedulers use exits if one available otherwise they try to hook IBM code (least preferred approach). |
|
Back to top |
|
|
vasanthz
Global Moderator
Joined: 28 Aug 2007 Posts: 1744 Location: Tirupur, India
|
|
|
|
Very informative, Thanks |
|
Back to top |
|
|
Joerg.Findeisen
Senior Member
Joined: 15 Aug 2015 Posts: 1306 Location: Bamberg, Germany
|
|
Back to top |
|
|
Manu SK
New User
Joined: 20 Mar 2020 Posts: 3 Location: India
|
|
|
|
Thanks everyone for all the information! Really appreciate your time and effort on this. One additional question if you don't mind please. When you say "Vendors may change IBM/DB2/CICS modules if they don't find any exits" does that mean non-IBM Mainframe Software vendors will have access to IBM/DB2/CICS source code? |
|
Back to top |
|
|
Apoorva
New User
Joined: 28 Jan 2020 Posts: 49 Location: India
|
|
|
|
Manu SK wrote: |
Thanks everyone for all the information! Really appreciate your time and effort on this. One additional question if you don't mind please. When you say "Vendors may change IBM/DB2/CICS modules if they don't find any exits" does that mean non-IBM Mainframe Software vendors will have access to IBM/DB2/CICS source code? |
I just know that non-IBM software vendors buy special logic manuals from IBM (that's not available on public domain) under non-disclosure agreement. These logic manuals may include module/program flow of specific functions that vendors are interested in but I don't think so they will also include source code. Nevertheless non-IBM vendors do lot of reverse engineering of DB2/IMS/JES/CICS/SMS etc modules/programs to accomplish what they want to do. The tools that they normally use for such reverse engineering are SLIPs, GTF traces (Generalized trace facilities), PER SLIPs (PER ==> Program event recording) etc. Please Google if you want more info on this. |
|
Back to top |
|
|
Manu SK
New User
Joined: 20 Mar 2020 Posts: 3 Location: India
|
|
|
|
Apoorva wrote: |
Manu SK wrote: |
Thanks everyone for all the information! Really appreciate your time and effort on this. One additional question if you don't mind please. When you say "Vendors may change IBM/DB2/CICS modules if they don't find any exits" does that mean non-IBM Mainframe Software vendors will have access to IBM/DB2/CICS source code? |
I just know that non-IBM software vendors buy special logic manuals from IBM (that's not available on public domain) under non-disclosure agreement. These logic manuals may include module/program flow of specific functions that vendors are interested in but I don't think so they will also include source code. Nevertheless non-IBM vendors do lot of reverse engineering of DB2/IMS/JES/CICS/SMS etc modules/programs to accomplish what they want to do. The tools that they normally use for such reverse engineering are SLIPs, GTF traces (Generalized trace facilities), PER SLIPs (PER ==> Program event recording) etc. Please Google if you want more info on this. |
Thanks for the explanation. |
|
Back to top |
|
|
|