Skip to Content

At their booth at TechEd 07 in Las Vegas, IBM demoed updated transactions for Basis and other SAP experts to view the underlying  hardware performance of a virtualized environment, as well as the totally cool transfer of a live running SAP system and database from one node to another, without user session interruption.  The system was running in Walldorf, and the control/display was in Nevada, USA!

ST06/OS06 is the local  operating system monitor, where you view the CPU, memory and other metrics  collected by an SAP software agent (saposcol) on an immediate or short history  period.  One of my mandates in visiting TechEd this year was to get ideas for better  solutions to the collection, storage and use of this data in both long term  capacity planning scenarios as well as immediate escalation issues.   See my ST10 blogfor an example of where we need this capability.

The problem with the  “old” ST06 is that it uses granular system resources that don’t exist any longer  in a virtualized environment, so if I see 90% CPU usage that might not be a bad  thing.

The screens we saw were for SAP 7.0,  so my next challenge will be how to connect our older systems via agents to a  newer system (maybe Solution Manager 4.0) to provide this information  before we upgrade. 

SDN forum:

I found one SDN forum Virtualization on AIX 5.3 and above with SAP on the topic of OS metrics in virtualized environments. 

Miscellany: 

OS07N ST06N OS06N, all lead to the same target.
Details are described in the SAP note 994025.
Agents to monitor older SAP versions are available down to 4.6 D kernel based systems.
The central display needs to run on an 7.0 based system [correction: central display can run on systems 6.20, 6.40 or 7.0 based, as described in Note
994025]

Notes of interest:

  • 994025 Improvements in monitoring virtualized OS environments
  • 1013331 CCMS agents and kernel: Patches 2007
  • 1019567 Corrections for new operating system monitor
  • 1084019 OS07N: New operating system monitor

And, aPDF document entitled “Operating System Monitor” – release 716.

 

Screen shots and research into our systems:

The newest SAP installation we are running is Solution Manager 4.0, so I used this to try to find the transactions as shown.  Alas, while we have a 7.00 system, we are not up to the patch levels required in the Basis, kernel and agent versions.  We are on PL 10, and need to be on PL26!?

However, there are notable differences from our 6.20/6.40 systems.  The most obvious is the change from flat screen displays to ALV grids.  Less useful is that we are missing some of the column titles, and one appears to be in German (“stunde”).  Time to patch!

620/640 – Old 700 / New
image image
image image
image image

 


Screen shots pulled from TechEd partner session LCM243:

 

 

image


image

 


 

Screen shot from PDF file – note that this lists an HP-UX system, showing the improvements are not only for IBM AIX.

image

To report this post you need to login first.

3 Comments

You must be Logged on to comment or reply to a post.

  1. Nikolaj Lyngfeldt
    Hi there.
    I agree that these is some nice improvements, however, most of this information is already available with various utilities/commands, like lparstat, topas, nmon and so on. So I’m not really excited. I think what is really needed is the overall CEC view and integration between SAP workload stats and the OS stats, per CEC and LPAR. I am trying to create this view by grouping the systems and customizing some alert monitor sets, but I am not yet sure if this will be the right solution – E.G. no total CEC values. I can do this with third party tools, but I really want to see it in CCMS. Has anybody done that?

    Cheers, Nikolaj

    (0) 
    1. In CCMS or these new transactions, do we have the option to look at the status of file level? How busy it is, etc.  What we have is file system and logical volume and hard disk drive level.

      Regards,
      Dwight

      (0) 
  2. Irene Hopf
    Hi Nikolaj,

    as you mention this data is available in various utilities/commands.
    But this is exactly the problem. By using these tools you did not get this information centrally and you did not get it inside the SAP system. The new monitor functionality in OS07n now provides all this data centrally and additionally provides the big benefit to store the history of the virtualization data in the central performance history of the SAP system.
    This allows SAP to provide their services like Early Watch in a standard way which has been a problem before (e.g. creating red traffic lights based on wrong data).

    Let me make some remarks regarding the CEC view, a topic that was mentioned also by several customers:

    – I agree that a complete CEC view is missing that shows on a single screen the relative load produced by each LPAR running on the server.

    – Partially you get the information about the CEC already today. OS07n already shows some data of the CEC, e.g. the total number of processors in the shared pool, the total available capacity and some more.

    – In addition there is another aspect of a CEC view – getting information about the CEC from inside an LPAR is always critical. This is also the reason why it is possible to grant or refuse this inside out view to an LPAR. This is handled on HMC level. You can see this in OS07n by the indicator “Pool utilization authority”.

    You can check out some good screenshots also in the SAP Info online article that was published earlier this year – http://www.sap.info/int/go/34932/

    I hope SAP will further improve the monitoring for virtualized systems in the future and will find also a good solution for the overall view on the complete physical server.

    Regards, Irene.

    (0) 

Leave a Reply