Skip to Content

During the past months HANA made a fairly big wave in the SAP community. Looked like everybody wanted to get the hands onto this and gather some experience with the new piece of technology. With that tons of questions arised about how to operate such a system.

The recently released note #1599888 – SAP HANA: Operational Concept answers these questions and clearly describes what’s supported and what’s not.

I’m pretty sure that this will make life a lot easier for everybody involved in HANA projects.

A big THANK YOU from my side to the authors of this note!

To report this post you need to login first.

15 Comments

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

  1. Stefan Koehler
    Oh Lars my friend – you are walking in the wrong direction ;-))

    Thanks for sharing that note / PDF with us. Well that paper confirmed all my main concerns about HANA as an operational database for BW or ECC itself.

    >> 7.2 Recovery
    No Point-in-Time Recovery? It must be a joke to do aggressive advertising for HANA as SAP BW or ECC database, but no option to do point-in-time recovery. Just think about user errors (deleting data accidentally or something like that)

    > 8 High Availability
    Just pointing to hardware partners is measly – the software itself need HA possibility in my opinion.

    Lars – come back to the good old world :-))

    Best Regards
    Stefan

    (0) 
    1. Mark FΓΆrster
      Hello Stefan,
      somehow I agree that a decent database needs a point in time recovery option.

      However, the classical point-in-time recovery is getting more and more a theoretical option. In practice, these monsters are avoided at almost all costs. If someone has even a slight chance not to do a point-in-time recovery for its production system, then this is the way to go. Period.

      In case of user errors it is useful to create a copy of production e.g. via a point-in-time recovery and then fix the user errors manually.

      Regards,
      Mark

      (0) 
      1. Stefan Koehler
        Hello Mark,
        i fully agree with you. Mostly it is tried to avoid a classical point-in-time recovery, but even if you try to reconstruct the data by a copy or some other technologies like flashback – the “core component / technology” is point-in-time recovery.

        … and even the HANA database has no software HA / disaster recovery features – i really have some doubts about running that in a critical environment as “universal database” for SAP BW/ ECC.

        By the way we use “classic point-in-time recovery” regularly by performing system copies (productive to quality). All external satellite systems need to be in consistent state with the SAP itself.

        Regards
        Stefan

        (0) 
      2. Lars Breddemann Post author
        Hi Mark,

        you’re so right!
        PiT-Recovery is effectively a checkpoint feature in distributed system landscapes.
        For user-errors I would expect to see a stronger use of ‘time-travel-data-access’ in the future: deleted or changed data is then simply selected as of the point in time before the change happened.
        Oracle can do stuff like this and HANA has the foundations for such features as well.

        Anyhow, there surely is a huge room to improve HANA DBMS-features and development is working to fill this room.

        Best regards,
        Lars

        (0) 
    2. Lars Breddemann Post author
      Hi Stefan,

      guess there’s no way ‘back’ for me or for HANA and that’s a good thing. πŸ™‚

      What’s lacking in the HANA DB is obvious to everybody involved and I know that development is working on these points.

      So, right now, for the current state, HANA is the side-show appliance that comes in fixed dimensions.
      It surely will become more interesting and more difficult to support a broader usage of HANA technology. But for the time being I got the impression that more ‘guidance’ better helps with taking the first steps in In-Memory computing.

      Best regards,
      Lars

      (0) 
      1. Stefan Koehler
        Hey Lars,
        well i am not able to drive you into your luck :-))

        >> So, right now, for the current state, HANA is the side-show appliance that comes in fixed dimensions.

        Oh well i got another statement about that: Are we putting the cart before the horse?

        SAP HANA will be in ramp-up for SAP BW in november this year – so SAP knows that HANA DB is not ready for usage in real environments (because of lack of features), but the customers can get it in ramp-up? I am scared :))

        Regards
        Stefan

        (0) 
        1. Lars Breddemann Post author
          Hi Stefan,

          just back to the office and already another comment from you πŸ™‚

          From my point of view (which is NOT the official SAP position!) it’s like that:
          HANA will be available for usage with BW for ramp-up customers.
          Such ramp-ups are in _very_ close contact to development during the ramp-up phase.
          Everybody involved knows exactly the limitations of the product during this period.
          Plus, the BW data does not necessarily need to be recoverable via point-in-time recovery.

          To me it’s not like we’re pretending to deliver features that aren’t there.
          Instead we make the features available to early-bird customers quite early but keep on developing the needed features for general usage pattern meanwhile.

          cheers,
          Lars

          (0) 
          1. Stefan Koehler
            Hell Lars,
            for sure – as you know i am interested in the whole database market so i am looking beyond one’s own nose to keep up-to-date – even when it is called MaxDB or HANA ;-))

            The SAP marketing beast is running out there and hit us too. My managers were at different marketing presentations (with/without hardware vendors) and they were thrilled about HANA. I am quite sure that they don’t know about these limits right now and even not that there is no way back to a “normal disk-based RDBMS” (afaik).

            Once you got your system running on HANA you are catched by the claw of the SAP beast πŸ˜‰

            Regards
            Stefan

            (0) 
            1. Lars Breddemann Post author
              Always good to keep an eye on what’s coming up next – no question about that!

              For HANA, well, yes, there are currently special features developed exclusively for HANA, like the flat Infocube structure.
              This likely won’t work very well on classic DBMS.
              But will you be locked in? No, since SAP keeps up the heterogenous system copy feature.

              Other HANA-specific stuff (calcviews/analytic views) – sure, you won’t be able to move that to another DBMS. But that’s no different to e.g. how partitioning works with differen DBMS or how STAR queries need to be optimized, etc.

              And (yet) nobody is selling OLTP on HANA to customers. So, there’s a lot of way to go – and I’m quite confident that our developers at least know about the rough direction πŸ˜‰

              Best regards,
              Lars

              (0) 
  2. Martin Maruskin
    Hi Lars,

    by any chance have you contributed to this doc (SAP HANA Operational Concept)?
    I’m just missing some explanations of abbreviations used the doc, e.g. “LT system (page no.3)”. For us who trying to to study what is HANA all about would be appreciated…
    Some typo errors are there as well: 171356 – SAP software on Linux not on “Linus” πŸ™‚ (page no.5)

    Thanks, Martin

    (0) 
    1. Lars Breddemann Post author
      Hi Martin!

      No, as I’m not part of our development team, I don’t write actual documentation or code or something like that.
      I simply write _about_ the stuff my colleagues create and how I cope with it πŸ™‚

      The abbreviation “LT system” refers to the origin of the replication feature – the “Landscape Transformation Service”. There the replication was used to reduce downtimes during upgrade operations to a minimum.

      Concerning the typo, I’ll forward your comment t the colleague who owns this document.

      Thanks for lending your ‘eagle eyes’ and pointing out!

      Cheers, Lars

      (0) 

Leave a Reply