Skip to content

Zasady Projektowania Środowisk PowerHA

Mój prawie sąsiad, Rob Y. określił jedną z moich poprzednich prezentacji HACMP/PowerHA zatytułowaną “PowerHA Concepts, etc” jako “przestarzałą”. Ta, którą przekazuję tym razem jest chyba najnowsza (zawiera materiał PoserHA 6.1 oraz 7.1) – Best Practices for Designing a PowerHA Enterprise Edition Solution on AIX. Zobaczyła światło dzienne pierwszy raz w czasie ostatniej konferencji AIX w Las Vegas, NV.

Posted in AIX, HACMP, Real life AIX.

Tagged with , , , , , , , .

7 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Andrew Wojnarek says

    You know I think I am the only person in the world that does not like HACMP. I believe the clustering should either be done at the Hardware level or the Application level.

    I have not had good experience with OS high availability. Even with HACMP you have downtime when it has to vary on the VG etc.

  2. MarkD:-) says

    I think there are more people that question HA.

    Over the ten years period of time I did see HA do what it was designed to do only four times.
    In all other times it was a hindrance instead of help, unfortunately.
    Even with it working and doing what it should, often the application protected by HA itself has to be paused,
    checked, verified (all transactions have been applied, nothing is missing, etc) so the primary benefit of
    HA versus let’s say partition mobility – some say the second requires manual intervention….. Then there are the issues with two nodes clusters, the list continues.
    This is good topic for a Friday evening in a bar over a brew or two.

    Have a good weekend,


  3. Anto says

    Hi i m interested to study AIX concepts and functions.. can u send basic HACMP concepts and Study materials?

  4. MarkD:-) says


    You can find more information, directly from IBM free bookstore “”


  5. Rob says

    Mark – a dog ate my homework. I had a comment , gone.
    Anyhow, I’ll be shorter. That is a good presentation. With sysmirror you can notify
    instead of fallover a resource group. Take manual intervention. There are probably
    some SLAs that keep HACMP going. That gateway application. Business wants a one
    hour unplanned outtage… dreads a 3 or 4 hour recovery. What do you design
    to meet a 1 hour outtage? I miss active-active-active-active, etc. clusters. Nothing lasts forever.


  6. MarkD:-) says


    I share your “pain”. To tell you the truth gpfs and load balancers look better and better (if only all applications were able to operate this way……). Often it is difficult to meet the 60 minutes outage because to move a RG + Start/Verify Application requires more then that. Overall, Hacmp/PowerHA seems to be adequate ten years ago. It needs a new look/approach to stay useful today – an opportunity is on the horizon.
    On another front, any news on the subject of time synchronization?

    All the best to you Rob,


  7. Rob says

    It is so funny you should ask about time. In the context of RSA hacked, I dribbled out
    this little emali to co-workers on Saturday:

    Back in the day, a few of us geeked out with s-key. The tcpip stack we used
    gave it away for free. No one was going to hack into my NTP servers!
    (Wanted to put it on production boxes, but co-worker nixed it).

    I simply generated about 300 keys, popped them in a spreadsheet that I would
    bring up on my phone (I had it elsewhere – just in case).

    More here:

    If interested, checkout opie for your ubuntu.

    login as: ry_skey
    Using keyboard-interactive authentication.
    Using keyboard-interactive authentication.
    otp-md5 491 ub2626 ext, Response:
    Linux ubuntu 2.6.35-28-generic #49-Ubuntu SMP Tue Mar 1 14:40:58 UTC 2011 i686 GNU/Linux
    Ubuntu 10.10

    Welcome to Ubuntu!


Some HTML is OK

or, reply to this post via trackback.

Copyright © 2016 Waldemar Mark Duszyk. All Rights Reserved. Created by Blog Copyright.