It is with the greatest pleasure that I am able to announce that Novell has just posted the documentation for setting up OpenAIS, Pacemaker, OCFS2, cLVM2, DRBD, based on SUSE Linux Enterprise High-Availability 11 - but equally applicable to other users of this software stack.

We understand it is a work in progress, and the uptodate docbook sources will be made available under the LGPL too in the very near future in a mercurial repositoy, and we hope to turn this into a community project as well, providing the most complete documentation coverage for clustering on Linux one day!

• So our new test cluster environment is a 16 node HP blade center, which pleases me quite a bit. The blades all have a hardware watchdog card, which of course makes perfect sense for a cluster to use.
• However, the attempt to set the timeout to 5s was thwarted by the kernel message

  hpwdt: New value passed in is invalid: 5 seconds.

• So in I dived into hpwdt.c, to find:
  static int hpwdt_change_timer(int new_margin)
{
/* Arbitrary, can't find the card's limits */
if (new_margin < 30 || new_margin > 600) {
printk(KERN_WARNING "hpwdt: New value passed in is invalid: %d seconds.\n", new_margin);
return -EINVAL;
}

• Okay, that can happen. Sometimes driver writes have to make guesses when the vendor is not cooperative or unavailable. So who wrote the driver?
  * (c) Copyright 2007 Hewlett-Packard Development Company, L.P.
• ...

I prefer to ignore christmas and the madness they call holidays, but would like to close the year with a series of three questions, starting today:

1. What can Open Source (and/or Linux) contribute to making the world a better place? Think of developing nations and the real large issues, as well as the slightly smaller ones.

Please feel free to e-mail me your answers to lmb at suse dot de, but this is not required to follow this experiment.

• An article by heise open covers the Linux Kongress, and also my presentation on convergence of cluster stacks, even though they represent my message it slightly more tentative than I intended it to be. But maybe I am too optimistic. For what it is worth, here is a picture of the slide where I outlined the components in the joint stack, which heise open calls a "good mix from all sources."
• It is possibly quite important that that is my understanding of the results and goals, and even though I believe we had good buy-in in the development community, this should not be understood as a promise or commitment (or lack thereof) by Red Hat or Novell or anyone else to deliver this in the Enterprise distributions in particular, nor that there will be any loss of support for current configurations. If I could speak for both Red Hat and Novell, I would be earning a hell of a lot more money. (Some initial feedback to my blog entry here made me add this paragraph; I did discuss this in the presentation, but it is not captured on the slide shown.)

• Lukas Chaplin of Linux-Lancers.com, a Linux recruiting and placement agency, has interviewed me about working from a home office. This is not yet as pervasive elsewhere as in the Open Source environment, which is really a shame.
• Of course, before going to Lukas you should first check whether Novell & SuSE can offer you a new challenge!

• It's been a while since I blogged, so I have two conference reports as well, starting with the Cluster Developer Summit in Prague, 2008-09-28 - 2008-10-02. (See the link for Fabio's report.)

  This Summit was organized by Fabio from Red Hat and hosted by Novell, with attendees from Oracle, Atix, NTT Japan and others, which Lon captured on this picture. It is my honest belief that within a year or two, we shall have one single cluster stack on Linux; totally awesome! Amazing how much progress one can make if one is not stuck to one's own old code, but willing to select the best-of-breed.

  I think we have come a long way in the last ten years; having explored several different paths through concurrent evolution, we are now seeing more and more convergence as there is less and less justification for the redundant effort expended. Dogs, cats, and mice eating together ... It also reinforced my opinion that small, focused developer events can be exceptionally productive.

• At Linux Kongress 2008 in beautiful Hamburg, there were many tutorials and sessions where Pacemaker + heartbeat were used to build high-availability clusters. In my own session, I presented the last year or so of development on Pacemaker and heartbeat, and of course summarized the results from the Cluster Developer Summit.

  I also learned about a neat trick Samba's CTDB plays with TCP to make fail-over faster; of course, thanks to this being Open Source, they were able to contribute it to the community instead of reinventing their own cluster stack. (Haha, just kidding, of course they rolled their own - this is Open Source after all.) However, it should be possible to copy it and integrate it as a generic function for IP address fail-over. Cool stuff.

  I also very much enjoyed dinner with James, Jonathan, Andreas, Lars (Ellenberg), and Kay - who lives in Hamburg, but whom I only see at conferences ... Refer to the working from home offices interview!