[cgl_discussion] Re: [dcl_discussion] ANNOUNCE: OSDL Clusters (foundational components)

Lars Marowsky-Bree lmb at suse.de
Wed Nov 26 14:22:18 PST 2003


On 2003-11-26T13:49:24,
   Mika Kukkonen <mika at osdl.org> said:

> Having said that, I do agree that rei-inventing the wheel is not to be
> done lightly. Rest assured, that our decision has been done after lot
> of pain and deliberation of what is currently available in this space.
> None of the existing _implementations_ implement the SAForum API's which
> these projects are aiming for, nor are they at the level several of our
> sponsors hope this kind of services should be.

I must have missed this evaluation and apologize. I would be very
interested in learning more about it; could you please point me to it?
This would certainly be an interesting comparison and help me
immensely. How are Linux HA, OpenSSI, Trans-Is etc unfixable
insufficient?

I think that SAF AIS (even though I have some minor issues with their
APIs, but when is an engineer ever happy ;) is a great step forward
overall, if the process behind it wasn't so closed. All my efforts to
get involved have not been very successful, we didn't manage to get SAF
more interested.

> In case of cluster communication (aka TIPC), we are definititely not
> starting from scratch. It has been the opinion here, that building the
> other two services on top of TIPC should be doable with reasonable work
> amount.

I somehow managed to miss that too. I really should subscribe to the
OSDL cluster list, it appears.

> > There are also reasonably strong reasons arguing against implementing
> > these in kernel space. I'm sure you have heard the summary of them:
> > Maintenance, mainline / vendor acceptance etc.
> Through our initiatives we have heard opinion of people who are planning
> to produce products using Linux clustering technology (i.e. people who
> are potential customers to the company you are working for ...), and
> their opinion differs from yours.

They are also free to have that opinion ;-)

Vendors have a reasonably strict "Mainline first" policy for 2.6. How's
the chances of getting it into the kernel proper there? I'm not asking
tongue-in-cheek, I'm truely curious, because so far our assessment has
been that it seemed to work well enough outside the kernel and was also
much simpler to maintain.

> > Could you please clarify the rationale?
> As SUSE is a member of OSDL and it's initiatives, may I suggest that if
> you want to hear rationale or even influence our decisions you start
> participating in those?

I was just trying to participate by asking that question. I do sincerely
apologize if it came across more than a little cross, but please let me
explain the reason for this.

I can only speak for SUSE, as well as the OCF project (slow as it may
be) and Linux HA, but I've not seen even a brief and short discussion
about this out in the open; the first I've read was the announcement
here, which did create the impression for me that OSDL is not actively
contributing to the Open Source community, but rolling their own. Yes,
reinventing wheels if one of the project fails is one of the strengths,
but as you say, only after some consideration of the other options; I
cannot say that I believe that one can have considered other options
without having discussed them with the respective projects.

I understand and actively appreciate the idea of pushing the community
towards some common goal. Trust me that I'm actively hurt quite a bit by
all those different projects, and in particular the infrastructure
levels like membership et al hurt the most. But, it would be good if it
looked and felt somewhat more like the community was involved, and not a
closed OSDL project ;-)

> > > the network.  Applications and other cluster services (checkpointing,
> > > resource managegers, etc.) would simply subscribe to membership events.
> > Implementation for data checkpointing and some others actually already
> > exist for Linux heartbeat, and all in user-space.
> And we are not planning to implement these.

Well, then, Intel has just contributed a patch for heartbeat to provide
the SAF AIS membership API. ;-)



Sincerely,
    Lars Marowsky-Brée <lmb at suse.de>

-- 
High Availability & Clustering	      \ ever tried. ever failed. no matter.
SUSE Labs			      | try again. fail again. fail better.
Research & Development, SUSE LINUX AG \ 	-- Samuel Beckett




More information about the cgl_discussion mailing list