Is there interest in a talk on clustering?

William Muriithi william.muriithi-Re5JQEeQqe8AvxtiuMwx3w at public.gmane.org
Thu Dec 17 16:28:14 UTC 2009


Morning Madi

2009/12/17 Madison Kelly <linux-5ZoueyuiTZhBDgjK7y7TUQ at public.gmane.org>:
> Before you answer, I want to mention a concern...
>
> I've been doing a fair bit on clustering lately, and have successfully
> gotten two-node clusters working designed to run virtual machines that can
> be moved between nodes on the fly. I think I could see this being of
> potential interest to TLUG as a talk.

I think there would be interest. I for one would be very happy to hear about
it

> However,
>
> It's a sufficiently involved topic that, given the two hour window of a
> talk, I don't think I'd be able to cover everything needed to actually
build
> a cluster. For this reason, I've been debating offering it up as a talk.
> I've been thinking about ways to make a talk useful, and am still not sure
> how I could do that... So the second part of this question is; Given a
> limited time frame, what would you guys like to hear about with regard to
> clustering? What would be helpful?

This is definitely not a simple talk. I think there is no way you can
present it in details all across, especially as it involve lots of
technologies.
I mean, it involve virtualization, networking (VRRP), LVM or cluster file
system, DBRD or SAN storage just to name a few. The only way you can deliver
it in time is to generalize some of the content.

If I was in your shoe, I would skim over simple stuff, things that tend to
work easily and spend most of the time were really hard to get going. When I
mean skim, I mean mention it just enough that one can tell how its related
to the main talk, alternatives and their advantages/disadvantages etc.

The meat of the talk I assume would be DBRD and HA. For these two services,
you can go deep as explaining every section of the configuration file, so
the lion share of time should go toward these two in my humble opinion And
to be frank, they also tend to be the hardest to get going, all the more
justifying the extra effort needed there. Ok, I admit cluster file system is
also a pain, but one has to make a choice.

You can also consider sending out the presentation before the talk - people
can look at it and be up to speed when you are presenting. Or else share
your notes on the project at the end of presentation and everyone can look
at the detailed stuff  later and fill up the gaps. In short, the listerners
also have to do a bit of homework as the time will not be enough for very
detailed talk.
>
> Anyway, I'm just throwing this out here at this point to get some feedback
> and to see if there is any real interest. If there is, and I can think of
> something useful to cover, I'd probably be looking at an April talk.
>
> Madi

Anyway, thats just my 2 cents
> --
> The Toronto Linux Users Group.      Meetings: http://gtalug.org/
> TLUG requests: Linux topics, No HTML, wrap text below 80 columns
> How to UNSUBSCRIBE: http://gtalug.org/wiki/Mailing_lists
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://gtalug.org/pipermail/legacy/attachments/20091217/f8439f0d/attachment.html>


More information about the Legacy mailing list