|
I understand what you're saying somewhat but I still don't see why it is
wrong for someone to develop the same thing but from a different angle.
I realize that Red Hat is in a position in the market where they must
develop tools for the "IT manager". It seems to be their focus and
honestly (and I'm probably wrong here but I might as go ahead with it and
take the abuse) I think Red Hat is trying to somewhat secure themselves on
their efforts in the whole "Enterprise", "High Availability" effort and
probably doesn't like the competition in this area, especially from people
at VA Linux.
This is what I assume will happy, whatever package seems to get used the
most will probably end up having the competing developer then contribute
ideas. With the ideas you mentioned for piranha in the future, I'll be
the first one to give it a try when the changes are available, but as of
right now, I tend to like the Ultra Monkey approach.
I mean what you're saying certainly doesn't make sense for the "open
source community". Both project are GPL as far as I know. There are
3,010,599 irc clients for Linux that have somewhat different approaches to
the same goal. Why is there about 50 distro's available all based off of
RPM but with slightly different approaches and different installs
available. Should there only be Red Hat?
I fully appreciate the work Red Hat has done in this area. Trust me.
When Steven Tweedie's ext3 is available in a stable form, I will be
jumping up and down. Red Hat has the team to pretty much do anything they
want, so keep rockin' and let the people who actually have to use the
stuff decide.
-jeremy
> This posting was more useful than the previous one. I enjoy real
> discussions of experience and facts rather than just arbitrary position.
>
> Jeremy Hansen wrote:
> >
> > Basically at this point I've actually found Ultra Monkey a lot easier to
> > get running as opposed to piranha. Also Ultra Monkey seems to stay with a
> > pure philosophy. I may be wrong but I believe Piranha in order tow work
> > properly needs Red Hat's patches present in the kernel.
>
> It only needs the LVS and ipvs patches.
>
> > TO me this is easier. I'm sure it's not for everyone and those who have
> > piranha working I'm sure will stay with piranha, but I also don't
> > understand Keith's email as if he's opposing alternative. He can't
> > honestly say he wasn't aware of Horms' ultra monkey...I don't know, maybe
> > he wasn't but it certainly was mentioned on this list before.
>
> I am not against alternatives, but I am against duplicate effort and waste.
> It often does more harm than good to have 32 packages that all do the
> same thing, when a co-ordinated community effort would have produced a
> superior product. Piranha was open-source from inception. Why do yet
> another flavor instead of joining and improving what's currently an
> active project?
>
> > It's all basically the same goal. I think initially it's good to have
> > alternatives and then the users will decide which works best from their
> > experience.
>
> Actually, I believe that it is better to gather user requirements and
> create something that meets their needs. Alternatives only for the sake
> of alternatives adds confusion rather value. I'm not saying that piranha
> is the best or only solution, but I'm not convinced that even more
> clustering alternatives are now adding value.
>
> > Whatever comes out to be the most useful then should be
> > worked on exclusively to make it the best.
>
> IMHO, this is what was going on.
>
> > My goal currently is to be able to present people with an alternative to
> > paying high amount of money for things like F5's bigIP. I was truly
> > amazed to learn that not only is this thing just a bsd box, but it cost
> > somewhere around the $15,000 mark.
>
> Piranha is free and completely GPL. It has dedicated staff, vendor input,
> and community involvement (which is being geared up to be better once the
> web site is done). Wensong, the creator of LVS, is also on Red Hat staff
> and is part of the project. The only business reason to do something different
> is because a vendor is trying to do some sort of proprietary product
> association.
>
> > I also don't like how things on Piranha seem to have to be configured from
> > the front end like the web. Then I need php, apache installed on the lvs
> > machine itself and there's no reason for that.
>
> You are not forced to use the GUI, it is just recommended to prevent problems
> or get support from us (which you can choose not to do). It is likely upcoming
> releases will have a command line configuration file "checker" for those
> that don't want to use GUIs.
>
> > Convert that stuff into
> > understandable command line things and I think it would open a lot more to
> > some people.
>
> Most IT shops want GUI interfaces and remote system management. Piranha is
> being expanded to support N nodes, a single GUI will may it easier. Plus,
> whether you like it or not, most IT shops are not unix shops -- their clients
> are Windows. A web front end is a one-GUI-fits-all solution.
>
>
>
--
http://www.xxedgexx.com | jeremy@xxxxxxxxxxxx
---------------------------------------------
|
