On Oct 18, 2012, at 5:06 PM, Joe Brockmeier <j...@zonker.net> wrote:
> On Thu, Oct 18, 2012 at 03:00:03PM -0700, Edison Su wrote: >> Nobody complains OVM doesn't work in 4.0 before, means nobody use and >> test it on 4.0 branch since half year ago when we starting to work on >> 4.0 release. > > "Since half a year ago"? > > Let's be realistic. How long has the 4.0 branch been in a state that > people could test it, if they're not developers? Two months? It's not > half a year by a long stretch. (Remember, we're talking about users and > folks deploying CloudStack, not folks who are watching git and compiling > it themselves.) > > When 3.0 was released, upgrading from 2.2 was not supported: > > http://www.cloudstack.org/blog/117-cloudstack-acton-released.html > > So let's recap: > > 4.0 hasn't been in a state to test realistically for more than two > months - and that's being generous. > > The 3.0 release was not recommended for upgrades - so 2.2 users who did > have OVM support wouldn't have tried to upgrade. Shortly after, the move > to Apache was announced and the open source version effectively froze at > 3.0.2. It's also not like every 2.2.x user has rushed to upgrade > immediately. > > At least one user has spoken up on cloudstack-user to voice support for > OVM. I don't know how many folks don't read the lists and aren't aware > of the potential for it to be dropped. Having been through this scenario > before when openSUSE dropped PPC, I'm guessing we will hear from folks > after the 4.0.0-incubating release actually hits the streets and the few > OVM users who were hoping to upgrade can't. > > If nothing else, I hope this is the only time this kind of feature > attrition happens without giving folks due notice to pick up the > feature. > -- > Joe Brockmeier > Twitter: @jzb > http://dissociatedpress.net/ I think I'm the one you're referring to speaking up on the users list. We currently are not using OVM under Cloudstack but it's on our near term project list (we were hoping in the next month to start that piece). The Oracle Management tools are an abortion at best, so Cloudstack was very attractive because we could pull all our virtualization management under one roof. So I'm 100% in favor keeping OVM support, however, not being a developer means I can contribute by submitting bugs, feature requests, etc but I'm not able to take any kind of development ownership of this feature. Not being a developer, I don't know the answer to this from a code stand-point, but OVM is a fork of Xen, how much different are they to support? My points were similar to what Joe has been saying, if it's dropping, that's fine but even in 3.x, the claim was there that it supported OVM (wether it worked or not doesn't matter, IMO). If it's not in 4.0, I also don't think that's a deal breaker for us, but I'd like to know that it will be in 4.0.x or even 4.1, just that the support will continue to be there. I also brought up the concern that if there is a customer currently using OVM, and they upgrade to 4.0, they'll have a broken environment. So the upgrade ought to include something that checks for OVM use, if it's there it fails to upgrade. Users may be upset that they can't upgrade, but atleast their environments won't be broken.
