> -----Original Message----- > From: Chip Childers [mailto:chip.child...@sungard.com] > Sent: Thursday, July 18, 2013 6:49 AM > To: dev@cloudstack.apache.org > Subject: Re: [DISCUSS] Upgrade path to ACS 4.2 from CCP > > On Wed, Jul 17, 2013 at 11:43:49PM +0000, Animesh Chaturvedi wrote: > > > > Folks > > > > We have an upgrade path from CCP 3.0.2 to 4.0 ACS and since then we > have diverged. I would like to propose adding upgrade path from later > CCP versions to ACS 4.2 which provides customers the choice to move to > ACS from CCP and also brings CCP closer to upstream ACS releases. > > > > The proposed upgrade paths are CCP3.0.4, CCP3.0.5, CCP3.06, CCP3.0.7 > to ACS4.2. Please provide your feedback. > > > > Thanks > > Animesh > > > > > > > > > > Generally... +1000. I think that this is a wonderful idea. > > I have a couple of thoughts: > > We will really need to implement the "checkpoint" > version concept that Hugo proposed earlier [1]. The reason that I say > this, is that the source and binaries for CCP aren't available to non- > customers of citrix / non-citrix devs. [Animesh>] Agreed I also like the checkpoint, as Hugo points out our dbupgrade are sequentially applied so not a coding issue but certainly matters to QA > > I obviously assume that Citrix is officially OK with this (and you are > basically representing that to the list). Given that "ok", I further > assume that Citrix employeed engineers would do the work (since they > have access to the CCP releases). > > Once the upgrades are in ACS, we would want to reduce the testing effort > around them going forward by ensuring that the targeted ACS release that > we test the upgrades to will be one of the "checkpoint" releases per > Hugo's proposal. [Animesh>] Agreed > > Last thought... this doesn't seem like a 4.2 change to me. Isn't it > too late to deal with testing it now? [Animesh>] Yes this seems late for 4.2, but it makes it easier to add support now. If there is objection we can track it for 4.3.
If not, and there are people > ready to roll, then no worries. I'm just pointing out what may be > obvious: you've hit feature freeze, and this seems like a significant > amount of work. > > -chip > > [1] http://markmail.org/thread/p34jbalr25pwocez
