On 06/18/2013 10:46 AM, Karli Sjöberg wrote:
tis 2013-06-18 klockan 10:06 +0300 skrev Moran Goldboim:
On 06/17/2013 02:51 PM, Karli Sjöberg wrote:
mån 2013-06-17 klockan 10:11 +0003 skrev Alex Lourie:
On Fri, Jun 14, 2013 at 12:26 PM, Karli Sjöberg <Karli.Sjoberg@slu.se <mailto:Karli.Sjoberg@slu.se>> wrote:
Hi!
Hi Karli,
More than two months have passed without any notices or updates on this issue, and looking at the bugreport absolutely nothing has been done since then. What´s confusing me is that the release management for oVirt 3.2 has this clearly printed: "MUST: Upgrade from previous release"
And you all seem too busy with getting the next release out of the door before fixing the current... What gives? Don´t you care about early adopters?
I apologise that this issue wasn't taken care of yet. It is clear that the priorities of developers shifted towards work on features in 3.3 release, which feature freeze is happening in about 2 weeks, so you can understand the time pressure they all have. There's no need to be angry about that. It is an open source project and as resources for the development are limited, we're trying to do our best to cover as much issues as possible. This can lead to situations as this issue, which may be left behind for some time.
I´m not angry, as I stated before, I´m more confused, and maybe a little disappointed that important(in my opinion) pieces of a projects continuum have been ignored like that. I´m having a hard time understanding the reasoning behind the priorities that have made this issue appear. In my mind, making sure that users of any project are able to follow between minor upgrades is a given, with a schedule like; 1) Feature freeze, 2) Fix release blockers (upgrading being one of them), 3) Release, in that order. Now what´s happening is that you are kind of hoping to drag this out until the problem goes away by itself when you decide it´s time to stop supporting 3.1.
That said, the process for upgrading the oVirt 3.1 on Fedora 17 to oVirt 3.2 on Fedora 18 is complex and non-trivial, that's why it is taking time to resolve all the issues that come up in the process.
If this matter is not extremely urgent, we are currently working on a tool that may make this upgrade easier, albeit, maybe, not completely automatic.
On the other hand, if this is a completely and utterly urgent, we may try to guide you through possible manual attempts to perform the upgrade. Mind though, that they weren't thoroughly tested, and not promised to go smoothly or even work, and we may have to think about possible alternative solutions on the go.
Straight answer: Am I ever going to be able to upgrade from 3.1 to 3.2?
My concern (fear) is about being left behind, abandoned. It makes me feel non-important. Just like any admin, the work put down to make Templates, to set up quotas, create VMPools, a dozen of guests. And then have to export everything (nullifying the idea of "Thin Provisioning"), redo every Role, reset every permission, quota, and so on, isn´t (in my opinion) viable in the long run for oVirt as a project to demand. That´s why I had the notion that upgrading surely had to be a high priority. Was I wrong about that?
upgrade actually was high priority task, but due to lots of platform (fedora 18) changes, it became a huge operation to support it (which comes on top new features/bugs/support etc.). in addition we have noticed that the interest from the community regarding it, is very limited.
from where i see i we have couple of options: -if there is enough interest from the community we can give it another go (though i think most of it moved long ago to oVirt 3.2)
You certainly have my interest at least;)
-we can help individuals by making a migration (3.1->3.2) document on fedora base systems- we'll probably need your assistance on testing side here
A document explaining how to upgrade would be completly acceptable for me. I´m not asking for a turnkey command that makes everything for me, I´ve managed to get quite far by myself any way. I would love to help testing the procedure, as I´ve stated before, we have set up a dedicated test environment dedicated for this specific upgrade task in mind, with engine, hosts and storage somewhat equal to the production system. Please, if there´s anything I can do to help, don´t hesitate to ask!
I think another consideration here is with the issues around upgrades with fedora (which breaks itself during upgrade), if recommending to deploy on .el6[1] (RHEL/CentOS/etc.) platforms wouldn't be better. [1] until other platforms are added - ubuntu is starting to look close.