Hello,
I don't run currently any oVirt deployement in production but I have a lab.
I used to run in production the old oVirt product (in rails).
My best wishes for the future release of oVirt are :
* GluserFS support as many of us
I choose oVirt because my first goal is to manage a virtualized datacenter with OSS. But we begin to look at private cloud deployement.
I think Aeolus would work with oVirt virtualization backend but AFAIK it only support redhat based linux which is not possible for us as we run
almost only debian server except for virtualisation layer. So we would like to deploy OpenStack but to rely on oVirt for KVM hypervisors.
* Fully supported stateless ovirt-node
Regards,
Romain
2012/6/28 Robert Middleswarth
<robert@middleswarth.net>
On 06/15/2012 06:23 AM, Dave Neary wrote:
Hi,
On 06/13/2012 04:28 PM, Dave Neary wrote:
So - how are you using oVirt? Why did you choose it over alternatives?
What do you like about it? and what would you like to see change,
whether that is in terms of technical, process, or marketing changes?
I'm here to help, but to do so I need your help first!
Thank you to all those who have replied, on and off list, so far. For those of you who sent me private messages, I'll be (anonymously) collating your feedback and forwarding it on.
The range of users who have replied so far includes:
* Sysadmin at small web hosting business
* Cost-sensitive IT department of an unrelated industry
* Hosting provider specialising in HA
* Running a private cloud
* Test lab set-up considering for production deployment
Well no one should be crazy enough to go live with a product they haven't at least ran inside a testing lab.
And the top features you've cited are:
* Stateless hypervisor
* Ability to migrate VMs
Number one reason I am working with oVirt
* RHEL and KVM
We are a debian based org so changing over to the RHEL based OS's is more a pain then a benefit. KVM is still kinda young compared to both Xen / Vmware it seems to work well but there aren't as many os's covered by the vitro drivers and there seem to be more bugs / race conditions but that has been steadily changing as it has been getting more mature
* Cost
* The ability to have your preferred OS as both hypervisor and guest as a first class citizen
* Aimed for data center use-case rather than cloud
This would be number 2 in the list.
And the top gaps you've identified so far:
* Insufficient resources (docs) to help with production deployment on ovirt.org
* Difficulty of configuration and getting started
* You'd like to see a more diverse contributor community
* Stability (unfortunately, I don't have any concrete examples of this from the commenter)
* History on resource usage in hypervisors and guests
* Integration with Gluster
* Offer choices of guest agents with other distributions than RHEL
I could have created this list myself. I have hit pretty much every one of these limits in the last few months working with the project. 3.1 adds limited support for Gluster and ovirt seems to be more stable dispute F17 instability.
As for the question of stability the file storage system in 3.0 can be a bit unstable. If your NFS share disappears for a few mins the file system tends to go offline and wont reactivate. Not sure about iscsi or FC since I don't have access to those file systems.
This is all giving me great insight into who's here - please keep it coming!
Cheers,
Dave.