oVirt?
- Thread starter voodooFX
- Start date
Notice: Page may contain affiliate links for which we may earn a small commission through services like Amazon Affiliates or Skimlinks.
I have oVirt deployed nested on top of my vSphere infra, works well, KVM goodness under the hood through a web portal wrapper (oVirt engine). Not entirely happy how storage worked out (I'd have to boot up and play w/ env but there were some oddities to what I was used to for shared/clustered stg in the vSphere world compared to how oVirt managed this type of storage). Seemed solid from what I tested, I think I had a plex and win7 instance deployed on it and it was zippy. No where near as polished as vSphere but what is?
Last edited:
I am trying to get an all in one setup going but so far not alot of success. It maybe me, or the non standard setup (I have only one server, and am trying to do a hosted engine install) but lots and lots of issues. To the point that I am almost ready to drop all and try esxi or xen server. My main problem whith any setup is that I want to use btrfs pools on the same server for data store. Proxmox or ovirt is the only two I know thay have support for that.
oVirt is a pretty cool system. I had it running for a while on a single CPU host without that many VMs, I can't claim much knowledge on the system, but from my perspective it works pretty well that said not everything is rosy either.
I experienced a few issues, namely the hosted engine crashing (for reasons that I never really got to the bottom of) and some issues with my gluster volumes. None of those were big problems, but for some reason it was hard to figure out what the exact problem was, from my experience the logfiles were cluttered with useless debug information and not that easy to find. But if you chose to use oVirt YMMW, I did not exactly spend a lot of time trying to solve these issues
For the record, I got bored waiting for oVirt to implement PCI passthrough functionality (it is to be implemented in oVirt 3.6, which is in RC1 at the moment) and have decided to see if I can make plain old virt-manager (yay for simplicity) do the trick perhaps with a few extra additions such as LXD and openvswitch. I may yet revert my decision and go back to oVirt when 3.6 has been released though.
I see no reason why you shouldn't be able to modify it slightly with a btrfs filesystem underneath the local gluster storage.
I experienced a few issues, namely the hosted engine crashing (for reasons that I never really got to the bottom of) and some issues with my gluster volumes. None of those were big problems, but for some reason it was hard to figure out what the exact problem was, from my experience the logfiles were cluttered with useless debug information and not that easy to find. But if you chose to use oVirt YMMW, I did not exactly spend a lot of time trying to solve these issues
For the record, I got bored waiting for oVirt to implement PCI passthrough functionality (it is to be implemented in oVirt 3.6, which is in RC1 at the moment) and have decided to see if I can make plain old virt-manager (yay for simplicity) do the trick perhaps with a few extra additions such as LXD and openvswitch. I may yet revert my decision and go back to oVirt when 3.6 has been released though.
I was able to setup an ovirt system using this guide: Up and Running with oVirt 3.5 | Gluster Community Website
I see no reason why you shouldn't be able to modify it slightly with a btrfs filesystem underneath the local gluster storage.
I have several issues with all of my test tries so far. Btw. I do mot want to use gluster. I want simple nfs share storage. Gluster for me is just another complication on top of everything else. But more importantly in every build I did I have issue with multipath. As soon as I run engine deploy I start to get multipath erros which only get worse as I add my multidisk btrfs pools. It is weird thing but I can not even install the centos 7 when all my disks plugged in in the system. Install give me an error and quits. I figure out it was my btrfs pools causing errors and now when I want to start anew I unplug all disks but the system and sysyem storage one, run the install and then plug in my btrfs pools back.
My other issue is that after I do an engine setup, I am loosing ssh access on main nic. I have 4 nics on thw system 2 intel and 2 nvidia mpc22. Can not remove intel as they are the only one cent os finds on install. The onboard nics need to be setup after with forcedeth driver from elrepo. So after I finish setup I loose ssh on both intell nics as I use them as main interfaces, and multipath gose crazy. So far I had not found a solution for aky of this.
My other issue is that after I do an engine setup, I am loosing ssh access on main nic. I have 4 nics on thw system 2 intel and 2 nvidia mpc22. Can not remove intel as they are the only one cent os finds on install. The onboard nics need to be setup after with forcedeth driver from elrepo. So after I finish setup I loose ssh on both intell nics as I use them as main interfaces, and multipath gose crazy. So far I had not found a solution for aky of this.
Hi, I'm Sandro Bonazzola from oVirt integration team. I suggest you to join the oVirt user mailing list ( Users Info Page ) for discussing your issues while testing oVirt.
We'll be happy to help you going over your issues.
Do you mind open a bug on Log in to Red Hat Bugzilla about this?
Same here, can you please provide full sos report of the system?
I would post the report if you tell me how to get it. I maybe attempting an advanced setup but I am practically at almost a noob level when it comes to Linux. As for filing report at red hat there is several reports there already. I am not thw only one and not the first one who have issues with multipath.
about the sos report, you can just use "sosreport -a" or use the ovirt tool called ovirt-log-collector, "yum install ovirt-log-collector" and then "engine-log-collector collect"
I figure it might be worth mentioning that oVirt 3.6 was released today: release notes