KVM / Red Hat Virtualization Management
-
@FATeknollogee said in KVM / Red Hat Virtualization Management:
@DustinB3403 said in KVM / Red Hat Virtualization Management:
@FATeknollogee said in KVM / Red Hat Virtualization Management:
@DustinB3403 said in KVM / Red Hat Virtualization Management:
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@EddieJennings said in KVM / Red Hat Virtualization Management:
@black3dynamite said in KVM / Red Hat Virtualization Management:
oVirt web management UI looks very similar to what RHEV uses.
I'm surprised Red Hat doesn't seem to be promoting whatever they use for managing RHEV. I figured they'd try to lure folks away from VMware or Hyper-V with a slick interface
Ovirt is the upstream of RHEV.
So you'd install 1-300 oVirt nodes (on your physical and individual servers) and create a gluster pool out of them?
I actually looked at this a while ago and got frustrated with their documentation because and very specifically it jumps across all of the options to getting ovirt to work.
Ah no, that's not how oVirt works.
Can you explain then?
Ovirt has their own iso. So presumably you'd install that on your hardware as a node over and over.
If you're talking Gluster, that usually refers to a "hyperconverged" install with HE (Hosted Engine).
These days, yeah. Didn't used to.
-
@FATeknollogee said in KVM / Red Hat Virtualization Management:
If you're talking Gluster, that usually refers to a "hyperconverged" install with HE (Hosted Engine).
Yes that is what I'm specifically talking about.
@FATeknollogee said in KVM / Red Hat Virtualization Management:
This'll give you some idea: https://www.ovirt.org/blog/2018/02/up-and-running-with-ovirt-4-2-and-gluster-storage.html
In the guide, he specifically installs oVirt Node on his 3 "physical" boxes, as I mentioned above.
-
@DustinB3403 ovirt has an orchestrator an a number of virtualization nodes.
Orchestrator is like XO, nodes are like xcp. You add the node iso on compute nodes and then you run the orchestrator either as a physical box a vm on a node or a number of nodes.
Gluster is used for SDS. You can use it with iscsi target as VSAN but this is an extra feature. It is not required to run ovirt.
-
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@DustinB3403 ovirt has an orchestrator an a number of virtualization nodes.
Orchestrator is like XO, nodes are like xcp. You add the node iso on compute nodes and then you run the orchestrator either as a physical box a vm on a node or a number of nodes.
Gluster is used for SDS. You can use it with iscsi target as VSAN but this is an extra feature. It is not required to run ovirt.
So in all of the documentation on their websites, when they are discussing installing a Node, they mean to create a new hypervisor.
And on any one of those hypervisors, setup Orchestrator to manage the whole thing with vSAN (since we're discussing hyperconvergence).
-
@DustinB3403 yes. Exactly. Installing a node= put the virt engine on the node. Like installing xcp or so.
Hyperconverged means each or some nodes exposes a gluster brick to create shared storage.
Than on that storage (couple of years ago exposed via iscsi, dont know anything about current setup) you create a special VM which is the orchestrator (XO) -
@matteo-nunziati simple enough.
-
@DustinB3403 it is like putting XO in a VM inside xcp. The major difference here is the gluster layer.
-
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@DustinB3403 it is like putting XO in a VM inside xcp. The major difference here is the gluster layer.
Which a lot of people do all of the time. Could you install it as bare metal on a separate host, absolutely. Should you, meh depends on what you need.
-
Which XCP-ng has a gluster layer solution of their own now as well. Though I don't know if it's available as an OSS approach.
-
@DustinB3403 said in KVM / Red Hat Virtualization Management:
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@DustinB3403 it is like putting XO in a VM inside xcp. The major difference here is the gluster layer.
Which a lot of people do all of the time. Could you install it as bare metal on a separate host, absolutely. Should you, meh depends on what you need.
You can do the same with ovirt. The gluster layer is for redundancy. In a First lab you can directly put ovirt into a single Vm into a single node.
-
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@DustinB3403 said in KVM / Red Hat Virtualization Management:
@matteo-nunziati said in KVM / Red Hat Virtualization Management:
@DustinB3403 it is like putting XO in a VM inside xcp. The major difference here is the gluster layer.
Which a lot of people do all of the time. Could you install it as bare metal on a separate host, absolutely. Should you, meh depends on what you need.
You can do the same with ovirt. The gluster layer is for redundancy. In a First lab you can directly put ovirt into a single Vm into a single node.
Just mind you have to put the orchestrator in a special state before node reboot. I didn't figured this out at the time and after the first reboot everything gone nuts.
I Abandoned the idea in favor of virtmanager + vanilla kvm. -
Red Hat has 2 (well, actually 4) products that manage KVM. The main ones are RHV/oVirt and RHOS/RDO (their Openstack distribution).
Besides these two there is the localhost-oriented libvirt with virt-manager, cockpit and boxes, and kubernetes oriented kubevirt.
So if you're looking for a vCenter cluster replacement, RHV is what you should be looking at (or if you prefer the less stable opensource version - oVirt). If you want to build something larger, at the scale of AWS, you need Openstack. For local stuff, e.g. virtualbox - virt-manager, boxes, and eventually cockpit will do the job, and if you want to run a VM in a k8s pod - kubevirt is for you.
-
@scottalanmiller RHV now, after a rebrand in 2016-ish
